It’s no surprise that, when faced with the words ‘artificial intelligence’, most people immediately think of the word ‘hype’ shortly thereafter. The technology has been much anticipated, and although the term has been around since the 1950s, the general public is still waiting to see it have a significant impact on their lives.

Recent numbers from the MMC report showed 40% of European startups classified as “AI startups”, showed no sign of using AI. Despite all of this, where there is substance, investment in the technology is booming like never before.

Tech Nation is proud to support and represent 29 of the most exciting AI companies in the UK. These companies have undergone a rigorous judging process, to help you cut through the hype, get to the core of their product and see what they’re really made of. Read more about them and how they were selected.

Why is AI booming? 

As the sophistication improves and potential for the application of AI broadens, confidence in the technology has massively increased. To get a holistic view of how and where AI is thriving, let’s look at worldwide investment, which has skyrocketed since 2017 and shown no signs of slowing down. Possibly, 2019 could be a record year for AI investment.

$3.5bninvested in AI in 2014

$4.5bninvested in AI in 2015

$6bninvested in AI in 2016

$16.5bninvested in AI in 2017

$19bninvested in AI in 2018

$16bninvested in AI in 2019

**To compare, in the year of 2010, AI barely saw $1 bn in investment worldwide

How is the UK comparing?

Investment in the UK’s AI technology has grown almost six-fold in the last four years, but how does it compare to the rest of the world? 2019 has so far seen the following numbers:

$9bnin the US

$3bnin China

$1bnin the UK

How many people do they employ?

300+Chinese companies

3500+US companies

800+UK companies

Why are our companies special?

Tech Nation ensures a thorough selection process based on expert criteria and judging panel. This year, we had the pleasure of hosting Marta Krupinska, Head of Google for Startups UK, Volker Hirsch (partner at Amadeus Capital), Dame Wendy Hall, David Kelnar of MMC, Tabitha Goldstaub (Chair of the UK Government AI Council) and Dr Mike Short (Chief Scientific Advisor, Department of International Trade) amongst others, as judges.

Companies who make it through the judging process are selected for the Applied AI programme, a highly specialised, six-month-long sector-specific growth programme that aims to provide peer to peer learning, expert classes and a long-lasting network. Judges had the following comments about the process and companies selected:

“It’s fantastic how the Tech Nation team managed to bring together some of the most brilliant minds in tech and AI for a lively discussion on who’ll be the next game-changing start-up. AI is all the rage now and a fantastic tool to support solutions for the world’s most pressing problems.”

Marta Krupinska, Google

“It was great to judge so many strong AI ideas that were tested and appeared ready to scale up. Tech Nation will have a tiger-by-the-tail to support some of these young companies with their hunger for success and imagination.”

Dr Mike Short CBE, Chief Scientific Advisor, Department of International Trade

"It was a pleasure to help select the cohort of Tech Nation's Applied AI programme. I was impressed with the quality and depth of the companies applying, which is a testament to the UK’s strength in this pivotal technology sector, which we at Amadeus Capital are keen and proud to support."

Volker Hirsch, Partner, Amadeus Capital:   

“I'm thrilled to see the launch of Tech Nation's first Applied AI cohort. It's great to see the strength of technical expertise and diversity of the companies. With 35% operating outside of London and nearly half of the cohort's participants having at least one female founder I think this represents the AI community well and exactly what we should be championing.”

Tabitha Golstaub, Chair, AI Council and Founder, CognitionX commented:

Less artificial, more intelligent

The Applied AI programme is doing things differently. It specifically looks for companies that aren’t just using AI for the sake of it, but apply it to solve real problems for businesses and society, with recognition and focus on social responsibility and ethics. Our companies specialities range between diagnostics and prevention of cancer and medical conditions, reducing fake news and misinformation, tackling climate change and waste management, autonomous vehicles and road safety, reducing burnout and improving mental health and identifying pay inequality.

Founder: Oliver Armitage, CSO

Location: Cambridge

Year founded: 2015

Core area of focus: HealthTech

Previous backing: Endure Capital

In a nutshell: BIOS is a leading neural engineering startup, creating the open standard hardware and software interface between the human nervous system and AI. Improving the quality of life for millions of people affected by chronic disease, they combine applied materials, machine learning, software and neuroscience with the experience of surgeons, clinicians and patients.

BIOS’ solution: Current care for chronic conditions is manual, time-consuming, constant, and often relies on at least one, if not many, prescriptions. BIOS aim to make the care of chronic conditions seamless and easy for both patients and clinicians. Their model will deliver SaaS-type healthcare, with new prescriptions delivered via algorithms, and a process that will cut costs, increase efficiency and deliver a high level of personalisation.

USP: BIOS has a number of major differentiators which gives it a significant advantage over potential competitors: An incredibly talented multi-disciplinary R&D team, the largest PNS (peripheral nervous system) neural data set in the world, with proven AI models for analysing neural data, a very strong IP portfolio and proprietary neural interface hardware.

Goals for the next 12 months:

• Overcome scaling pains
• Raise further investment
• Find talent 

Experience: Founders Oliver Armitage and Emil Hewage are technically credible and specialised in a rare blend of technologies. Oliver’s PhD was in BioEngineering and he is specialised in tissue interfaces and engineering materials. These materials allow technology to be fused with the body, and has already developed cutting edge patented tissue interface technology for BIOS. Emil studied computational neuroscience for his PhD research at the University of Cambridge. 

Traction: Winner of MassChallenge (London) in 2015 and successful completion of Y-Combinator in 2017. Bios was awarded significant grant funding, including Innovate UK, SBRI and i4i Connect. They secured major equity funding of over $4.5M and partnerships with UC London, University of Cambridge, New York University, Nvidia, Digital Health.London and IEEE. They own the largest PNS (peripheral nervous system) neural data set in the world with proven AI models for analysing neural data. 

Website | Twitter

Founder: Bhavagaya Bhaski, CEO

Location: London

Year founded: 2017

Core area of focus: HealthTech 

Previous backing: None

In a nutshell: C the Signs is an award-winning platform founded by doctors. Using artificial intelligence mapped with the latest evidence, the tool is used by Doctors to identify patients at risk of cancer at the earliest and most survivable stage of the disease. Covering the entire spectrum of cancer, C the Signs can identify which cancer(s) a patient is at risk of and, what test, investigation or referral they need in less than 30 seconds. The tool is currently being used by over 1000 doctors in the NHS and trials to date have demonstrated a 30x improvement in cancer detection rates.

C the Signs’ solution: Founded by two doctors, C the Signs is an award-winning digital tool that uses artificial intelligence mapped with the latest evidence to identify patients at risk of cancer at the earliest and most survivable stage of the disease.

USP: Covering the entire spectrum of cancer, C the Signs can identify which cancer(s) a patient is at risk of, and what tests, investigation or referral they require, all in less than 30 seconds. The tool is currently being used by almost 1000 doctors in the NHS and trials to date have demonstrated a 30x improvement in cancer detection rates.

Goals for the next 12 months

• Growth
• Internationalisation
• New product launch

Experience: The C the Signs team consists of experienced clinicians, designers, data scientists, engineers and a psychologist.

