Nurturing Adoption: Supporting Growth of Early-Stage AI Businesses in Scotland

This video shares a talk from the Scottish AI Summit, addressing how best we can support the commercialization of AI technologies to meet real-world needs.

The University of Edinburgh has programmes to nurture data-driven entrepreneurs at different stages on their journey from the seed of an idea to scaling globally. As well as supporting both technology transfer from academic research and home grown businesses, the University is also attracting high growth global AI start-ups.

In this panel session, Katy Guthrie and Laura Bernal from the University’s Data-Driven Entrepreneurship programme will be joined by founders of some the companies they have supported, across different sectors including MedTech and ClimateTech.

The discussion will focus on the opportunities and challenges for companies using AI, how scaling AI driven business can benefit the people of Scotland and the world and what can be done to stimulate activity and address issues.

Introductions – (3m:15s)

Katy Guthrie, the Manager for the University of Edinburgh AI accelerator program commenced the summit by stating the purpose of the panel, which is to explore how best to support the commercialization of AI technologies.

The university has entrepreneurs at different stages on their journey from the seed of an idea to scaling globally, as well as supporting both technology transfer from academic research and home-grown business. The University is also attracting high-growth global AI start-ups.

She introduced the other members of the panel who are founders of AI-driven companies at different stages. They’ve all taken entrepreneurship programs at the university too.

After the introduction, at about 4m:20s Katy listed the topics to be discussed which are:

• The role of AI as an enabler.
• The areas of opportunity.
• Challenges faced and how well to overcome them.

She describes how they had a recent program that supported 12 AI-driven companies over a period of 6 months and it’s helped companies to build commercial skills, connect them into networks, and to learn from each other.

At 6m:47s Laura Bernal talks about how they supported Ph.D. students and early career researchers who are looking to commercialize IP or research, and that their program helps companies to be more business-minded. They help scientific background people to think about business because they believe academia can have a relationship with industries. She was formerly a founder before she joined the program.

At 8m:14s they are joined by the other panel members introducing themselves and what they do. Xiaoyan Ma, founder and CEO of Danu Robotics introduced herself and talked more about her company.

Then at 9m:00s, she hands over to Lorenzo Conti, a founder and managing director at Crover who talked about his Edinburgh-based startup which focuses on helping grain storage operators like grain merchants, port operators, and cooperatives.

On to Debbie Wake (10m:28s), a medical doctor and clinical academic with the University of Edinburgh. She is also a Co-founder of a digital health company MyWay. She gave a brief talk on how they manage conditions like diabetes, drug response and diagnostic prediction as well as how they offer tools to Education platforms.

Joseph Twigg, CEO of Aveni, introduces his company (11m:56s), also an Edinburgh-based startup, where they focus on natural language processing, delivering and helping with solutions in the service industries. They derive insight and automation from customer conversation.

Scaling AI in Scotland

From 13m:40s Katy moves on to the detail of the discussion, with the main perspective being to identify why it is important to support the scaling of these businesses, which is to boost economic growth and productivity because these companies are creating high-value jobs.

They also support companies that tackle societal issues in Scotland. They have a strong capability in AI in Scotland which is attractive to companies who want to work not just at the University of Edinburgh but in Scotland.

They are often seen as being great at inventing things but not as good as commercializing them. For example from a review it was identified that the UK is ranked third in the world for research in AI but only 11th in its ability to realize Impact from it. So they want to change that and make Scotland seen as an excellent place to build a high-value AI-driven business.

It’s partly about supporting homegrown businesses and also attracting AI start-ups from elsewhere. The key point is about building the ecosystem.

In Scotland, they’ve got lots of big corporates in industries and they need innovative smaller businesses like start-ups that can bring new ideas. Developing and fostering that ecosystem is a key part of what they do at the Bayes Centre of the University of Edinburgh with the data-driven entrepreneurship support program at the university.

The Role of AI as the Enabler (16m:23s)

Here Katy asks the panel why they use AI, to explore what role the technology plays as an enabler.

For example at 17m:04s Joseph said he worked in the financial services industry for 15 years and identified a number of potential scenarios for improved efficiencies through using the technology.

