If you are layering agents onto an existing workflow to make it slightly faster, you are playing the wrong game — Christensen's framework tells you where the real opening is, and it is not where your competitors are looking.
Clayton Christensen's The Innovator's Solution moves disruption theory from diagnosis to prescription. The core move is to stop competing for established customers and instead find people who have simply given up — the non-consumers who lack the skill, money, or access to use existing solutions at all. Pair that with jobs-to-be-done thinking, where you treat a product not as a thing but as something a customer hires to make progress in a specific situation, and you get a targeting discipline that cuts through feature debates. Christensen also argues for staying integrated when performance is still short of what the market needs, and only modularizing once the architecture is good enough to hand off.
Agents obliterate the skill and access barriers that once kept non-consumers out of high-complexity work — legal review, financial modeling, technical writing, medical triage. That is your non-consumption opening. Jobs-to-be-done thinking becomes the design brief for your agent: what specific progress is a user trying to make, and what gets in the way. Christensen's integration rule applies directly to trust architecture — where agent output quality is still uneven, you keep humans in the loop not because you are cautious but because that is where your product earns its margin and its credibility.
- Find the population that stopped trying, not the one already served
- design each agent around a single job with a clear definition of done
- stay integrated at every capability layer where performance isn't yet reliable enough to modularize.