Traction

• Clinical study 2017 - 2018 (publication this year - 30x improvement cancer detection rates p<0.05).
• 7 NHS CCG Contracts (over 2M patients- rolling out to a further 23 NHS CCGs Jan 2019 - 10M patients).
• Over 25 third sector charity Partnerships. Registered Class I Medical Device, CE Marked.

Website | Twitter

Founder: Matthew Baker, CSO

Location: Glasgow

Year founded: 2016

Core area of focus: HealthTech

Previous backing: Mercia

In a nutshell: ClinSpec Dx use artificial intelligence to analyse blood for early detection of cancers and other diseases. They have developed the world's first cost effective blood test for the early detection of brain tumours. Studies have shown that their technology can bring savings to healthcare service providers as well as improved quality of life for patients.

ClinSpec Dx’s solution: ClinSpec Dx are pioneering a new spectroscopic method for blood serum analysis, allowing same-day detection and grading of a range of diseases, with first applications to brain cancer.

USP: Low cost, rapid diagnosis of brain cancer for the first time, using novel hardware that can improve patient survival and quality of life, whilst bringing savings to the health services.

Goals for the next 12 months

• Complete clinical study of 600 patients in NHS Lothian
• Begin large cohort clinical trial in multiple sites across the UK
• Secure Series A investment to complete regulatory roadmap and to launch test into the market

Experience: The core team has a strong track record in scientific achievement and commercial realisation. They combine the experience of commercial and academic backgrounds in order to push the translation of this technology.

Traction: Preliminary results from their primary clinical study were recently published in Nature Communications. In Q2 2019, they made their first sale on novel hardware for non-clinical applications. ClinSpec Dx was awarded Innovation Award at the Scottish Life Sciences Awards 2019.

Website | Twitter

Founder: Fouad Al-Noor

Location: London

Year founded: 2016

Core area of focus: HealthTech

Previous backing: EF, AI Seed Fund

In a nutshell: ThinkSono is an ultrasound AI company. AutoDVT is the world's first software that enables nursing staff to detect 'deep vein thrombosis' (DVT). DVT is a blood clot in the leg and is considered the number one cause of preventable hospital death.

ThinkSono’s solution: AutoDVT solves the DVT diagnosis problem by allowing all healthcare staff to detect DVT in just 10 minutes. It eliminates the need for a specialist (e.g radiologist or sonographer). Since AutoDVT enables nursing staff and junior doctors to detect DVT, it offloads a huge burden on not only themselves but also reduces cost and improves patient outcomes at the same time.

USP: Worldwide, there are very few ultrasound AI companies. And currently, only ThinkSono has a solution to detect DVT.

Goals for the next 12 months

• Carrying out clinical trials. 
• Ensuring their technology is as robust as possible. 
• Sign commercial deals with ultrasound companies. 

Experience: ThinkSono have a strong team of more than 20 members, including two technical founders (hardware and software), four machine learning staff (post-docs), radiologists, haematologists and emergency doctors. 

Traction: So far, ThinkSono has published a patent for AutoDVT and filed a new patent along with collecting the world's largest labelled dataset for DVT. They have letters of intent representing more than 50 hospitals and published a peer-reviewed paper at MICCAI conference. They are in formal discussions with the two largest ultrasound companies in the world (which represent 48% of the entire ultrasound market) and are ready to start clinical trials to show the efficiency of the technology at the end of 2019. 

Website

Founder: Ben Holland, CTO

Location: Cardiff

Year founded: 2017

Core area of focus: HealthTech

Previous backing: - 

In a nutshell: Antibodies are a very successful and valuable class of drugs. But discovering and developing them into new drugs is a laborious and unreliable lab-based process. Antiverse discovers and optimises new antibody candidates faster, meaning only better targeted and further optimised candidates are included in expensive clinical trials. With the top-selling antibody bringing in sales of $19.9bn in 2018, the potential is huge.

Antiverse’s solution: By taking the discovery and optimisation process in silico, Antiverse take data about a target molecule, such as the structure or sequence, from a customer and return the sequences of novel antibodies that bind to that target. They also provide antibodies optimised to possess whatever other properties the customer requires. 

USP: Nobody else is offering antibody generation fully in silico. Rather than require discrete lab work for every project, Antiverse do the lab work up front to build and train a fully-developed model capable of generating and optimising antibodies. Once mature, their models will be able to perform the whole antibody discovery and optimisation process in silico, at unmatched speed and accuracy.

Goals for the next 12 months

• Affordably and sustainably recruit more Machine Learning professionals.
• Raise investment. 
• Business/customer development, especially structuring of deals.

Experience: Co-Founder Murat has worked as a bioinformatician and software engineer. Co-founder Ben has an MEng from Oxford with a specialisation in Information Engineering, and has since worked extensively on mathematical modelling and neural networks.

Traction: Antiverse have generated new antibody sequences in silico, which are currently being tested by a major pharma partner. They are in discussions with another major pharma company about a potential licencing/co-development deal. They are working on a pilot project for a third pharma company, and have a Letter of Intent for another. 

Website

Founder: Elizabeth Clarke, CEO

Location: Manchester

Year founded: 2014

Core area of focus: Marketing & Advertising

Previous backing: -

In a nutshell: Dream Agility uses Machine Learning and Visual AI to manage every aspect of Google Ads. From using Visual AI to attribute rich data from images, to MLoD (machine learning on demand) to disrupt the agency relationship, pass on massive savings and increase uplifts.

Dream Agility’s solution: Creating Google ads can be time consuming, labour intensive and expensive. Dream Agility’s visual AI accurately processes thousands of images a day and produces the data required to process SKU, or that Google requires for its ads. Using Dream Agility’s software not only saves on time and man-power, but will lead to an increase in sales as a result of having more relevant accurate and consistent data. It also uses Machine Learning to make Google ads more effective, thus decreasing spend and increasing conversion.

USP: Agencies are diametrically opposed to the end advertiser in the use of MLoD . Using it to reduce the waste will impact on their revenues if they work on a percentage of spend model. Dream Agility allows users to get the maximum impact, all while decreasing spend, and it can be used alongside any other bidding tech without interfering with the suggestions it makes. It's the perfect partner tool.

Goals for the next 12 months:

• Refining their visual AI product.
• Overcoming market objections by agencies.
• Achieving direct adoption from end advertisers
• Completing a raise with the aim to scale at speed

Experience: Dream Agility’s Lead Data Scientist’s PhD thesis involved applied machine learning techniques in particle physics research. He was involved with the Higgs boson discovery at the CERN laboratory in Switzerland and has since gained experience using Machine learning applications with medical data to help improve decisions regarding cancer treatment. Elsewhere, Dream Agility have a team of proficient developers. The company was founded on the commercial brains of a CIMA qualified Economist with an MBA from Manchester and a Best Selling Flirting Expert who was instrumental in the flotation of a technology transfer company with a 400 patent portfolio for £50m who made FTSE 250.