Debbie highlights how they didn’t start as an AI business, but rather as an educational web site for diabetes information, and they came to realize how important data was to that process, and it became clear that AI could play a key role in facilitating better personalized access to that data.

For Lorenzo (20m:00s) they were working with a very novel and complex technology that involves hardware locomotion through a very kind of unstable environment data and they employed AI to tackle several technical challenges in a different part of their product, and to reach the required level of performance. In a very efficient way, AI helped them in the prediction of the robot’s movement around the environment, and since then they have employed it in other areas like the accuracy of their sensors.

At 23m:41s Katy asked Laura if she has any other companies looking up to AI for solutions to early-stage academic problems.

In response to this she mentioned a company that didn’t start as an AI company but later employed AI to model how the results of diagnosis will perform and how effectively it could put them to an advantage over their competitors. She said she has seen many companies that are AI-driven and need AI to increase their competitive advantage.

Finding Opportunities

Laura takes over at 25m:11s and asks the panel if there are other areas where AI can be applied.

For Xiaoyan it will have a huge opportunity in traditional industries such as agriculture risk management and manufacturing. AI can help people in the space to improve their operational efficiency, profitability, and sustainability, and can help in changing the procurement processes.

For Lorenzo (29m:35s), what is needed for the AI solution to work is assistance on how to manage funding.

There’s a need for access to another company’s customers’ data and regulatory bar set very high to train models (Joseph 32m:58s), and AI also needs to make a product that compensates for the risk companies are taking.

If you could do something different this time what it be? (Laura 34m:14s)

According to Joseph, partner selection would be the first thing to do differently, especially in the first critical phase of developing and validating the concept.

Debbie seconds this at 35m:25s, describing their challenge with Information Governance in the NHS, and how bringing in an expert consultancy to address this early on for them would have saved them up to two years of work.

At 38m:00s Laura asks Katy what opportunities she has seen and the ones coming up next. Katy replies that she sees many, in key areas like Natural Language Processing, and that you can view them through a lens of industry vertical or horizontal technology specialisms, and that there is potential to cut across multiple use cases.

She cites an example of a body scanning app that took part in their accelerator, which can be used for e-commerce scenarios where it can be used to match customers sizes to reduce returns, and it can also be used in Healthcare.

Challenges Faced by AI-Driven Business

At 43m:45s the panel moves on to the final topic: Challenges Faced by AI-Driven Business.

These include the volume of data (Debbie 44:23), and when dealing with AI there’s high-risk complexity often that brings in regulatory burden, therefore there’s a need for additional research evaluation risk assessment.

Joseph makes a critical point at 45m:42s, that the Scottish startup support ecosystem is great, but it’s not so great when it comes to encouraging large enterprise organizations to work with those startups. For example providing them incentives to choose to work with a local startup vs AWS.

From 49m:18s, Debbie talked about what can help:

• Support within the Scottish ecosystem.
• Support from using wider support.
• Support from Scottish Enterprise. They support in a lot of ways around market scoping internationally and have amazing networks and obviously the AI accelerator program.

From 51m:58s Lorenzo adds that there’s a need for support programs that truly understand deep hardware innovators, a challenge globally not just in Scotland.

At 54m:45s Katy asks Joseph: Should early-stage businesses worry about the public perception of AI, or should they just get on with making the best possible solutions?

Lorenzo believes it’s a lot to do with perception, that most people assume AI is intelligent robots when really what is in use is the type of business applications improved by algorithms that have been discussed on this panel. Similarly Joseph adds that while AI is used to augmented processes like those in financial services it is still ultimately a decision made by a human.

Debbie also added that though changing public perception is really hard but as SMEs she thinks they’ve got restricted resources as to how to do that. Working with policymakers and bodies to try and support the industry image is a good thing.

Wrap Up

At 57m:32s: Katy handed over to Laura to round off the summit.

Laura started by accepting the fact that there are always going to be challenges in the process. Therefore she encouraged us to think about how we could help companies that are developing in AI and how we can start changing that infrastructure to overcome those challenges as an ecosystem in general.

In her closing remark, she said the AI accelerator would be opening applications for their next intake in April to start the program in September. And also if anyone wants to scale an AI company should contact Katy through her LinkedIn: Katie Guthrie.