Most professional workflows could theoretically benefit from AI tooling. That observation is close to useless. The question is which ones are actually worth building for — where the opportunity is real, the moat is buildable, and the timing is right. Four criteria separate the good targets from the ones that just look good.

Document volume is high and high-stakes. This is the baseline. If the workflow doesn’t involve large volumes of complex documents where the stakes of getting something wrong are meaningful, AI doesn’t add enough value to justify adoption friction. The document requirement is also a capability filter: domain-specific document analysis is hard to do well, which is why most general-purpose AI tools fail at it.

Existing tools are separate SaaS platforms, not workflow-native. If professionals are already using an AI tool embedded in their primary workflow, the opportunity is gone. The valuable gap is the workflow that currently requires toggling between four separate applications — a VDR for documents, a spreadsheet for modeling, a data platform for comps, an email thread for review notes. When the workflow exists across disconnected tools, adding a workflow-native alternative has a clear value proposition: time, context, and reduced error from switching.

No major AI-native incumbent has established dominance. This is the most important criterion and the one most often overlooked. Some professional verticals that appear unserved are actually dominated by AI companies that aren’t consumer-visible but are deeply embedded in enterprise buying cycles. The tell is whether major law firms, investment banks, or healthcare systems have already signed multi-year contracts with an AI vendor for this specific workflow. If they have, you are competing against switching costs, not against an empty market.

The professional has a clear time-pressure deadline. Urgency drives adoption. A professional who needs a document review completed before a board meeting has a different relationship to friction than one doing exploratory work with no deadline. Deadline-driven workflows also tend to have clearer quality metrics — the output either meets the standard required for the decision, or it doesn’t. That clarity makes evaluation and adoption easier.

These four criteria together form a filter. Applying it to any given professional workflow quickly separates the ones worth pursuing from the ones that are theoretically interesting but practically hard. Workflows that score 4/4 are rare. When one appears, it’s worth moving on quickly — because the same criteria that make the opportunity good are the same ones that will eventually attract competition.

+++