Accrual Acquires Puzzle to Bring AI Close to Practice Management

Accrual acquires Puzzle in a deal that brings the target’s AI-native ledger and month-end close capabilities under one platform. The transaction, announced on 2 September, is the first acquisition for the San Francisco company since it launched seven months ago with $75 million in venture funding. Financial terms were not disclosed.

The deal gives Accrual something it did not build itself: a general ledger and close workflow already serving more than 7,000 organisations in production. Puzzle, founded by Sasha Orloff in 2019, built its platform around an AI-native ledger, a product called AI Close, and collaboration tools branded Cowork and Chat. That technology now moves into Accrual’s broader system, which until this point focused almost exclusively on tax preparation and review for large US accounting firms. The combined entity positions Accrual as one of a small number of vendors attempting to unify tax and client accounting on a single AI-native stack.

Orloff and members of the Puzzle team will join Accrual as part of the transaction. Accrual co-founder and CEO Cosmin Nicolaescu, previously CTO at Brex, described the acquisition as a way to accelerate a client accounting services roadmap that was already in development. The company’s CAS offering is currently in early access with a handful of firms. Nicolaescu built Accrual alongside co-founder and CTO Siddarth Chandrasekaran.

The strategic logic is straightforward. Accounting firms have historically operated tax, audit, advisory and client accounting as separate technology stacks with separate vendors, each requiring its own implementation, training and data model. Accrual wants to consolidate those functions into a single platform where data structured during tax preparation flows directly into close workflows and advisory analysis. Tax was the entry point. Client accounting services, powered by Puzzle’s ledger and close technology, is the second.

To be sure, the gap between announcing a unified platform and delivering one is considerable. Accrual launched in February 2026 and has so far focused on individual tax returns. The company claims its platform processed roughly 15,000 returns during its first tax season, ingesting approximately 1.5 million pages of documents and generating 99.7% of all worksheet fields using AI. Fewer than 1,000 fields across that entire volume required human correction, according to company-reported figures. Those numbers have not been independently verified.

The firm client list lends some credibility. Armanino, a Top 20 US firm, deployed Accrual across six offices during the 2026 tax season and processed thousands of returns through peak, with roughly 70% of its individual returns carrying partnership, trust or S-corp income. H&R Block, Aprio, Creative Planning, BMSS and Stephano Slack are also listed as adopters. Accrual says preparation time drops by more than 85% and review time by up to 60% on the platform. The company claims every 50 complex returns processed effectively adds the capacity of one accountant without increasing headcount.

Puzzle brings a different profile. Where Accrual targeted the largest firms from the outset, Puzzle built its initial customer base among startups and small businesses before expanding into the accounting firm market. The platform had raised $66.5 million across three funding rounds, with its $30 million Series B in November 2023 led by S32 and XYZ Capital. General Catalyst, which led Accrual’s $75 million round, was also a Puzzle investor. Orloff founded Puzzle after building LendUp, a consumer lending platform backed by $325 million in debt and equity financing, and its credit card spinout Mission Lane.

The overlapping investor base raises questions about whether this acquisition reflects genuine strategic fit or a rationalisation of portfolio companies by a shared backer. General Catalyst’s presence on both cap tables does not prove the latter, but it warrants scrutiny. Accrual itself was a Puzzle customer, using the platform to run its own books since early in the company’s history. Whether that user experience informed the acquisition thesis or merely provides a convenient narrative is difficult to assess from outside.

Under the deal structure, Puzzle will retain independent operations for startups and small businesses that use its general ledger directly. The accounting firm technology and the team behind it move to Accrual. Puzzle will no longer market a standalone product to CPA firms. Instead, it will work with accounting partners to deliver combined technology, accounting and advisory services to its small business customers. The carve-out creates a clean division: Accrual serves firms and their clients, Puzzle continues serving its established base of businesses that manage their own books.

The timing is significant for the broader profession. Client accounting services is the fastest-growing revenue line at many mid-market and large US accounting firms. Firms that once viewed CAS as low-margin compliance work now treat it as a scalable advisory channel, and the talent shortage across the profession makes automation of close and ledger maintenance a near-term priority rather than a long-term aspiration. Vendors including Botkeeper, Karbon, Financial Cents and Canopy have built platforms targeting that shift. Accrual’s bet is that bolting CAS capability onto an AI-native tax platform creates a more compelling offering than standalone CAS tools can match.

Integration will determine whether that bet pays off. Accrual expects to begin embedding Puzzle’s ledger and close capabilities following the transaction’s close, which is anticipated in the coming weeks. Broader availability is planned by the end of 2026. For firms evaluating the combined platform, the relevant question is not whether the technology exists but whether two products built by separate teams, with different architectures and different customer bases, can be unified into a single coherent workflow quickly enough to deliver on the promise.

The acquisition also signals something about the AI accounting market’s maturity. Puzzle raised $66.5 million to build an AI-native ledger from the ground up. Seven years after founding, its most viable path forward for the firm-facing side of the business was absorption into a better-capitalised competitor. That outcome is not unusual in enterprise software, but it suggests the standalone AI accounting category may consolidate faster than its funding cycle would imply.

Nicolaescu has described Accrual’s ambition as building one intelligence layer across tax, audit, client accounting and advisory. Acquiring Puzzle fills one of those four boxes. Three remain open.