Accrual is bringing Puzzle’s AI-native ledger and close technology into its growing CAS platform. But Puzzle itself isn’t going away.
On September 2, Accrual announced that it had agreed to acquire Puzzle’s accounting-firm technology and business.
The deal includes Puzzle’s AI-native ledger and the technology behind AI Close, Cowork and Chat. Puzzle founder Sasha Orloff and members of the team behind its accounting-firm products will also join Accrual once the transaction closes.
The companies haven’t disclosed the purchase price. And since this is still an agreement, the deal needs to go through the usual closing process over the coming weeks.
But here’s the part that’s easy to miss.
Puzzle will continue independently, focused on startups and small businesses.
So, what exactly is changing?
It helps to first understand what the two companies were building.
Accrual is fairly new. It launched in February 2026 with $75 million in funding led by General Catalyst and deliberately started with tax preparation and review for accounting firms.
Tax, Accrual says, gave it one of the toughest places to prove its AI model: compressed seasons, high stakes and deadlines that don’t move.
The idea is also different from simply giving accountants an AI chatbot.
Accrual’s agents work from the client record, source documents and the firm’s own methodology to prepare more of the engagement, while calculations, open questions and source support remain available for review. The accountant still owns the judgment and final decision.
But Accrual never planned to stop at tax.
Its larger goal is to build one intelligence layer across tax, audit, client accounting services, or CAS, and advisory. And CEO Cosmin Nicolaescu says customers had been asking Accrual to bring the platform into CAS since it launched.
And that’s where Puzzle starts to make sense. Puzzle came at accounting from another direction. It was built around the books themselves.
Think about what happens every month for a bookkeeping or CAS client. Accounts need to be reconciled. Unusual transactions may need investigating. Adjustments have to be made. And eventually the books need to be closed before anyone can rely on those numbers for reporting, tax planning or advisory work.
Puzzle spent years building technology around that process through its AI-native ledger, AI Close, Cowork and Chat. The company says its platform serves more than 7,000 startups, small businesses and accounting firms.
There’s another connection worth knowing.
Accrual has used Puzzle to run its own books since early in the company’s history.
So Accrual isn’t buying technology it has only seen in a product demo. It already knows Puzzle as a customer.
And the timing fits.
Accrual’s CAS product was already in early access before this deal. So Puzzle isn’t giving Accrual a new direction. It gives the company technology and people that could help it get where it was already going much faster.
Cosmin made much the same point in his LinkedIn announcement: Accrual’s customers were already asking for CAS, and as the two teams spent more time together, continuing to build separately made less sense.
That matters because tax and CAS run on different rhythms.
Tax has busy seasons and fixed deadlines.
CAS brings a firm into a client’s financials month after month.
And once those monthly books are clean and approved, that information becomes the starting point for financial reporting, forecasting, tax planning and advisory work.
Today, that context can sit across ledgers, spreadsheets, statements, emails and separate workflow tools. Accrual’s idea is to keep more of it connected so a reviewer can see the source support, what changed, why it changed and what still needs attention without rebuilding the engagement from scratch.
Or as Cosmin Nicolaescu put it, combining the two teams could bring “production-ready AI to CAS faster.”
Well, Puzzle isn’t disappearing. And Sasha Orloff’s own explanation probably makes the split easiest to understand.
He said that over the past year, it became clear that Puzzle was really running two businesses inside one company: one for accounting firms and another for small businesses.
Both were real opportunities, he wrote, but they needed different homes if they were going to win.
That means the accounting-firm technology and team move to Accrual, while Puzzle 2.0 continues with a sharper focus on startups and small businesses.
Puzzle 2.0 will be led by Ozgur Uzuner as CEO and Jason Mitchell as CTO. Orloff is joining Accrual full time but will remain on Puzzle’s board.
And Puzzle 2.0 isn’t simply carrying on with the exact same model.
Source: Post | LinkedIn
Puzzle says it learned that customers wanted more than software. They wanted accounting and tax services powered by AI and automation, but still led by people. At the same time, accounting firms using Puzzle were asking the company to help them grow.
So Puzzle 2.0 plans to combine its technology with accounting and tax services, its customer-acquisition engine and a smaller group of accounting-firm partners.
For existing customers, however, the immediate message is continuity. Puzzle says their data stays with Puzzle, while pricing, terms and support contacts remain unchanged.
And Accrual has been clear about something similar on its side.
It says the combination is about technology, product and talent rather than consolidating customers.
Probably not that they need to change anything tomorrow.
The transaction still has to close, and Accrual’s CAS product remains in early access. After closing, the company expects to begin bringing Puzzle’s ledger and close capabilities into its platform, with broader availability planned by the end of 2026.
But the direction is worth paying attention to.
Much of the early conversation around AI in accounting has been about tools that help professionals answer questions or finish individual tasks faster.
Accrual and Puzzle are working toward something broader: AI that helps prepare the work itself, while keeping the evidence, methodology and professional review connected around the engagement.
For Accrual, Puzzle gives it an important piece of that plan: the ledger and recurring close work that sits underneath much of what accounting firms do throughout the year.
And Puzzle gets a clearer path too.
Accrual can focus on building technology for accounting firms across tax, CAS, audit and advisory.
Puzzle 2.0 can focus more directly on the accounting and tax experience startups and small businesses are asking for.
So, for now, the deal looks less like one company disappearing and more like two businesses getting clearer paths.
If you have a take on what this could mean for accounting firms, we’d love to hear it.
And if this helped you understand the deal, share it with someone who likes to keep up with what’s changing in accounting.
A practical guide for CPA and tax firms on Form 8962, the premium tax credit reconciliation, its repayment caps, and how to prepare it without missing the numbers that matter.
A practical guide for CPA firms on Form 8865 filing categories, attribution rules, common mistakes, and how to prepare it correctly.
A practical guide for CPA and tax firms on Form 5498, why it shows up after the filing deadline, and how to review it without second-guessing a finished return.