End Close, a San Francisco payments reconciliation startup founded in 2026, has raised a $3.5M seed round led by Foundation Capital to automate transaction matching.
Key Takeaways
- End Close raised a $3.5M seed round led by Foundation Capital, announced November 17, 2026.
- Y Combinator, Liquid 2 Ventures, Sunflower Capital and Angel Invest also joined; valuation is undisclosed.
- Founded in early 2026, the company uses AI agents to resolve payment exceptions.
Lead
End Close, a San Francisco startup building payment reconciliation software, announced a $3.5M seed round on November 17, 2026. Foundation Capital led the round. Y Combinator, Liquid 2 Ventures, Sunflower Capital, Angel Invest and a group of individual angels also took part. The company did not disclose its valuation.
Sean Bolton (CEO) and David Newell (CTO) founded the company in early 2026. The money will go toward engineering, go-to-market and implementation hires.
What Does End Close Actually Build?
End Close sells software that checks whether every payment a company moves ended up where it was supposed to. The product targets high-volume money-movement businesses, including card programs, neobanks, payroll providers, marketplaces, remittance apps and payment processors.
The system reconciles records across banks, processors, card networks, ERPs and data warehouses. A deterministic, rules-based matching engine handles the bulk of transactions. AI agents then investigate and resolve the exceptions that rules cannot match, which is where finance teams traditionally lose the most time.
The company says it can resolve 99% of exceptions automatically on day one and reach 99.999% within six weeks. Those are the company's own figures, and no customers have been named publicly to corroborate them.
Why Does Reconciliation Still Need Fixing?
Reconciliation remains a manual chore at many payments firms because data arrives in different formats from different counterparties and rarely lines up cleanly. Teams typically run it as a month-end close, so errors surface weeks after the money has moved.
End Close pitches continuous reconciliation as the alternative. Matching transactions as they happen would shrink the window in which a missing or duplicated payment goes unnoticed. Bolton framed the problem plainly, saying he has "spent most of my career watching smart people lose days to reconciliation, which is just checking that every payment ended up where it was meant to."
The pitch has a built-in tension. Finance and compliance teams tend to distrust probabilistic systems near the ledger. By keeping deterministic matching as the core and confining AI to exception handling, End Close is betting that auditors will accept a hybrid design more readily than a model-first one.
What Does the Round Say About Investor Appetite?
A $3.5M seed from a firm of Foundation Capital's standing, alongside a Y Combinator slot, signals that investors still back narrow, back-office AI products with a clear buyer and a measurable pain point. The round is small by recent AI standards, which suggests the company has been funded to prove a wedge rather than to chase a platform thesis.
The undisclosed valuation and the absence of named customers limit how much can be read into the terms. There is no prior round to compare against, since the company is less than a year old. The real test will be whether the claimed resolution rates hold in production at payment volumes where a single percentage point of unmatched items means thousands of tickets.
The founders are National University of Ireland graduates, and the story drew coverage in Irish media as an Irish-founded company, though the business is headquartered in the United States.
What Comes Next for End Close?
The company plans two product releases. RecBench is a benchmark for measuring how well AI models perform on reconciliation tasks. A standalone AI agent for simpler reconciliation work is also planned.
The benchmark is a notable choice. Publishing an open yardstick lets End Close define how accuracy is measured in a category where every vendor quotes its own numbers. The standalone agent could widen the funnel to smaller companies that do not need a full platform, though it may also invite comparison with general-purpose AI tools that finance teams already use.
Hiring is the near-term priority. The company is recruiting engineers, go-to-market staff and implementation specialists, the last of which matters because reconciliation deployments depend on integrations with each customer's banks, processors and ERPs.
Outlook
End Close enters a crowded corner of finance automation with a small round, a hybrid rules-plus-AI design and accuracy claims it has yet to prove publicly. Named customers, independent results from RecBench and the performance of the standalone agent will show whether the 99.999% figure survives contact with real payment flows. The next funding round will likely turn on those answers.



