If I were to summarise the deals I’m seeing, there are three key trends.
1. Tools for managing models, understanding model bias, archiving models, etc.
2. Tools for making AI easier to do for roles of all kinds. This means apps that make AI easy for marketing and sales, platforms that make AI easy for data scientists, tools that make AI easy for people designing and building robots, etc.
3. Marginal improvment applications – AI tools that are a little better than what is on the market, fixing design flaws in early AI apps or peeling off market segments from broadly horizontal applications.
What is noticably absent are two things. First, top-to-bottom transformational AI applications. I think the reason for this is the cultural adoption challenges of really changing workflows to accommodate these apps and get all the benefits, and the economic challenges AI applications can have, initially, versus traditional software.
The second thing that is absent is really revolutional technology breakthroughs. There are people working on some, for sure, but I haven’t seen much that really matters. It’s not like AI was 2015 – 2018 when all kinds of companies came out doing things you couldn’t even believe were possible.
So we are in a lull. It’s not an investment lull because many of these companies started today in the fields above will be very valuable. But it’s an innovation lull, or at least it feels like one. And I wouldn’t call it an AI winter because it isn’t driven by a lack of technology. It’s driven by the need of the market, and the operational infrastructure of work more generally, to catch up to what is offered.
That also means it’s a great time as an entrepreneur to start thinking about what is next. In the five years, we’ve gone from buyers and companies misunderstanding AI to embracing it, and now it’s really beginning to penetrate workflows. It’s early, but the changes in attitude of AI buyers and the increased adoption in the workplace are going to provide tons of opportunities. Put your thinking cap on, because now is a great time to start building for the 2022 – 2025 AI wave.