Traction

• The Fragrance Shop saw Christmas sales up 3.6% year on year, with online growth of 19.3% and a 28.7% uplift in Black Friday sales. 
• Chemist Warehouse, Australia’s largest pharmacy group, cut Google Ads costs whilst at the same time increasing conversion value by 178% over their previous year’s performance. 
• My Generator, Australia’s largest online collection of power generators, saw sales increase by 425% year on year. The impact was so great that it captured the attention of national television, all while reducing their Google AdWords spend (64% reduction).
• Idle Man UK’s revenues went up 53%, ROAS improved by 138% YoY in December, and ROAS over all territories up 239%.

Website | Twitter

Founder: Steve Erdal, Director of Linguistics

Location: Gateshead

Year founded: 2017

Core area of focus: Customer service

Previous backing: North East Innovation Fund

In a nutshell: Wordnerds teach computers to understand language, make decisions and find actionable insights. Using AI and advanced linguistics, they are turning the internet into the world’s biggest focus group.

Wordnerds’ solution: Wordnerds is a Cloud-based SaaS platform. It gathers huge quantities of text data from social media, online reviews, survey and forum data, and customer service messages, to bring out the actionable insight so it is available at a glance to the user. It focuses on the structures and patterns of language as well as vocabulary, meaning stakeholders can cut through the noise and understand what consumers really feel.

USP: Traditional solutions tend to rely on n-grams (the "bag of words" model), where the only input of the AI is the presence of a word, and its physical proximity to other words, treating a word like a number and not picking up on human factors like misspellings, politeness or sarcasm. Wordnerds works with the structure and pattern of the language, rather than the vocabulary, creating a deeper understanding of their meaning, and a product that focuses on the end-user, and jargon-free customer communication. 

Goals for the next 12 months

• Balancing ideas and focus to create the perfect product-market fit.
• Getting their products in front of the right audience and people. 
• Streamlining processes to prepare for scaling. 

Experience: Pete Daykin is a serial entrepreneur with 15 years' experience of running Web Development and Digital Transformation Agencies in the North East. Pete has a BA from the University of Cambridge. Angela Daykin has spent the last 15 years in Finance Director roles in technology companies. Steve Erdal is a writer, linguist and Product Owner at Wordnerds. After studying Corpus Linguistics during an MA at Newcastle University, Steve formed his own linguistics-led marketing business in 2014. He has produced a range of linguistics-based research for a variety of HE organisations, including Oxford, Edinburgh, Durham, York, Queen’s Belfast, Bath and Cardiff.

Traction: Wordnerds was born out of a Nissan hackathon, where the team’s solution beat dozens of UK and global competitors. This won them the right to develop their product and their brand. They now work with a range of organisations, creating ROI from improving their customer service. Their MRR has moved from £3.5k to 22k and continues to rise. In addition, they won a World Innovation Award at the World Rail Festival in Amsterdam.

Website | Twitter

Founder: William Ritchie, CEO 

Location: London

Year founded: 2018

Core area of focus: HR Tech

Previous backing: Ascension Ventures

In a nutshell: Chosen AI is a people analytics platform that uses advanced Natural Language Processing and Deep Learning to help employers identify and optimise the potential of their existing employees. They have built a global contextual skills knowledge base that connects with existing HR systems to generate predictive insights that help align people strategies to business strategies.

Chosen AI’s solution: Chosen AI has built a unique technology and data platform that uses Natural Language Processing and Deep Learning to identify people’s proven skills and the context behind where and how they acquired them, making it possible to understand that the same job title at different companies often involves learning and using entirely different skills. As opposed to existing solutions, Chosen AI’s big data approach also takes into account people’s entire employment history, rather than just analysing their current role. 

USP: Chosen AI uses advanced natural language processing to extract skills from existing employee data (CVs, cover letters, performance reviews, work) and a global contextual skills knowledge graph to predict skills where data is missing. Their approach ensures skills are always representative of the current situation and have complete employee coverage without the need for any employee input.

Goals for the next 12 months:

• Building a repeatable and scalable enterprise sales process.
• Scaling the team, specifically design, product, sales and data science
• Customer education and content marketing

Experience: Chosen AI’s CEO Will Ritchie spent 4 years in Tech M&A as an investment banker, advising the corporate strategy of some of the best-known global tech businesses. Before founding Chosen AI, he successfully launched and scaled Joblift. Charis Sfyrakis, Chief Scientific Officer, has 10 years of experience in high-performance computing and big-data science. Ashleigh Otter (COO) previously managed the accelerator programme at Founders Factory, mentoring over 50 early-stage businesses on their growth. Dr Jeffrey Ng, Chief Scientific Advisor & Company Director, has a doctorate in Machine Learning and Computer Vision.

Traction: The team has spent the last 3 months doing intensive customer discovery to ensure they fully understand the end-user and value proposition of their technology. Off the back of these interviews, they’ve had interest for PoCs, without starting an active sales push. This includes a Fortune 500 FMCG, a Fortune 500 Pharma, a Fortune 500 Insurer, a 15,OOO person Italian FMCG and a Fortune 500 FMCG. 

Website

Founder: Dr Zara Nanu, CEO

Location: Bristol

Year founded: 2017

Core area of focus: HRTech

Previous backing: None

In a nutshell: Gapsquare enables companies worldwide to track pay disparity, pay equality and pay gap data instantly and sustainably, building fairer workplaces and simpler processes around fair pay. Gapsquare offers AI insights around building an inclusive workplace, and moving pay gap and equality analytics in-house and within the control of the company

Gapsquare’s solution: Their flagship product, Gapsquare FairPay®, is an easy-to-use cloud based software as a service. It allows businesses and organisations to analyze and track pay disparity statistics and produce instant reports on equality and diversity data, ready to be presented to teams. Companies running their HR and Payroll data through their software get instant in-depth knowledge around pay gap and pay equality data, helping them to both comply with, and go beyond government regulations.

USP: Gapsquare offers an integrated approach to pay analysis, allowing organisations to continuously monitor their data around disparities and/or any potential legal issues which may arise. Their software embeds an equality and diversity narrative that would normally come through costly consultancy. Within the tool, you effectively have an on-demand consultancy service as well as the ability to embrace continuous monitoring.

Goals for the next 12 months

• Rapid market takeover.
• Ensuring AI insights are directed at the right level within organisations using the platform.
• Bias and ethics in AI - both algorithms and data.

Experience: Gapsquare’s CEO, Dr Zara Nanu, is an expert on how diversity and inclusion can shape more dynamic and productive teams and a more engaging and empowering workplace. Chief Data Officer and Co-Founder Ion comes from a background in artificial intelligence and machine learning. The VP for Partnerships and Growth, Sian, has a background that spans both the corporate and third sectors, and their Chief Technology Officer, Ant Kennedy’s background spans from rapid prototyping, growth engineering and architecture to transitioning ML research to client-facing products.

Traction: Gapsquare have had traction with large customers including Serco, Babcock, London Mayor’s Office, IDC and London Metropolitan Police have also signed a 3 year commitment. Their sales channels are law firms, and large software providers of HRM such as Oracle. They have over 30 organisations signed up to their platform and performing pay gap analysis/equal pay audits.

