Thomas Dowling has already sold one accounting startup. The price tag—$10 million in cash plus $2 million in goodwill to Opendoor—wasn't life-changing money in venture terms, but it bought him something more valuable: credibility and a clearer view of what finance teams actually hate doing.
Now he's back with FullSeam, and this time the pitch is more ambitious. Forget better software dashboards. Forget alerts that tell CFOs what's wrong without fixing it. Dowling and his co-founders—Geoff Segal and Aaron Coppa, the same team from TaxProper—are building AI agents that log into QuickBooks, chase down missing purchase orders, and reconcile bank statements while the humans sleep.
It's automation, yes, but not the kind that requires ripping out your existing tech stack. FullSeam's software works inside the tools companies already use. No migration projects. No change management consultants. At least, that's the promise.
Software That Actually Does the Work
The concept sounds almost too straightforward. FullSeam's agents integrate with the usual suspects: QuickBooks, Stripe, Xero, NetSuite, Salesforce, and a handful of others whose logos populate the company's website. What those agents do once connected is where things get interesting—or, depending on your tolerance for letting AI handle money, unnerving.
On accounts receivable, the software doesn't just track who owes what. It follows up with customers. Submits invoices into their AP portals. When someone flags a missing purchase order, the agent resolves the exception and drafts the follow-up email. For cash application—matching incoming payments to the right invoices—the AI does the matching and, crucially, explains its reasoning.
The accounts payable side is similarly hands-on. Vendor bills arrive via email, portals, PDFs. The AI extracts details, matches bills to purchase orders or contracts, suggests general ledger codes with confidence scores, and posts entries directly into accounting systems. Bank reconciliation follows the same pattern: compare statements to ledger data, auto-match what's clear, flag what isn't, maintain an audit trail.
Every move is logged. Companies set approval thresholds and business rules. Exceptions—and there are always exceptions in finance work—get routed to humans with context already attached.
Whether this holds up at scale, beyond the early customers FullSeam has attracted, is the open question. Finance is full of edge cases. A missing PO number. A vendor who changed their name. A payment that should have hit but didn't. The art of accounting often lives in those moments.
The Founders Who've Done This Before
Dowling, a Rhodes Scholar, isn't exactly new to the tedium of financial automation. TaxProper, the startup he and his co-founders built before this one, automated property tax workflows for real estate firms. Opendoor bought it in November 2022 for $10 million in cash plus $2 million in goodwill, according to SEC filings—a tidy outcome, if not the kind that makes headlines.
The trio spent some time at Opendoor post-acquisition, then registered FullSeam as a Delaware entity in New York on December 4, 2025. By winter, they were part of Y Combinator's batch, backed by partner Brad Flora. No other funding amounts have surfaced publicly, which is typical for companies this early. YC's standard check gets you in the door; the rest depends on what happens next.
Early Metrics, With the Usual Caveats
Two customers have emerged in company-published case studies, both from late December. Belfry reported that FullSeam auto-validated 95% of invoices before they reached the finance team, slashed billing disputes by 40 to 60%, and gave back more than 10 hours per month. Deferred, another early adopter, automated all referral payouts and shifted from monthly to real-time disbursements while saving similar time.
These are company claims, not third-party audits. Still, the specificity—10 hours here, 95% there—suggests the kind of ROI that finance leaders hunting for headcount relief tend to care about. Time back. Fewer disputes. Faster cycles. The math either works or it doesn't.
No Rip-and-Replace (Really?)
The "no workflow changes required" pitch is FullSeam's main differentiator, at least on paper. Instead of forcing companies to route data through a new platform, the agents log into existing tools. The setup, according to FullSeam, involves connecting systems, configuring rules, deploying agents. Most teams are supposedly live within days.
If that holds true, it's a meaningful contrast to the traditional enterprise software playbook: months-long implementations, workflow redesigns, armies of consultants. Finance automation projects stall all the time because the juice doesn't seem worth the squeeze. FullSeam is betting that lowering the friction—no migration, no retraining—will get CFOs over the finish line.
Whether "days, not months" scales beyond a handful of mid-market companies remains to be seen.
The Crowded, Noisy Market
FullSeam is hardly alone in chasing AI-powered finance automation. Pilot announced an "AI Accountant" in early February, billing it as fully autonomous with human oversight. Vic.ai expanded AP autonomy and payment intelligence in its Q1 product updates. HighRadius rolled out what it called "Agentic AI" across receivables, payables, treasury, and closing functions sometime in early 2025. Workday and BlackLine both shipped finance agents by mid-year.
The difference, FullSeam insists, is execution over assistance. "AI operators, not workflows," the marketing copy reads. The pitch is that FullSeam acts within the existing tech stack rather than adding another layer. For CFOs exhausted by tools that surface insights but still require someone to do the actual work, that might resonate.
Or it might be a distinction without much difference. The market will decide.
What Comes Next
FullSeam's website offers no public pricing—just a demo request form. The go-to-market motion suggests a focus on mid-market or enterprise customers, given the complexity of the workflows the agents are designed to handle. It's unclear whether pricing is per agent, per module, or based on transaction volume. The company hasn't said.
For finance leaders trying to make sense of the AI automation wave, FullSeam represents a particular bet: that the next leap in productivity won't come from smarter dashboards or better alerts. It will come from software that logs in, does the work, and only bothers humans when something genuinely requires judgment.
Whether that vision holds up depends on how well the agents navigate the messy, exception-filled reality of finance operations. The demo is always clean. The real world, less so. And that gap—between what the AI can handle smoothly and what still needs a human touch—will determine whether FullSeam becomes a category winner or just another well-intentioned automation tool that couldn't quite deliver on the promise.
