AI, ML & the Consulting Problem
Artificial intelligence and machine learning have fully transitioned from academic concepts, to venture capital investment trends to a suffix on the title of an executive in a large enterprise. Just like we ended up with heads of big data or data science, we’re now starting to see heads of AI and Machine Learning.
It’s not a problem. It just tells us where we are in the cycle – early.
The technology is out of the labs. Watson showed us that, and Google’s AlphaGo took it a step further. These victories for Artificial Intelligence have unleashed a surge in investment. According to McKinsey, companies invested up to between $20-30b in AI initiatives in 2016, with 90% attributed to research and development and the balance for acquisitions. An additional $6-9b went to startups through venture capital and direct corporate investments.
Attaching machine learning and AI to the titles of executives plants a flag in the enterprise. They’re findable via LinkedIn. They’re the obvious choice for a keynote or a panel presentation. It’s a signalling function: “we’re a big company, and we care about new technology. Why don’t you come over and tell us about how you might be able to help.”
The title, however, misses one thing: the use case. What exactly does the Head of Machine Learning and AI do? Is there a head of excel? No. When an organization hires a Head of Machine Learning, they’re hiring a hammer and asking them to look for nails. It’s a decision to fund a technology in search of a problem – a typical feature of the early stages of the technology adoption curve.
There’s nothing wrong with the approach, but it does have important consequences. Just as organizations are open to new technologies, the multitude of possible applications can be overwhelming. Therefore, the experience for an enterprise can feel like chasing butterflies. They’re everywhere, desirable, seemingly easy to catch but also just out of reach.
Enter the first phase of adoption: consulting. Consulting firms are ideally suited for searching conversations about business problems and the various technologies that could solve them. We saw the same effect with the wave of “big data” investing. Big data opportunities became engagements more akin to consulting and proofs of concept. Palantir is a case in point. This naturally favored well capitalized venture backed companies that had the credibility to and were willing to invest in consultative engagements and consulting companies that are organized as such.
What is the head of AI and Machine Learning looking for? In all likelihood, they’re looking for a proof of concept and are leaning heavily on consultants. It’s not a bad thing. It’s just a function of how early we are in the adoption of these technologies. But it is something that should keep entrepreneurs up late worrying about, for they can test and experiment longer than a founder has funding.