Website | Twitter

Founder: Nigel Winship, CEO

Location: Bristol

> Year founded: 2018

> Core area of focus: HRTech

> Previous backing: None

In a nutshell: People Matter Tech’s mental wellness platform is a safe online space that measures and improves mental wellness at work, without surveys. Their AI learns who you are and how your digital environment impacts your burnout risk, offering personalised recommendations.

People Matter’s solution: They underpin data science with a proprietary model called PEAK. This model is based on established behavioral science, primarily the job demand resource model. The PEAK model underpins and provides a solid frame of reference for their machine learning work. 

USP: People Matter's approach is award-winning (Tech Nation Rising Star 2019), and disruptive. No other solution on the market analyses work-based communication data. They combine HR data and data gleaned from a mental wellness app to identify signs of burnout. This facilitates a read on corporate culture and organisational mental health.

Goals for the next 12 months

• Successfully delivering their first product to a FTSE 100 business, then building out customers 2 and 3. 
• Securing funding to facilitate ambitious growth plans in 2020/21. 
• Growing commercial pipeline to facilitate growth in 2020.

Experience: Nigel Winship, co-founder and CEO is an experienced technology entrepreneur who has built and sold two successful technology companies since 2001. Amy King, co-founder and Research/Customer Success Director, is a passionate business psychologist who has worked with FTSE 100 organisations around the world. She has spent the past 8 years consulting with clients to help transform the world of work.

Traction: People Matter signed a research partnership with Experian in October 2018, presenting initial research findings in March 2019. They commenced Okina product build in March 2019 and signed draft commercial terms with Experian in April 2019. The project kicked off on 6th June 2019.

Website | Twitter

Founder: Dr Darko Matovski, CEO

Location: London

Year founded: 2016

Core area of focus: BI & Analytics, Marketing & Advertising

Previous backing: AI Seed Fund, IQ capital

In a nutshell: causaLens is building the next generation of autonomous predictive technology for complex, dynamic systems and presence in key global markets. With a team that consists of scientists and engineers at the top of their game, their technology processes data to understand how the global economy will develop.

causaLens’ solution: The next generation autonomous predictive technology for dynamic systems.

USP: causaLens specialise in time-series and dynamic systems. Their technology adapts to a dynamic environment and is fully autonomous (from 'dirty data' to optimal model in production). Production ready 'out-of-the-box' and training and inference run at the same speed (even in high frequency use-cases).

Goals for the next 12 months

• Hire 50 PhDs and engineers
• Scale across industries
• Grow customers from financial services

Experience: causaLens’ co-founders have experience in top research institutes and quant funds. They are a full-time team of 20, the majority of which have PhDs from top schools.

Traction: From launching their first product in August 2018, causaLens have already attracted some of the most sophisticated banks, hedge funds, asset managers, exchanges, market makers & data providers as customers.

Website | Twitter

Founder: James Dean, CEO

Location: London

Year founded: 2015

Core area of focus: Construction

Previous backing: Force over Mass, Roundhill Capital and Tencent

In a nutshell: SenSat builds digital simulations of the real world to help computers solve complex problems. Their simulated reality platform Mapp® allows companies operating in physical domains, such as infrastructure construction, to make more informed decisions based on real site data.

SenSat’s solution: SenSat builds technology to translate the real world into a version understandable to AI. This helps offline industries analyse their environments to learn how things work and improve the way we make decisions.

USP: Huge scale spatial data integration.

Goals for the next 12 months

• Add value to their data lake
• Advance machine understanding of real world context
• Achieve £500k MRR sales on Mapp®

Experience: The team has a rich portfolio of qualifications, experience and expertise, with areas of knowledge spanning astrophysics, satellite sensing and earth monitoring, geospatial technologies, clinical psychology and years of experience in media, sales and SaaS. 

Traction: 32 Tier 1 Infrastructure clients.

Website | Twitter

Founder: Jamie Potter, CEO

Location: London

Year founded: 2016

Core area of focus: Industrial Automation and Productivity

Previous backing: EF, Backed VC, Join Capital

In a nutshell: Flexciton uses AI to optimise production scheduling for manufacturers. A factory production schedule dictates what should be manufactured, how and when. Creating a production schedule for just one week typically presents more than a trillion different options. Flexciton software searches through all of those options to generate the best schedule possible.

Flexciton’s solution: Flexciton’s AI creates a digital twin of the user factory which is capable of simulating the behaviour of that factory. It can then calculate a production schedule by searching through the trillions of different options to find the most efficient. This optimised schedule is provided to the user in a simple web application. On average this has resulted in a 10% increase in efficiency for each factory (this could be by increasing output, reducing inventory, reducing cycle time etc), and one client made savings of around £50m per year per factory.

USP: The time taken to calculate production schedules and the efficiency of those production schedules (and the resulting factory efficiency). Flexciton’s AI technology can produce optimised and hyper-efficient production schedules extremely quickly. 

Goals for the next 12 months

• Productising in a way capable of providing fast growth.
• Providing support for a mission-critical enterprise solution which runs a factory.
• Growing quickly whilst continuing to hire top-class talent and ensuring they all work well together.

Experience: Flexciton has world-leading technical experts in their domain. Prior to founding or joining Flexciton, the team published over 140 academic peer-reviewed papers on deep tech, from case studies with ABB, Unilever, Heineken, BASF and more. They also have a world-leading expert who has published a textbook on production scheduling in manufacturing. The Head of Product comes from a scheduling optimisation company where he spent the last 18 years and was previously their CTO.

Traction: Flexciton has experienced overwhelming demand for its technology for the past few years. Whilst the product is still relatively early, paying customers include Seagate Technologies, Unilever, Burberry, Alex Begg, Camira Fabrics, Finnebrogue Foods and Adient. The commercial product will be ready to be bought by other non-early adopter clients in early 2020 and there is a significant pipeline of prospective customers interested.

Website | Twitter

Founder: Dr Richard Ahfeld, CEO

Location: London

Year founded: 2016

Core area of focus: Industrial Automation and Productivity

Previous backing: Pentech, Ascension Ventures

In a nutshell: Monolith AI is a team of engineers, data scientists and software developers that work together to accelerate engineering through the use of AI. They have developed a software that combines AI, tailored visualisation tools and high-performance computing to enhance the productivity and knowledge of engineers, predict results ahead of time, and accelerate product development.

Monolith’s solution: Monolith’s product is an online platform that easily gathers all user data in one platform, giving a much better understanding of their data and their product. It does this through the use of interactive data exploration tools and learns from past simulations or experiments to instantaneously predict the behaviour of a new product through the use of ML tools. This facilitates decision making by easily sharing results with other teams and clients

USP: Monolith focuses on solving Engineering problems with AI and as a result, they have developed a platform with a unique set of features specifically developed for engineers. The platform features AI algorithms for CAD (3D geom) data, uncertainty and visualisation tools tailored to clients. 

Goals for the next 12 months

• Start more pilots.
• Converts pilots into recurring revenue.
• Continue to develop and update technology to keep a competitive advantage.

