In a recent report, London-based venture capital firm MMC was unable to find any evidence of AI use in close to half of the “AI start-ups” in Europe. Among start-ups that did use AI, the majority preferred purchasing third-party AI solutions rather than building their own.

The report, sponsored by Barclays, has sparked debate over whether AI has simply become a fashionable term to attract publicity and funding, and also begs the question of whether these ‘AI’ companies are filling the pockets of big software firms at the expense of developing their own IP.

“AI” attracts up to 50% more funding

MMC’s review looked at the “activities, focus, and funding” of 2,830 purported AI start-ups in 13 EU countries. In only 60% of these companies was there evidence of AI having a contribution to the firm’s business processes.

“We looked at every company, their materials, their product, the website and product documents,” David Kelnar, head of research for MMC, told Forbes, “In 40% of cases we could find no mention of evidence of AI.” In such cases, he added, “companies that people assume and think are AI companies are probably not.”

The report showed that AI-branded start-ups have attracted between 15 and 50 percent more funding than traditional software companies since 2015. This may explain why start-ups prefer branding themselves in this way. However, it’s worth noting that not all of the companies were actively claiming to use AI; some were simply failing to correct misclassifications made by third-party sites.

Bespoke solutions and differentiation

The study reported that nearly half of the companies surveyed favoured buying AI solutions from third parties, while a third intended to build custom solutions. The remainder looked to outsource custom solution development to a third party or wait for AI to be embedded in their favoured software products.

Seeing as only a third of the surveyed companies chose to develop AI solutions in-house, it’s important to highlight the reasons why this may be.

Key challenges for start-ups looking to develop AI solutions as identified by MMC:

  1. Recruiting AI talent
  2. Accessing good quality training data
  3. Developing production-ready AI that would work in the real world

Despite custom-built AI systems (e.g. Google’s Deepmind) yielding potentially greater benefits and attracting more investment, some start-ups may view the ambition of producing bespoke solutions from scratch as a costly and unrealistic approach. One alternative to this is for start-ups to collect valuable data and use open-source machine learning libraries to achieve a similar goal. Although this approach may not result in the development of IP which is as valuable as that produced by pursuing an entirely bespoke solution, it carries a lower risk. If embarking on this approach, companies should be conscious of the obligations imposed by many of the common open-source libraries, concerning software reuse and contribution back to the open source community.

Spending on off-the-shelf AI products and spending on developing custom solutions in-house should not be seen as mutually exclusive. Investing in both could allow firms to create IP that enhances business value and is attractive to investors. Even if the concepts and applications developed are not capable of being patented, there are many other forms of IP protection available which can allow companies to protect their valuable efforts from misuse by third parties and demonstrate value to investors.