Middle market AI companies have an opening, FTV's Bernstein says

Middle market AI companies, rather than the largest software incumbents or the newest AI-native startups, may be best positioned to turn generative AI into durable business gains, according to Brad Bernstein, managing partner at FTV Capital. Bernstein framed the opportunity around companies with existing customers, domain-specific data, enough capital to experiment and less organizational drag than legacy vendors. He did not disclose new funding, revenue or headcount figures for the category.

The argument is a useful read on how growth equity investors are sorting software companies in the AI cycle. Bernstein said FTV speaks with thousands of operators and founders each year, and sees the strongest setup in scaled, vertical or workflow-heavy technology businesses that can add AI to systems customers already rely on.

His view cuts against the default assumption that AI economics will accrue mainly to hyperscalers and dominant software platforms. He pointed to Microsoft Copilot and Salesforce agents as real threats, but argued that horizontal products are less suited to specialized, regulated and exception-heavy workflows.

Why could middle market AI companies benefit?

Bernstein's answer is that AI needs operating context, not just model access. In his view, a company that owns the workflow and system of record can become the place AI agents query for decisions, history and exceptions, rather than the software displaced by those agents.

He also cited a PwC finding that three-quarters of AI's economic gains are being captured by 20% of companies. That concentration, he argued, raises the cost of waiting for founders and operators still treating AI as a side project.

Bernstein used recent AI missteps to show that scale is not enough. Klarna, valued at $6 billion in 2024, said its OpenAI-powered chatbot could handle millions of customer conversations and perform the work of 700 customer service employees. Customers disliked the rollout, and Fast Company reported that Klarna was hiring humans back in 2025. He also cited Jasper, whose valuation came under pressure after ChatGPT made its core offerings less differentiated.

What traits does FTV say matter?

Bernstein listed five traits he sees in stronger middle-market technology companies: disciplined self-assessment, speed, ownership of complex workflows, technical capacity and a balance sheet that can support experimentation. The common thread is whether AI improves a defensible customer process or consumes engineering budget without changing the business.

Pricing is one test. Bernstein highlighted Intercom's 2023 move to charge 99 cents per resolved conversation for Fin, its AI agent. He said Fin became Intercom's core product, and CNBC reported that Salesforce recently bought it for $3.6 billion. The broader point is that agentic software can pressure seat-based models and shift vendors toward outcome pricing.

He also pointed to Toast, where product teams used AI to reduce documentation and process work, then connected product changes to external communications with large language models improving instructions for agents.

Two FTV portfolio companies served as examples. Bernstein said Agiloft is building around contract decision workflows, including approvals, negotiations and redlines. He said ReliaQuest, founded in 2007, has worked across more than 1,000 customer environments and shifted some SOC analysts into product development roles as automation increased.

The investor case is clear enough: AI favors companies that already sit inside critical work. The unproven part is how many middle-market vendors can change pricing, product and go-to-market execution quickly enough before larger platforms package similar functions into existing suites.

This story draws on original reporting from Crunchbase News.