Experience: As Monolith wants to combine AI with Engineering, their team is a combination of experts in both fields, with half of the team having PhDs. The team's extensive experience in AI and engineering was acquired through years of research at world-leading universities (Cambridge, Oxford, Stanford, Imperial College London) and collaboration projects with some of the world’s most advanced engineering companies (NASA, Rolls Royce, Airbus).

Traction: Monolith has piloted with clients from various engineering backgrounds, including product packaging, car performance and aircraft structural prediction, ensuring structural integrity for different flight scenarios.

Website | Twitter

Founder: David Levine, CEO and Founder

Location: Manchester

Year founder: 2013

Core area of focus: Retail

Previous backing: Maven Capital, NPIF Equity

In a nutshell: DigitalBridge is a Guided Design platform that enables retailers to help customers design and buy their dream bathroom or kitchen. Integrating directly into a retailer’s website, DigitalBridge guides customers through the entire process, from concept and design to visualisation and completion. By reducing friction and difficult decisions, they enable retailers to shorten sales cycles and increase conversion and revenue.

DigitalBridge’s solution: DigitalBridge has built a mobile-first, web-based, consumer-focused tool that is embedded into a retailer's website (not mobile app), which guides the retailer's consumers thought the bathroom and kitchen design journey.

USP: Unlike their main competitors, DigitalBridge Aspect is primarily focused on the consumer. Creating a compelling experience that allows the consumer to design their dream bathroom or kitchen with little input from a professional designer requires a very deep understanding of the complexities involved.

Goals for the next 12 months

• Setting up a US office
• Recruiting additional staff
• Growing whilst retaining the culture of the business.

Experience: DigitalBridge has a very experienced senior management team, including CEO David Levine who has run the business since its founding and was previously Global Head of Connected Car at Vodafone Group. The board includes individuals such as Stuart Marks who runs corporate accelerator L Marks, Jonathan Well (ex Chief Digital Officer at MissGuided) and Mike McGrath who is chair of a number of eCommerce businesses.

Traction: DigitalBridge has two live deployments: B&Q in the UK (planmybathroom.diy.com) and Castorama in France. They currently process over 40k consumer projects every month.

Website | Twitter

Founder: Kirsty Lloyd-Jukes, CEO

Location: Oxford

Year founded: 2017

Core area of focus: Mobility and Autonomous Vehicles

Previous backing: Oxford Capital

In a nutshell: Latent Logic ensures that autonomous vehicles can coexist and interact safely with humans in the real world, at scale. A spin-out of the award-winning Computer Sciences department at Oxford University, they use machine learning to build realistic simulations of motorists, cyclists and pedestrians - critical for the development, testing and certification of autonomous vehicles

Latent Logic’s solution: Building virtual humans is really difficult. Humans don't always follow the rules of the road (90%+ of accidents are caused by human error), and we all behave differently - across different countries, cities, time of day, weather conditions - every one of us has a different risk profile. Realistic virtual humans, displaying bad behaviours as well as good ones, are critical for anyone building an AV, or needing to test it and prove its safety. Latent Logic is building state-of-the-art machine learning to solve this problem, creating realistic, diverse and massively scalable virtual humans.

USP: Latent Logic has 4 major aspects to their USP. The realism and diversity of their behaviours due to deep-learning from real traffic data, their expertise in Imitation Learning at scale, their desire for objectivity, and their access to data. Latent Logic have developed strong partnerships with traffic camera owners, and gone through GDPR and data-sharing agreements, which new entrants to the market will not have.

Goals for the next 12 months

• Sign up their next three big AV developer customers
• Raise Series A
• Publish next paper in an Applied AI top conference 

Experience: Their CEO has strong credentials in automotive and entrepreneurialism and set up AA’s award-winning connected car product, Car Genie. Their CTO, Joao has 10 years experience in robotics and machine learning, and his passion for best practises is critical for commercialising AI. Their CSO, Shimon, is also a full-time professor at Oxford University in Computer Sciences, and a renowned expert in Imitation learning. The team values transparency, open-mindedness, practicality and pioneering, 

Traction; Latent Logic is in the process of delivering two major projects with the UK Department of Transport - OmniCAV and VeriCAV. OmniCAV is the world’s first to include rural as well as urban roads in a simulation to test an AV. They are also mid-way through delivering their first paid customer pilot, which is an integration of virtual humans into an AV developer' simulation platform. Finally, they published their first paper this year, which was accepted at the prestigious International Conference of Robotics and Automation.

Website | Twitter

Founder: Maya Pindeus, CEO

Location: London

Year founded: 2017

Core area of focus: Mobility & Autonomous Vehicles

Previous backing: Global Brain Corporation, Anthemis, Wayra

In a nutshell: Humanising Autonomy is setting the global standard for how Automated Vehicles interact with people. The company has developed an intent prediction technology that predicts the full range of pedestrians and the behaviour of vulnerable road users across different environments and cities. As a critical perception technology, the software integrates with driver assistance systems, autonomous vehicle stacks, and smart infrastructure systems for real-time accident and near miss prevention, improving the safety and efficiency of Urban Mobility Systems across cities worldwide. 

Humanising Autonomy’s solution: Humanising Autonomy has developed an intent prediction platform that is able to recognise and predict pedestrian and vulnerable road user behaviour in real-time. The technology integrates with ADAS systems, Infrastructure systems and Autonomous Vehicle stacks for real-time accident and near miss prevention, as well as cloud based video analytics.

USP: Identifying dangerous situations more than 2 seconds earlier than a human driver. Proprietary predictive features, such as ‘risk index’, ‘intention to cross’, and ‘acute danger’ with respect to pedestrians, cyclists, and other vulnerable road users. Real time analysis and prediction of events around the vehicle. Customisable features that are adaptable to specific application requirements, environments, and cultures.

Goals for the next 12 months

• Revenue conversion to large recurring licensing contracts
• Integration with commercial fleets today

Experience: A combination of the best engineers and scientists in AI, computer vision, and behavioural psychology from Imperial College London, Cambridge University, LSE and other academic institutions. 

Traction: Customers are leading global mobility providers including OEMs, Tier 1s, Fleet Operators, and Transportation Authorities. Humanising Autonomy has deployments across Europe, the US, and Japan, with technology tested and validated with multiple customers including Daimler Mercedes Benz and Airbus.

Website | Twitter

Founder: Hami Bahraynian, CEO

Location: London

Year founded: 2016

Core area of focus: Mobility & Autonomous Vehicles

Previous backing: IQ Capital, Seedcamp

In a nutshell: Wluper is building Conversational AI at the intersection of Natural Language Processing, Dialogue, and Knowledge Bases, aiming to bring them together by leveraging deep learning. Existing voice-based assistants can only understand one-way commands, not conversations, so developing goal-driven dialogue systems is required to achieve significantly improved user experiences.

Wluper’s solution: Wluper is building a Conversational AI technology that understands everything transport-related. By narrowing down the problem and focusing on one domain, they can make clear assumptions about what the user is talking about, which allows them to understand more complex and naturally asked questions, multi-intent queries where the user asks for different things in one go, and, ultimately, follow up questions enabling a true conversation. They believe it’s exactly the conversational nature of a back and forth which will make peoples’ relationship to new technologies more personal and more intimate.

