Fin Turns Help Docs Into Agents

This exposed why old bots rarely became a real budget line, they saved labor only after a company spent labor to script, tune, and babysit them. Intercom’s own second generation Resolution Bot could reach about 50% resolution, but only when teams put in heavy setup work. The strategic break with Fin is that it turns existing help docs into a working support agent fast enough that automation can be bought as a product, not built as an internal project.

• Earlier bots worked best in narrow, repetitive environments like food delivery, where most tickets fit a small set of known problems. For the average SaaS company, question volume was too messy and too varied, so hard coded flows and fuzzy matching never paid back the implementation effort.
  1 sacra 3 sacra
• The economic shift is concrete. Third generation AI support agents can resolve conversations for about $1.40 to $1.50 each versus roughly $10 to $15 for a human handled resolution. That is why pricing moved from seat licenses toward per resolution charges like Fin at $0.99 per resolved ticket.
  1 sacra 2 sacra 4 sacra
• Once the bot is good enough, the product surface gets much bigger than chat. The winning system ties together the bot, help center, ticket inbox, agent assist, reporting, and customer data, so a human can answer an edge case once and the system gets better the next time. That is the core difference between an AI wrapper and a full support platform.
  2 sacra 3 sacra

This market is heading toward a split between standalone AI agents and integrated help desk platforms that bundle automation with human workflows. Intercom’s path is to use Fin to grow beyond seat based support software and become the operating system for AI first customer service, while newer players like Sierra and Decagon push the whole category toward higher autonomous resolution and broader adoption.