Enterprise Software and the Innovator's Dilema

I was recently interviewed on Ticker News to talk about the coming disruption AI and AI-written software will bring about in the enterprise software market. Enterprise software providers are finding themselves in a classic “innovators dilemma” where the original innovators will not be willing to cannibalize their own business in order to compete with the threat posed by AI, namely that AI is going to re-write all enterprise software at a fraction of the cost. 

An example of this kind of disruption is that of the rise of mini-mills in the steel industry, in which small, nimble players used electric arc furnace technology to produce steel from scrap at much lower costs than traditional integrated steel mills. Initially focusing on low-margin products like rebar, mini-mills improved their technology over time, moving upmarket to higher-quality steel products that had been the stronghold of integrated mills. These incumbents, heavily invested in expensive blast furnace operations, failed to adapt due to sunk costs, organizational inertia, and reluctance to cannibalize their own business. As a result, mini-mills like Nucor gained dominance, reshaping the industry and forcing many traditional steelmakers into bankruptcy or irrelevance.

Can you imagine SalesForce or Oracle offering to have their AI re-write Oracle EBS or SalesForce for pennies on the dollar? These enterprise software providers are way too invested in their existing products and existing revenue streams to even consider competing with themselves in this kind of way, hence the dilemma!

How quickly will this happen? Some think that it’s going to be very soon, but I think this is one of those situations in which technologists are over-estimating how quickly things will change in the short-term but perhaps even under-estimating how dramatically they will change in the long-term. 

It will be a while (10 years?) before Enterprise CIOs fire up the army of AI “Developers” to re-write their ERP systems. That’s largely because CIOs are inherently risk averse. The old saying “no one ever got fired for choosing IBM” (I’m dating myself) is still true, only today it is “no one ever got fired for choosing SalesForce.com”. It may take 15-20 years before AI-developed business applications are really “mainstream”. But at some point there is a dramatic shift that puts these existing businesses out of business.

Having AI re-write core business applications is, today, a risky proposition. But that is going to change in a big way. Imagine a world in which the low-risk option is to have AI developers re-write your core applications instead of buying off-the-shelf software. Imagine a world in which the term “developer” no longer refers to a human being. Both are going to happen.

See what you think of my interview and let’s discuss!