USP: The main differences and competitive advantages lie in Wluper’s NLU 2.0, i.e. different Semantic Representation, allowing quick annotation and learning as well as the necessary learning algorithms. That, plus endeavours in knowledge bases as well as the overall system design, allow Wluper to perform better from a tech stack perspective and build and introduce an improved user experience.

Goals for the next 12 months

• Hiring good talent
• Transition into an applied tech startup, rather than one that is purely research
• Taking our results, insights, and technology and applying it into real life

Experience: Wluper has a skilled team of researchers, engineers and entrepreneurs with backgrounds in Mathematics, Machine Learning, Software Engineering, and Computational Linguistics. Experience spans from an array of organisations, including Imperial College, UCL, University of Cambridge, Oxford, and Edinburgh, Apple's Siri team, Amazon Alexa, Nuance, Volkswagen, among others.

Traction: Although Wluper is still in semi-stealth mode and hasn't done any marketing and sales, they have still been approached by car-manufacturers, Tier-1 suppliers, and tech suppliers, all expressing interest in their approach and technology. They are currently in conversations with Bosch, BMW, JLR, TomTom, to name a few, and aim to identify proof of concept, where concrete needs and technical requirements exist.

Website | Twitter

Founder: Bola Adegbulu, CEO

Location: Exeter

Year founded: 2016 

Core area of focus: Mobility & Autonomous Vehicles

Previous backing: EF, AI Seed Fund

In a nutshell: Predina uses AI to build a dynamic risk map platform (STARI) that predicts and prevents the risk of road crashes based on contextual (referred to as spatio-temporal data) and previous accidents. STARI makes road users aware of contextual risk and provides the safest route from A to B in real time. They license their core technology to automotive mapping partners for ADAS and AV use cases and insurance telematics.

Predina’s solution: Predina uses data from over 2 million accidents and 28 contextual variables, like weather, visibility and traffic flow, to dynamically predict where and when accidents are likely to happen. Their API solutions also includes the journey risk assessment for a given route and maps out the safest route for a journey. For mobility providers, their B2B2C app further reduces the risk of accidents to their users by providing journey risk reports, audio alerts and the safest route from A to B.

USP: Instead of using just weather information or static information about the road layout, Predina’s solution combines data from previous accidents and 28 spatio-temporal variables including weather, road layouts as well as other variables. This provides a more accurate, predictive and dynamic model able to predict road accidents in real time and up to 48 hours ahead of time.

Goals for the next 12 months

• Raising their next round of financing (or grants)
• Converting their pilots to recurring revenue
• Monetisation and user growth

Experience: Predina’s team have a rare combination of deep technical expertise specific to spatio-temporal problems, automotive data domain expertise and experience building beautiful, user centric B2B2C products, all of which are essential to the success of the company.

Traction: Predina’s solution achieved between 70% - 83% in predicting the risk of over 1000 real world accidents. In a real world trial of their solution, it reduced accidents by between 25% - 53%. They successfully completed a proof of concept with a commercial fleet (Telematics) who invested in the company. Their pipeline includes leading mobility providers, telematics providers and ADAS/AV mapping providers.

Website | Twitter

Founder: Mads Holmen, CEO

Location: London

Year founded: 2014

Core area of focus: Marketing & Advertising

Previous backing: 01 Ventures, Informa and Howzat

In a nutshell: Bibblio helps content owners navigate the attention economy by matching the right content to their audience through AI; increasing engagement and maximising the return from content. Bibblio's AI engine runs across three tools: Circulation, Syndication and Promotion. Each targets important metrics for content owner teams.

Bibblio’s solution: Content recommendation and personalization

USP: The intelligent decision layer between content and users

Goals for the next 12 months

• $2m+ ARR
• Tier 1 conversion (+20 Tier 1 annual licenses)
• Successful US market entry

Experience: The team has experience in media, tech and data science, helping publishers adapt to a world beyond 3rd party cookies and programmatic ads.

Traction

• Live on 1300 sites
• 100m monthly users served
• Significant traction with a range of well-known media brands (Informa, Ziff Davis, Haymarket etc)
• Investment from Informa, 01 Ventures and Howzat

Website | Twitter

Founder: Dhruv Gulati, CEO

Location: London

Year founded: 2017

Core area of focus: Media, Entertainment and DemocraTech

Previous backing: Mark Cuban

In a nutshell: Factmata builds natural language processing to understand online content for what is being said, and how. Their goal is to develop a better understanding of online information and rebuild trust in the internet.

Factmata’s solution: Today, organisations find it incredibly difficult to understand what is being said about them online, or trending ideas that might give them an edge in developing new products. Factmata helps organisations detect claims spreading online about any issue or topics they would like to monitor. Factmata Intelligence Reports not only extracts the claims, but it’s threat level, influence level, who is spreading it, and where the claims were discovered originally.

USP: Being able to deliver unique insight, rather than just media tracking to simply "keep on top of things" is a unique and new point of view that Factmata embraces. By synthesising online content into claims, arguments and rumours, customers can derive real information that can change how they message online, decide what products to build, what campaigns to run and more. Most existing tools simply detect brand mentions and sentiment, but do not synthesise and summarise the key takeaways from the day or week, like Factmata’s tool does. 

Goals for the next 12 months

• Close SaaS deals, which have typically long sales cycles, before the end of 2019. 
• Hit recurring revenues at level to justify a Series A. 
• Pick which organisations to buy expert media intelligence (brands, PR agencies, defence, government policymakers, consultancies, news companies).
• Keep maintaining their main goals - continuing to acquire regular training data for our algorithms cheaply, and run two revenue streams (moderation and expert media intelligence).

Experience: Factmata’s team experience consists of building large data annotation tools along with world-class machine learning engineers from places like Creative AI, DigitalGenius and Nuance with experience in complex NLP applications, along with data engineers with experience building large scalable platforms around email marketing and more. Their advisors lead Facebook AI research and are professors at Cambridge NLP lab, and we are well connected to the state of the art research happening in fields like automated fact-checking, hate speech and argument mining.

Traction: Factmata currently has eight algorithms for moderation built out, with all above 90% accuracy and unique test datasets created for each. They also have developed 20,000 rows of unique manually annotated training data obtained from expert journalists, in house specialists and advocacy groups. They have a pilot starting with the BBC and Mediacorp Singapore, and $8k MRR deal signing with Taboola.

Website | Twitter

Founder: Ali Tehrani, CEO

Location: London

Year founded: 2018

Core area of focus: Media, Entertainment and DemocraTech

Previous backing: AI Seed Fund, Luminous Ventures

In a nutshell: Astroscreen uses AI to protect brands and defend democracy from harmful social media manipulation campaigns, also known as astroturfing. Social networks have become the new cybersecurity attack vector and Astroscreen is building the solution.

Astroscreen’s solution: Astroscreen has a series of detection methods that can spot anomalies in social networks, generating alerts. By detecting signs of social media manipulation at its earliest instance, they can give brands and governments time and context to mitigate the adverse effects of social media manipulation.

