Why AI is one of my favourite, but also most challenging sectors to invest in

Learnings from one year of sourcing AI startups at Founders Factory

Andrew Ng is raising a $150M fund solely focused on Artificial Intelligence. Salesforce set aside a whopping $50M to invest in AI startups, as did Toyota announcing the launch of its $100M fund. CVC alone has seen a $1.8Bn invested in AI companies during the first half of 2017.

In short, AI is hot and every investor is placing their bets.

At Founders Factory we dedicated an entire sector to working with early stage deep tech and AI companies, together with our investor CSC Group, one of China’s largest PE funds and most active tech investors.

Within my role on our sourcing team, I get the opportunity to talk to entrepreneurs from a wide variety of businesses and fields from cancer diagnosis to protein research in the cloud; complex data visualisation in 3D; satellite image processing or autonomous marketing assistants using natural language processing. Despite this incredible variety, we see that these companies all face very similar challenges, specific to deep tech early stage businesses.

  1. Lack of access to proprietary datasets

The holy grail to commercial impact and competitive advantage lies in the access to proprietary datasets, due to their specific nature and barrier to imitation. While the creation of large enough datasets from scratch is incredibly time intensive and costly, vast amounts of data can be found hidden and protected within large corporations, who serve millions of customers every single day.

As example, easyJet transports over 70M passengers each year and has a record 20M downloads for its app. Aviva is one of the world’s largest insurance providers, servicing over 33M customers globally of which 7.5M are digitally active, and CSC Group has access to a network of over 100 Chinese universities.

At Founders Factory we can activate this potential, helping founders get access to datasets through pilots and commercial projects with our large network of corporates. Incumbents naturally safeguard their data and assets, and it is especially hard for small companies to gain enough trust and reputation for joint projects. We are uniquely positioned to bridge this gap to otherwise unattainable datasets.

2. Complicated university spin-out processes

Research institutions and universities provide an ideal breeding ground for talent and advanced technology, and valuations are literally increasing per PhD or data scientist on the team. As an obvious consequence, starting a business instead of continuing a career in research has become an attractive alternative.

Over the last year, I had the chance to meet with very impressive individuals and got insight into PhD projects with the potential to have a true impact. Developing these projects into scalable and investable businesses, however, frequently proves to be a more challenging and lengthy process than anticipated by founders and investors alike:

  • Many universities do still lack standardised, simple and founder-friendly procedures around spin-outs and especially IP rights. This can result in long and expensive negotiation processes or leave the founding team with unfavourable conditions for raising investment and hence future growth.
  • Especially in biotech and healthcare, clinical trials are costly and validation takes a long time, a fact that especially early investors and founders have to anticipate and embrace.
  • Once a company is fully formed and the technology is ready for first commercial pilots, it often becomes challenging to remain strict around building a scalable solution and staying away from the customisation trap.

Our Chief Data Scientist Jeffrey has gained a lot of experience from spinning out his own company Cortexica and shares some of his learnings here as well.

3. Developing a scalable product and model

A common challenge we observe among AI businesses is a rocky trajectory from initial pilots to a scalable model or technology platform. Especially companies with complex algorithms at their core find themselves trapped in consultative business models with heavy upfront integration work. Similarly, many of these businesses initially start deploying their technology on pilots in various industries without clear focus, leading to a need for customisation for each individual client.

This is partly due to the nature of the product itself as well as the clients’ needs, while in some cases also a function of missing commercial experience on the team.

These challenges are likewise concerning for founders and investors in these businesses, and we especially probe founders’ ability and thinking around them during our interviews.

Join us at Founders Factory

With our large operations team, however, we do not only have the resources but are also excited to tackle these challenges together with promising early stage teams and support them in their journey. We have developed our own methodology of working closely with deep tech companies, led by our Chief Data Scientist Jeffrey Ng and drawing from his experience in building and scaling businesses like Cortexica and BenevolentAI. Over the course of six months, the goal is to sharpen the value proposition, find product market fit, incorporate regular feedback from actual clients, work out a scalable model and — as a consequence — raise funding.

With the learnings from our first AI cohort behind us, we are currently looking to select the next exciting companies to work with. Here is a short summary about what we are looking for and how companies can join the Factory. Or feel free to reach out directly if you would like to find out more about how we work with companies in our AI sector.

About Founders Factory

We’ve established a whole new model for business acceleration and incubation, focused on day-to-day support from a dedicated 60+ operations team and a completely bespoke approach rather than relying on intermittent mentors and rigid program structures, and provide unprecedented support from our corporate investors — Holtzbrinck (EdTech & Science), Guardian Media Group (Media & AdTech), L’Oreal (Beauty), Aviva (FinTech & Insurance), CSC (AI) and easyJet (Travel).