Would You Use An Outside AI Program To Automate A Core Competency?

How AI could change what is "core"

Happy Sunday and welcome to Investing in AI. I’m Rob May, co-founder of the AI Innovator’s Community, which does events in Boston and New York so please sign up if you are interested.

I’ve been thinking a lot today about how AI business models, like turning services into software ,might impact how we think about core competencies in business. In classic business strategy we outsource things that aren’t core if they can be done more cheaply by someone else, and we keep things in-house that help define our core business model and value proposition.

It seems that the last wave of tech innovation sold software that aided work, but wasn’t core to most companies. Building a CRM, for example wasn’t something most companies would benefit from doing internally. But AI is different. If you consider Sarah Tavel’s “sell work, not software” piece as a viable future for AI (and I do) then the question becomes, when you sell “work”, what will people outsource and what will they not?

The difference between AI and what came before it is, AI is more algorithmic. And algorithms can be more core. A high frequency trading firm doesn’t usually sell their software to others because it’s more valuable to own it and use it yourself to trade. I think AI might take more companies in this direction. What I mean is, as more things become algorithmic and AI driven, and software can do more human tasks - can you “sell work” or, will people consider it too core?

If you outsource your core competencies to someone else, particularly a 3rd party AI company, then that tool is getting better and regardless of what it does or doesn’t do with your data, it may enable more people to compete with you. So the question is - when a company comes to you with an AI product that can help automate something you consider core, do you buy it?

There are two possible ways this can be resolved in the future. One is, yes, companies buy it, and the source of competitive advantage changes to areas not yet affected by AI. The second is, no, you do it in house. This latter outcome is the one I’ve been thinking about and what it means for investing in AI. I find it difficult to wrap my head around this so let’s try with two examples that come to mind.

If you were at Bain, Mckinsey, BCG, or some place similar, would you buy a “consulting agent” from a software company? If you were a PE firm would you buy an analyst agent from a software company? If every part of your value chain can be done with AI eventually, is it safe to offload it all to third parties? It probably depends on how valuable the “stringing it all together” part is to your customer base. But also, this would make “stringing it all together” easier for everyone else.

I’ve honestly been grappling with where the key vectors of competition and competitive advantage are in an AI world. I’m not there yet but, I’m leaning towards the idea that AI will level the playing field for areas like operations and innovation. I keep coming back to the things AI can’t change, and I think, counterintuitively, those may be some of the best places to invest to make money off of AI.

The fixed capacity in the modern world is attention. It seems like commanding someone’s attention is the most valuable thing to have or be. That generally cements incumbent positions, except in cases of extreme novelty. Maybe the AI people are right, and attention is all you need.