USP: Their main competitive advantage is the deeptech DNA of the startup, having been founded to commercialise PhD research, and having doubled down on this since. Their second USP is the approach to solving this problem; namely, a signals approach, whereas competitors are focused on building one or two silver-bullet algorithms, where there are none.

Goals for the next 12 months

• Hiring
• GTM
• Fundraising

Experience: Astroscreen has a team with varied skills, consisting of advisors, founders and employees that have tackled this problem previously. If you look at the most widely cited papers in this space, they are either written by a team member or an advisor of Astroscreen. From a commercial point of view, their founders have previously built a venture-backed startup in a similar space.

Traction: Having focussed primarily on R&D to date, the Astrosceen team are now working on a few pilot projects, including one designed to monitor the election of a G7 nation.

Website | Twitter

Founder: Mark White, CTO

Location: London

Year founded: 2016 

Core area of focus: Marketing & Advertising

Previous backing: Force over Mass, Wayra

In a nutshell: Codec helps brands identify the audience networks that matter most to them, while giving rich insights into how they can win their engagement and grow. This is performed by real-time analysis of hundreds of millions of real content interactions.

Codec’s solution: Brands use Codec’s intelligence platform to discover, understand, and connect with audience networks that represent high-opportunities for growth in a GDPR friendly way.

USP: Codec’s proprietary contextual data is generated from the analysis of millions of online content interactions. Their AI and Machine Learning algorithms find meaningful communities of people connected by shared passions, values, and behaviours through the content they choose to engage with online. Insight into these groups empower brands to connect more meaningfully and effectively with consumers.

Goals for the next 12 months

• Double MRR 
• Drive up margins to 75%
• Continue to scale through the world-class brand groups they already count as customers

Experience: Codec’s leadership team has tremendous experience in building and scaling businesses. Their proprietary technology is built by a team of expert marketers, data scientists, cultural strategists, and investors.

Traction: Some of the biggest brands such as Unilever, Procter & Gamble, L'Oreal, Samsung, Nespresso, UBS, Walgreens, Boots and many more already trust and rely on Codec's technology to identify and connect with the right audiences. 

Website | Twitter

Founder: Iggy Bassi, CEO

Location: London

Year founded: 2016

Core area of focus: AgTech

Previous backing: Future Positive Capital and Astanor Ventures

In a nutshell: Cervest helps businesses, governments and growers adapt to climate volatility. They use machine learning to generate real-time streamed ‘climate signals’ to answer questions linked to climate uncertainty, land, and natural resources. It informs decisions that help mitigate risk, safeguard food and economic security, and protect our planet.

Cervest’s solution: Cervest is pioneering Earth Science AI to empower businesses, governments and land managers to adapt to climatic and extreme events. Using its research-led machine learning, Cervest streams ‘signals’ that predict and quantify climate impact.

USP: The power of their detection platform. Beyond their potential to quantify solutions to Earth’s most pressing and complex problems, Cervest has the ability to transform the day-to-day and strategic decision-making of managers operating in risk, insurance, investment, infrastructure, CPG, national security and public policy, as well as farmer co-operatives.

Goals for the next 12 months

• Launch beta platform in Q1/2020 for real-time climatic extreme events with an initial EU28 coverage.
• Test scientific capability for multiple use cases and users, to ensure the development of the most valuable product/service within key sectors.
• Develop category leadership within the nascent multi-billion climate services market.

Experience: The Cervest team combines outstanding talent from academia and enterprises, including University of Cambridge, Imperial College, Harvard University, Cornell University, The Alan Turing Institute, Cornell, Google, and Microsoft.

Traction: Cervest has just raised £3.7m in a round led by Future Positive Capital and Astanor Ventures, oversubscribing on its original target of £2.5m. It has tested its early-stage product with global firms including McVitie's and Syngenta. It has also worked closely with the Colombian Government.

Website | Twitter

Founder: Camille Rougie, CEO

Location: London

Year founded: 2017

Core area of focus: Fintech 

Previous backing: SpeedInvest, EF, AI Seed Fund

In a nutshell: Plural AI is building a computational engine or data science platform for finance. Their systems mine web data to help business users answer complex, business-relevant questions. They are working towards becoming the default interface for knowledge work, allowing business users to access complex data science without knowing how to code.

Plural AI’s solution: Plural AI is a new kind of search engine, specifically for financial analysis. Current search engines are great at retrieving existing information, but only for generic questions which have been answered before and present issues like outdated information, unreliable sources and require manual work to produce the data. Plural AI’s approach is much more computational in nature and they aim to understand users’ questions and produce bespoke, on-demand answers to them by assembling data points from various parts of the web. 

USP: Plural AI’s platform is extracting information in a fully automated way, without any human input (other than to create data sets to train the models), whereas the majority of competitors rely on the work of human analysts. Their approach is fundamentally more scalable, as they can deploy systems across the web to mine an ever-growing range of concepts. Furthermore, they also build predictive models, based on their own data sets. 

Goals for the next 12 months

• Monetisation of early customer base (going from trials to PoC, negotiating agreements and filling the top of the pipeline). 
• Managing product delivery/focus on how to build the right features (e.g. managing pace and customer expectations, custdev cycle). 
• Hiring (data science/engineers in particular) and scaling the tech team

Experience: Plural AI’s founding team met through the deep-tech Entrepreneur First incubator program. Their Oxford-educated CEO, Camille, has 8 years’ experience in finance (IBD, PE, and fintech start-ups), whilst CTO Jaron, holds a PhD in Artificial Intelligence from Queen’s University Belfast and has over 20 years’ experience in graph theory, NLP, and knowledge-based systems.

Traction: So far, Plural AI has one paying client that has been using their product for 12 months, and just renewed their contract, in addition to a mid-market private fund trialling it, and a paid PoC with Big 4 firm. Product-wise, they built a knowledge graph containing ~15m nodes, which processed 500m+ web pages or other documents relating to them. Their models were built in-house and outperformed published benchmarks on text mining and prediction tasks. 

Website | Twitter

Founder: Nicolai Baldin, CEO

Location: London

Year founded: 2017

Core area of focus: IT systems 

Previous backing: Seedcamp

In a nutshell: Synthesized accelerates data-driven projects with high-quality synthetic datasets that mimic the original data, thereby unlocking deep insights whilst keeping original data locked down. Synthesized datasets allow partners to carry out practical analytics and research faster and more efficiently without ever needing to deal with sensitive data.

Synthesized’s solution: Synthesized help companies unleash the potential of their data by addressing the technical or compliance based issues associated with data sharing. They do this by providing them with high-quality synthetic data that mimics the structure, but not the individual data points. This enables most practical analytics and research to be done without ever needing to touch sensitive data. Synthesized has already developed and tested the core technology and is looking to accelerate product distribution in the UK.

USP: Unlike anonymization, encryption and other data manipulation techniques, Synthesized’s approach enables partners to share ultra-high-resolution information in a prepared format tailored to machine learning and analytics, without disclosing any information about individual data points.

