From Tax Season to Every Month

When building Accrual, we deliberately started with tax.

Tax is one of the profession’s hardest operating environments: compressed seasons, high stakes, fragmented source data, and deadlines that do not move. Every number has to tie back. Every unresolved question has to surface. And every return ultimately has to withstand professional review.

Those constraints shaped Accrual from the beginning: agents prepare the work, the tax system performs authoritative calculations and filing, and professionals retain judgment and control. That model has now operated through real busy seasons, at real firms, on real client work.

But an accounting firm does not exist only during tax season. The same clients need their books closed, questions answered, plans evaluated, and decisions supported throughout the year. From the beginning, Accrual was built for that larger relationship.

Our ambition is to build one intelligence layer across the firm: tax, audit, client accounting, and advisory. Not by treating those practices as interchangeable, but by giving each of them the same foundation: the full client record, the firm’s methodology, connected source support, and a clear path through review.

Clients experience one firm and one relationship, even when the work crosses several practices. Yet inside many firms, each engagement begins in a different system and rebuilds context the firm already has. Solving tax alone would address only one part of that problem.

Today, we are taking the next step. Accrual has entered into an agreement to acquire Puzzle’s accounting-firm technology and business. Puzzle founder and CEO Sasha Orloff, and the team behind Puzzle’s work with accounting firms will join Accrual following the deal closing.

Why CAS

Tax has its season. Client accounting work happens every month.

The month-end close sits at the center of the ongoing firm-client relationship. It becomes the foundation for reporting, tax planning, advisory work, and many of the decisions clients make throughout the year.

But the work is still frequently spread across ledgers, spreadsheets, statements, emails, and disconnected workflow tools. Teams chase source data, reconcile accounts, roll schedules forward, draft adjustments, investigate variances, and prepare reporting. Reviewers often enter late and have to reconstruct what changed, why it changed, and what remains unresolved.

That is the kind of work we believe should change.

The prior close should become the starting point for the next one. Source support should remain connected to the reconciliation or adjustment it supports. Missing evidence, unusual activity, and judgment calls should reach the reviewer with the relevant work already in view. Once the books are approved, the close should become the starting point for advice, not a file the team has to rebuild somewhere else.

Accrual had already begun building toward that model, and our CAS offering is now in early access with accounting firms. Puzzle allows us to move much faster.

Why Puzzle

Puzzle did not add a chat interface to a legacy ledger. It built an AI-native ledger and accounting workspace around what the technology can now do, with products including AI Close, Cowork, and Chat.

More importantly, Puzzle arrived at many of the same product principles we did. Accounting work is not valuable because an AI-generated answer looks plausible. It is valuable when the source support is visible, the calculation or procedure can be followed, open questions are explicit, and an accounting professional can review and approve the result.

Puzzle has put those ideas into production for thousands of startups, small businesses, and accounting firms. Accrual has also used Puzzle to run its own books since early in our history. We know the product both as builders and as customers.

Bringing Puzzle’s accounting-firm technology and team into Accrual allows us to combine its ledger and close capabilities with Accrual’s client record, firm methodology, permissions, and review model. The goal is not to place two products beside each other. It is to build one working layer for the engagement.

Together, the teams can direct more product and engineering investment toward modernizing client accounting and advisory work, giving the profession a faster path from isolated AI tools to production workflows built around firm methodology and review.

Why now

Over the past year, AI crossed an important threshold for accounting. It became capable of doing meaningful preparation work in production – not independently and not without controls, but with the evidence, methodology, and professional review required to make that work useful.

That changes the pace at which the industry can move.

Firms cannot wait for a platform to be built one service line at a time while their capacity constraints grow more acute. The pipeline of new CPAs is not keeping pace with the professionals leaving the field, and clients are asking their firms to do more, not less.

We could have continued building every part of the CAS platform ourselves. But firms need the capability sooner, and Puzzle has already spent years developing the technology and operating experience required to deliver it. Bringing the teams together gives our customers a stronger answer materially faster.

Sasha and we reached many of the same conclusions independently: AI has to work at the engagement level, it has to fit the firm rather than sit beside it, and it has to give professionals more leverage without removing their judgment.

When two teams independently arrive at the same view of the industry, that is usually a good reason to stop pursuing it separately.

What this means, and what doesn't change

This combination is about technology, product, and talent, rather than consolidating customers.

Puzzle’s business serving startups and small businesses that use its AI-native ledger directly will continue to operate independently. The structure gives each business a clear mandate: Accrual will remain singularly focused on accounting firms and the clients they serve, while Puzzle continues supporting its established base of businesses.

For firms already using Accrual, nothing changes about the platform or support they have today. Our CAS offering is already in early access, and that work will continue. Following the close, we expect to begin incorporating Puzzle’s ledger and close capabilities, with broader availability planned by the end of 2026.

What changes is what we can build next, and how quickly we can deliver it.

To Sasha and the team joining us: welcome! To Puzzle’s customers and everyone who helped build the company: the work matters, and the technology developed for accounting firms is about to reach many more of them.

We are building Accrual one real workflow, one production engagement, and one practice at a time. Tax proved the model under pressure. CAS brings that model into the work firms do with their clients every month. Puzzle gives us a stronger foundation and a faster path forward.

We started with tax. We are building for the whole firm.