Show HN: We Built an AI Agent to Recover Past-Due Accounts

We built an AI agent platform to automate accounts receivable and debt recovery for businesses. The idea came from a commercial lender friend who struggled to find a collection agency willing to take on their $300K in unpaid loans—most agencies only work with $1M+ accounts. That made us wonder: How many businesses are simply writing off smaller debts because collections isn't accessible?

Originally, we were selling compliance AI software to fintechs, banks, and collection agencies, but the sales cycle was painfully slow. While working through that, we saw how inefficient collection agencies were and realized there was a better way. Instead of agencies, we started working directly with SMBs, SaaS companies, and commercial lenders—businesses that often struggle to recover unpaid invoices.

So far, the reaction has been strong: decent close rates, real traction, and clear ROI. But top-of-funnel is a challenge because our ideal customer profile (ICP) is broad.

We’re wondering how others in B2B AI automation handle this:

Do you niche down aggressively, or let the market pull you toward a vertical? What’s worked for you when refining ICP for a broad solution? Would love to hear thoughts from other founders who’ve tackled this problem!