Goals for the next 12 months

• Solve marketing and PR challenges to get the right messages across to decision makers in their segment.
• Qualifying customers.
• Closing opportunities.

Experience: Their leadership team combines over thirty years of first-hand experience working on machine learning products in leading enterprises and research institutions. The founding team members invested two years building the core product for data synthesis upfront by developing proprietary algorithms and novel methodologies.

Traction: Since its product release, Synthesized has successfully completed two pilots with a Tier 1 insurance company and Tier 1 retail bank, and one pilot with a seed-stage company. They are currently in contractual discussions with those companies and are also initiating another pilot with an international technology conglomerate. Furthermore, they agreed two partnership agreements with a leading software technology company and an international consultancy company. 

Website | Twitter

Founder: Mohammad Rashid Khan, CEO

Location: London

Year founded: 2016

Core area of focus: Computer Vision

Previous backing: LocalGlobe, EF, Global Founders Capital

In a nutshell: Calipsa is on a mission to make the world a safer place. By building technology that transforms the detection and prevention of crime, Calipsa is empowering security professionals worldwide to make better decisions, in real-time. It uses cloud-based AI technology to identify the cause of a CCTV alarm, filtering out over 85% of false alarms and forwarding only true alarms to human operators to review.

Calipsa’s solution: Calipsa has developed an AI platform that analyses CCTV data to assess risk and automatically notify a human operator enabling them to take preventative action. On average, the platform is able to autonomously handle 80% of the workload of a CCTV control room, allowing for business growth without having to scale the workforce.

USP: Calipsa have developed ML algorithms that do not require the analysis of a video feed, instead decisions can be made with just 3 frames. This eliminates the need for a local server to process the entire video which can be expensive and removes the need to send an engineer to the site to configure.

Goals for the next 12 months

• Hiring A+ people.
• Retaining culture as they scale.
• Educating and influencing the market.

Experience: Both founders have advanced ML degrees from Tier 1 universities, combined with previous industry experience. Calipsa have assembled an amazing team of engineers, scientists and commercial people.

Traction

• 22 B2B customers (UK, USA, Ireland, Denmark)
• £46k MRR, up from £12k in Dec 2018
• 30k+ cameras monitored

Website | Twitter

Founder: Mikela Druckman

Location: London

Year founded: 2018

Core area of focus: Computer Vision

Previous backing: None

In a nutshell: Greyparrot is improving and automating the recycling process by enabling waste composition analysis and sorting of waste types. They use deep learning to power next-generation robotics and smart systems for waste management.

Greyparrot’s solution: Greyparrot’s recycling recognition API allows identification of different types of waste. They combine recent advances in deep learning to solve complex visual tasks coupled with the decline in cost of robotics. This allows further automation that is economically viable, making waste sorting more intelligent. 

USP: Companies that use traditional computer vision techniques with no AI currently are not able to solve quality control tasks, which are still manual. Greyparrot’s deep learning based solution could solve problems like separating plastics bottles from clear plastics, as it recognises plastic, but also the type of plastic (PET, HDPE and so on) as well as detecting contamination. With the imminent plastic crisis we’re facing today, recycling this material is more important than ever. 

Goals for the next 12 months

• Scalable data collection
• Choosing the exact use case to focus on
• Finding the right investor for the next phase of growth

Experience: Greyparrot was founded by the former computer vision team at Blippar and has a track record of building world-leading recognition systems. They also have extensive experience in working with large clients and scaling emerging technology solutions like AR/AI. The team have complementary skill-sets combining both commercial and technical expertise and bring a strong network of universities, corporations and influential public bodies. 

Traction: Greyparrot launched its first waste API prototype, based on self-acquired and labelled data. They developed a deep learning model capable of recognising cans, plastics (PET, clear bottles, milk bottles and black plastic), cardboard, card, newspaper and magazines with more categories being added. They’ve also integrated their API into robotics in partnership with Middlesex University for the following demo and established pilot partnership with Egbert Taylor in UK and ACI chemicals in South Korea. 

Website | Twitter

Founder: Rafie Faruq, CEO

Location: London 

Year founded: 2017

Core area of focus: LegalTech

Previous backing: EF, Connect Ventures

In a nutshell: Genie AI uses machine learning to curate and share relevant legal knowledge within a firm or legal team, empowering lawyers to draft with the collective intelligence of the firm. Their intelligent contract editor, "SuperDrafter," analyses thousands of variations of the same clause, deducing market standards so clients can negotiate the best deal every time.

Genie AI’s solution: Genie AI helps lawyers draft, negotiate and review with intelligent insights from their past data. It uses machine learning to recommend clauses, in real time, whilst lawyers draft. They can search for any clause and filter by market, client, sector, and other smart searches. Using their AI or their intelligent contract editor, SuperDrafter, Genie AI enables lawyers to always get the best clause.

USP: Genie AI are patent-pending. Their artificial intelligence is the only AI in legal tech that does not require manual tagging. It is ready to use from day one. They are also secure. To protect client data, they are building an AI-anonymizer that automatically redacts sensitive information from any data on the platform.

Goals for the next 12 months

• Scaling the company in line with demand
• Doubling the size of our team in line with the scaling timelines
• Launching a comprehensive sales and marketing strategy

Experience: Genie AI’s team of 12 people includes postgraduate engineers who are highly experienced in enterprise software development and machine learning. Their Head of Product is an experienced lawyer the magic circle. Co-founders were taught by Google Deepmind on their MSc in Machine Learning degrees and achieved top 98% and 92% in recent Kaggle competitions (among over 2000 data scientists). Advisory board includes world-leading professors in machine learning, and Lord Neuberger, the former President of the Supreme Court.

Traction: Genie AI has raised a £1.2 million seed round and has won an £800K grant from InnovateUK since it was founded. Their intelligent contract editor, “SuperDrafter,” is currently piloting with three large law firms, with interest from over a dozen others. They're also backed by leading angel investors and advisors, including former President of the Supreme Court Lord Neuberger and Professor Jun Wang at UCL.

Website | Twitter

Tech Nation empowers tech entrepreneurs to grow their businesses through growth programmes, digital entrepreneurship skills, a visa scheme for exceptional talent and by championing the UK’s digital sector through data, stories and media campaigns. 

Our first AI initiative is the growth programme Applied AI, launched in June 2019. We received almost 140 applications, and after an intense judging round from an independent expert panel, we on-boarded 29 AI companies with the potential and ambition to scale. 

In addition to delivering ground-breaking and innovative products, the Applied AI 1.0 cohort focus on ethics and improving the world around them. Their products are focussed on solving practical issues that real people or businesses face every day, aligning their visions, values and goals with our own.

Our Vision

To make the UK the best place to imagine, start and grow a digital business.

Our Mission

To empower ambitious tech entrepreneurs to grow faster through knowledge and connections; to build a UK economy fit for the next generation.

Our Values

• Originality: we break new ground
• Meaningful Collaboration: we nurture relationships
• Real Impact: we make it count

To know more or set up a meeting with any of the AI companies, please contact Harry Rhys Davies, Applied AI Lead.