Company Profile · Fintech & AI · New York
Concourse
AI agents built for corporate finance teams - that show their work.
AI agents built for corporate finance teams - that show their work.
Finance teams sit on a paradox: some of the most analytical people in a company spend their days moving numbers between spreadsheets. Concourse was built to reclaim that time. Its AI agents connect directly to a company's existing financial systems and let finance, accounting and data teams retrieve, analyze and report on their numbers by asking in plain English.
Rather than replace the tools a company already runs, Concourse sits on top of them as a reasoning layer. An agent can read from an ERP such as NetSuite or QuickBooks, a data warehouse like Snowflake, and a billing platform like Stripe at the same time - then answer a question that spans all three. There is no migration, no rip-and-replace, and, according to the company, onboarding for early-stage teams takes about five minutes.
The output is not a raw chat transcript. Concourse generates presentation-ready deliverables - PDFs, Excel models, PowerPoint decks - and can run reporting on a recurring cadence. Crucially, every result carries an audit trail: users can trace a figure back to the source data and the underlying SQL and Python the agent wrote to produce it. In finance, where numbers have to be defensible, that transparency is the product's spine.
The company frames the payoff plainly. Teams using Concourse, it says, cut manual work by more than 75% and produce roughly six times more analysis than before - turning reports that once took days into work that lands in minutes.
How do we enable CFOs to cycle through much more analysis of that information? That is the mandate we have.Matthieu Hafemeister, Co-Founder & CEO
Concourse's customers are corporate finance, accounting and data teams - and the company is deliberately size-agnostic. A free self-serve tier onboards two-person startups; a high-touch enterprise offering serves Fortune 500 finance departments. In between sit usage-based Starter and Professional plans built for teams scaling in the middle.
Named customers span that spectrum: Fortune 500 security firm Palo Alto Networks, communications unicorn Front, and high-growth Western brand Tecovas, alongside Lightmatter, Unify, CurbWaste, Instabase, Maximus and Pave. The company reported a 13x increase in its customer base over the year leading into its Series A.
The problem Concourse targets is old and familiar. Finance data is scattered across ERPs, warehouses, billing tools and CRMs, perpetually out of sync. The traditional response - hire more analysts, build more spreadsheets - eventually buckles. As Hafemeister puts it, teams end up choosing between rigid software and flexible spreadsheets and get "the worst of both worlds."
Concourse's bet is that AI agents finally collapse that trade-off, delivering the flexibility of a spreadsheet with the rigor of a system of record - and freeing finance leaders to spend time on the business rather than on data plumbing.
Every team has the same problem around data, getting access to it, and making decisions off of that information.Matthieu Hafemeister
Query financial data across every connected system in natural language for instant, ad hoc analysis and charts.
Run reporting workflows on a custom cadence so recurring analysis produces itself instead of a person producing it.
Generate presentation-ready outputs - PDF, Excel and PowerPoint - directly from agent results.
Trace every result to its source data and the underlying SQL and Python logic - the transparency finance compliance demands.
Agents pull from multiple data sources at once, answering questions that span the ERP, the warehouse and billing - not one silo at a time.
Transparency panels surface the exact SQL and Python behind each number. Where generic LLMs are a black box, Concourse is auditable.
Agents are measured against financial-compliance standards, not general chat quality - built for a domain where being wrong has consequences.
It layers on top of existing systems rather than replacing them, so adoption doesn't hinge on a multi-quarter implementation.
SOC 2 Type II, SAML/SSO and custom agent development for teams on NetSuite, Snowflake and SAP.
Built by operators who scaled fintech Jeeves to a $2.1B valuation and lived the finance-data problem firsthand.
It's not allowed, effectively, to produce numbers that cannot be traced back, and that are not effectively accurate.Matthieu Hafemeister, on why finance AI must be auditable
The alternatives: finance teams reaching for generic assistants like ChatGPT, incumbent FP&A and reporting platforms such as Anaplan, Pigment, Cube, Datarails and Mosaic, and a wave of newer AI-for-finance startups. Concourse's wedge is reasoning across the whole stack with an audit trail, rather than dashboards or static models.
Early-stage teams onboard in about five minutes via QuickBooks, Ramp or Stripe - no custom setup.
Growing teams pay as they use, with 7-day free trials to prove value before committing.
High-touch support for NetSuite, Snowflake and SAP users, plus SOC 2 Type II, SSO and bespoke agents.
The hybrid model - free at the bottom, contracted at the top - lets Concourse land small teams and expand into the enterprise as they grow, while opening the same enterprise-grade agents to companies that could never staff a dedicated AI team.
That positioning places Concourse in the fast-forming category of agentic AI for the back office. It is a bet that finance, long served by rigid planning software and improvised spreadsheets, is ready for a reasoning layer that speaks the language of the finance team and respects the discipline of the audit.
Matthieu Hafemeister and Ted Michaels, both from fintech Jeeves, launch Concourse with Y Combinator backing.
Co-led by a16z and CRV, with Y Combinator and BoxGroup. Profiled by TechCrunch.
Led by Standard Capital; agents open to finance teams of any size.
A former partner at Andreessen Horowitz who went on to scale fintech Jeeves. He calls prompting "probably the number one skill you need to have in the next 10 years to be successful."
Formerly Jeeves' head of finance, with an MS in Computer Science from Georgia Tech. He and Hafemeister met at Jeeves and helped scale it to a $2.1B valuation.
Being able to just work on the same goal together, no egos, just solving problems - it's really fun.Matthieu Hafemeister, on co-founding Concourse
Concourse builds AI agents for corporate finance teams that connect to a company's existing ERPs, accounting systems, data warehouses and billing platforms, letting teams query, analyze and report on financial data in natural language - with a full audit trail behind every result.
It was founded in 2023 by Matthieu Hafemeister (CEO) and Ted Michaels, who previously worked together at the fintech Jeeves. The company is based in New York and is a Y Combinator alum.
Concourse raised a $4.7M seed round in October 2024 and a $12M Series A in January 2026 led by Standard Capital, with a16z, CRV and Y Combinator participating.
Concourse reasons across multiple connected financial systems and exposes the underlying SQL and Python behind every result, so calculations are traceable and auditable - a requirement in finance that generic LLMs do not meet.
Named customers include Palo Alto Networks, Front, Tecovas, Lightmatter, Unify, CurbWaste, Instabase, Maximus and Pave - spanning Fortune 500 enterprises, unicorns and high-growth startups.