Agents Drive Vertical B2B SaaS

AI pushes SaaS toward narrower products because the hardest part is no longer making a dashboard, it is wiring deeply into the exact workflow and data model of one industry. ServiceTitan won by fitting how plumbers, HVAC shops, and electricians actually schedule jobs, quote work, dispatch techs, collect payments, and upsell in the field. Agents make that specialization more valuable, because the best agent needs live data and write access across the systems that run a specific business, not a generic wrapper over many apps.

• ServiceTitan started with one trade cluster, then expanded after it had the core workflow. Its product tied together office scheduling, technician mobile workflows, payments, financing, inventory, and marketing. That is what vertical means in practice, one system shaped around how a contractor makes money day to day.
  2 sacra
• Ampersand's broader point is that incumbents still own the main databases, but outside developers can build agents around them. A vertical agent might pull signals from Gmail, Slack, Gong, and a field service system, then trigger next steps inside the system of record. Horizontal platforms will cover common use cases, while the long tail gets built by specialists.
  1 sacra
• Healthcare shows the same pattern. AI scribes can use generalized integration layers to get basic coverage across EHRs, but the winning products go much deeper into one system or specialty. In markets with messy data models and old software, depth beats breadth because the buyer wants the job fully completed, not partially automated.
  3 sacra

The next wave of B2B software is likely to look like agent layers wrapped around entrenched systems of record, with more category leaders emerging from ignored verticals. The biggest winners will be products that own a concrete workflow in one industry first, then expand outward once they control the data flows, the automation points, and the monetization hooks around that workflow.