Accrual is stepping into accounting with the kind of ambition that usually shows up in infrastructure companies, not point solutions, announcing its public launch alongside a $75 million funding round aimed squarely at rebuilding how preparation and review actually happen inside firms. The round, led by General Catalyst with participation from Pruven Capital and Edward Jones Ventures, signals a clear belief that accounting is finally ready for AI-native systems rather than bolt-on automation. That distinction matters, because the pressure on firms right now isn’t cosmetic. It’s structural. Regulations pile up, client data arrives in every imaginable format, and expectations for speed keep rising, while the underlying workflows still depend on fragmented tools and manual handoffs that feel oddly untouched by the last two decades of software progress.
What Accrual is proposing is not “faster data entry,” but a rethinking of the workflow itself. Preparation and review are treated as a single continuous process rather than two loosely connected stages. AI agents act as a preparer that actually reads everything clients send in, from K-1s and 1099s to spreadsheets, emails, photos, and those massive statements nobody wants to scroll through. The system organizes data as it comes in, flags what’s missing, generates targeted follow-up questions, and produces draft returns that are already structured for professional judgment rather than cleanup. The subtle shift here is important: review stops being a forensic exercise and becomes an informed decision layer built on clean, traceable inputs. You can almost hear some senior reviewers quietly exhale at that thought.
That philosophy comes straight from Accrual’s leadership. CEO and co-founder Cosmin Nicolaescu frames accounting as an interconnected system that has been artificially chopped into tasks by legacy software. Instead of forcing firms to rework the same information across multiple tools, Accrual aims to let expertise compound over time, with each step reinforcing the next. Over the past year, this hasn’t been theoretical. The platform has been tested with firms including H&R Block and Armanino, along with several other Top 100 firms, and the early results read less like marketing fluff and more like operational relief. Preparation time dropping by more than 85 percent, review time cut by up to 60 percent, and every fifty complex returns effectively adding the capacity of a full accountant without adding headcount. In an industry quietly strained by talent shortages, that last point lands hard.
What’s also notable is what Accrual isn’t trying to do, at least for now. It integrates directly with existing tax engines, supporting prior-year data, carry-forwards, and final filing workflows instead of insisting firms rip and replace systems they already trust. That pragmatic approach lowers resistance and keeps the focus on where AI actually adds value. John Karls of Armanino put it bluntly: returns coming out of Accrual are already at manager-review quality, and the system is catching issues that might have led to amended returns later. That’s not just a productivity win, it’s a risk story, and accounting firms listen very closely when risk gets smaller.
From an investor perspective, the appeal is obvious. Accounting is massive, mission-critical, and deeply conservative for good reasons. Marc Bhargava at General Catalyst framed the opportunity as rebuilding core workflows that have remained largely unchanged for decades, and betting on a team experienced in regulated environments to do it without breaking trust. That last word keeps coming up, and it should. AI in accounting doesn’t get a free pass on accuracy or explainability, and Accrual seems keenly aware that adoption will only happen if professionals feel amplified, not replaced or exposed.
For now, Accrual is focused on individual tax returns, but the roadmap clearly points outward. The company is actively onboarding firms across the U.S., expanding its team, and positioning itself as long-term infrastructure rather than a seasonal tool. If this approach scales the way early results suggest, it could quietly reset expectations across the profession. Not with flashy promises, but with something accountants tend to respect even more: fewer late nights, fewer surprises in review, and systems that finally behave as if they understand how accounting actually works.
Accounting Has Been Asking for This for Years
Anyone who has spent time inside an accounting firm knows the uncomfortable truth: this market should have been overhauled years ago. Not incrementally patched, not wrapped in prettier dashboards, but fundamentally rethought. The sheer volume of manual coordination, document chasing, reconciliation, and repetitive review work that still defines accounting workflows today would be unthinkable in most other mission-critical professions. And yet it has persisted, season after season, normalized by regulation, tradition, and a quiet resignation that “this is just how accounting works.” AI arriving in this space isn’t a disruption so much as a long-delayed correction.
The core issue isn’t that accountants resist technology. It’s that the rules of the game have been frozen in time. Accounting has become a compliance machine built on layers of legislation written for a paper-first world, where every exception, attachment, and footnote adds friction that software is forced to absorb rather than question. Firms end up compensating with human labor instead of systemic clarity. Staff spend countless hours translating chaotic real-world inputs into forms that satisfy regulators rather than meaningfully improve financial understanding. Review becomes a defensive exercise, less about insight and more about making sure nothing explodes later. That kind of system doesn’t reward expertise; it exhausts it.
AI-native platforms finally expose how artificial much of this burden is. When preparation and review can be unified, when documents can be read holistically instead of line by line, it becomes obvious that a large share of accounting work exists only to bridge gaps created by outdated regulatory structures. The technology isn’t just faster, it’s revealing inefficiencies that were previously hidden behind human effort. Once software can consistently flag missing data, reconcile inconsistencies, and surface risk early, the question naturally follows: why are we still forcing the same complexity into the system in the first place?
This is where governments enter the conversation, and frankly, they are overdue too. Legislators tend to treat accounting complexity as a necessary evil, assuming that more rules automatically mean more accuracy or fairness. In practice, excessive complexity often achieves the opposite. It increases error rates, raises costs for individuals and small businesses, and funnels advantage toward those who can afford armies of professionals. Simplifying accounting legislation where possible would not weaken oversight; it would strengthen it by making compliance clearer, more transparent, and easier to validate both by humans and machines. AI doesn’t replace regulation, but it absolutely changes how regulation should be designed.
What’s striking is how rarely policymakers acknowledge that accounting is now an information systems problem as much as a legal one. When rules are written without regard for how data is collected, structured, and reviewed, the burden doesn’t disappear, it just shifts downstream into firms and taxpayers. AI makes that misalignment painfully visible. If governments continue to legislate as if every return is assembled manually from scratch, they will lock inefficiency into law just as the tools to eliminate it become viable. That would be a self-inflicted drag on productivity at a time when economies can least afford it.
The accounting profession itself understands this, even if it rarely says it out loud. Partners don’t want their best people buried in mechanical preparation. Clients don’t want to pay for hours that add no strategic value. And younger accountants are increasingly unwilling to accept a career defined by seasonal burnout and rote tasks. AI doesn’t threaten the profession; stagnation does. A modernized regulatory framework, paired with AI-native systems, could finally allow accountants to operate as analysts and advisors at scale rather than highly trained data janitors.
The overhaul now underway isn’t radical, it’s corrective. The real risk is not that AI moves too fast, but that legislation moves too slowly, preserving complexity for its own sake. Accounting has waited a long time for tools that reflect the reality of how work actually happens. Governments should take the hint. Simplify where possible, modernize where necessary, and let both professionals and technology focus on what actually matters: accuracy, insight, and trust, not paperwork theater.
