T:0 vs Puzzle: The New AI Accounting Landscape

Ten days changed the AI accounting market for US startups. On 2 September 2026, Accrual announced it would acquire Puzzle's accounting-firm technology and business, with Puzzle continuing independently for direct startup and SMB users. On 1 September 2026, one day earlier, T:0 exited private beta and launched publicly as an AI-native accounting platform built by Airwallex, backed by a $320M Series H at an $11B valuation. Together the two announcements reshaped the platform choice facing every VC-backed startup that has been staring at their QuickBooks Online subscription and wondering what to do about the August 2026 price increases.

If you are that founder, or the accountant advising them, the question is now direct: T:0 or Puzzle. Both are AI-native. Both are US-only. Both target the same core segment of Delaware C-corps on modern fintech stacks running Stripe, Mercury, Brex, Ramp, and Gusto. Both promise to automate the categorisation and close work that QuickBooks Online has historically required manual attention on. Both include an option to bring a human CPA into the loop for the monthly close.

But they are structured very differently, priced very differently, and built by very different companies with different strategic pressures. Puzzle has three years of platform maturity and a certified advisor ecosystem. T:0 is 10 days old but arrives with Airwallex distribution, cross-border payments infrastructure behind it, and pricing that packages the platform and the bookkeeper together in a way Puzzle does not. Neither is universally better than the other. Which one fits depends on specific factors that this post walks through in detail.

I write this as a Certified Puzzle Advisor. That means I have a genuine bias toward Puzzle, and the fair thing to do about that is to name it upfront rather than pretend I do not. What follows is my honest read of both platforms based on public documentation and my Puzzle certification, calibrated as fairly as I can manage. Where T:0 clearly wins, I say so. Where Puzzle clearly wins, I say so. Where the answer is genuinely "it depends," I say that too, and explain what it depends on.

💡 Key Takeaways

• T:0 launched 1 September 2026 as an AI-native accounting platform built by Airwallex, currently US-only

• Puzzle continues independently for direct startup and small business users after the 2 September Accrual acquisition

• Pricing structure differs materially: Puzzle sells the platform alone at $30-$360/mo, T:0 sells platform ($199) or platform-plus-bookkeeper ($599+) bundles

• T:0 packages a CPA into the monthly plan; Puzzle requires engaging a Certified Puzzle Advisor separately

• Puzzle has three years of platform maturity; T:0 is 10 days old with no independent tests published yet

• T:0 claims a hybrid architecture where the LLM writes accounting rules and a deterministic engine applies them (vendor claim)

• T:0 claims 12,000+ integrations (vendor claim, not independently verified)

• Both platforms target US-based VC-backed SaaS startups on modern fintech stacks

• Neither is currently built for e-commerce operations with Shopify, Amazon, inventory, or multi-state sales tax needs

• Verified 11 September 2026: both platforms will change rapidly in coming months

What Actually Happened in Those 10 Days

The strategic backdrop matters because both moves reflect the AI accounting market entering a consolidation phase. Understanding what happened helps you read the platform comparison in the right context rather than as isolated product decisions.

1 September 2026: T:0 launches

Airwallex publicly announced T:0's exit from private beta on 1 September 2026. The platform had been in private beta since around the $320M Series H raise on 25 June 2026 that valued Airwallex at $11B. The launch covered US startups only. Coverage from Finextra, FF News, and TechJournal was substantially press-release rewrites without added reporting or hands-on testing. That gap has not been filled yet by any independent review.

The T:0 pitch is a continuously-closed ledger built on a four-agent architecture: the Accountant for daily categorisation and reconciliation, the Scout for reading your business model and building a custom chart of accounts, the CFO for conversational questions on cash and runway, and the Integrator for connecting banks, payroll, cards, and billing. The company claims 12,000+ integrations and a hybrid technical architecture where an LLM interprets business context to write explicit accounting rules that a deterministic engine then applies to transactions. Both claims are vendor-stated and remain untested in public.

2 September 2026: Accrual announces Puzzle acquisition

The next day, Accrual and Puzzle jointly announced that Accrual would acquire Puzzle's accounting-firm technology and business, including the AI-native ledger, AI Close, Cowork, and Chat products. Puzzle founder and CEO Sasha Orloff and members of the Puzzle team would join Accrual. Critically, Puzzle continues operating independently for startups and small businesses that use its platform directly. The deal was expected to close in the coming weeks with broader Accrual integration by end of 2026.

Accrual's context is worth naming because it clarifies why the acquisition happened. Accrual launched in February 2026 with $75M in funding led by General Catalyst, focused on being "the intelligence layer for modern accounting firms." Their customer base includes Armanino, Aprio, H&R Block, Creative Planning, BMSS, and other Top 100 firms. The acquisition gave them proven AI-native ledger technology to accelerate expansion into Client Accounting Services, and it gave Puzzle's firm-focused business scale within a larger platform.

For a full analysis of what the Accrual-Puzzle deal specifically means for existing Puzzle users, my complete acquisition analysis walks through it thoroughly.

What both moves together mean

The pattern across those 10 days is the AI accounting market completing a phase transition. For roughly three years, AI-native accounting platforms competed on product quality and feature velocity against QuickBooks Online, Xero, and Sage as legacy incumbents. Puzzle, Digits, Zeni, Truewind, Pilot, Fondo, Rillet, Campfire, and others built genuinely new products that solved real problems the legacy platforms did not address well. That phase produced platform diversity but not consolidation.

The T:0 launch and Accrual acquisition together signal the next phase. Well-capitalised players with distribution advantages (Airwallex's payments customer base, Accrual's Top 100 firm relationships) are moving into the market with strategies that combine product with distribution rather than competing on product features alone. The independent AI-native platforms now face pressure that they did not face six months ago.

For founders making platform decisions today, the takeaway is straightforward. The category is not going away. Multiple platforms will exist for years. But the players that were most established six months ago are not necessarily the best answer six months from now, and the newer entrants are worth genuine evaluation rather than dismissal based on age alone.

The Companies Behind the Platforms

Before comparing pricing and features, understanding who built each platform and why matters more than usual because both are early-stage products where the founding team's judgment shapes the roadmap significantly.

Puzzle

Puzzle was founded by Sasha Orloff, previously the founder and CEO of LendUp. The platform launched publicly in 2022 with Series A funding, subsequently raised additional rounds, and by mid-2026 served over 7,000 customers spanning startups, small businesses, and accounting firms. Puzzle's positioning has consistently been "the modern AI-native alternative to QuickBooks Online for VC-backed startups," and the product reflects that focus. Deep native integrations with Stripe, Mercury, Brex, Ramp, and Gusto. Real-time burn and runway dashboards. Automatic revenue recognition under ASC 606. A Certified Advisor program with three tiers (Preferred, Design, Exclusive) that supports startups working with independent CPAs.

The Accrual acquisition of the accounting-firm side of Puzzle's business changes the corporate structure but not the product for direct customers. Sasha Orloff's departure to Accrual removes founder-level product vision, which is a meaningful change and creates uncertainty about long-term direction that honest analysis has to acknowledge. New product leadership will emerge. The roadmap may evolve. The platform functions today as it did before the announcement, and Accrual's CEO explicitly stated no immediate required change for customers.

T:0

T:0 was founded by Lance Co Ting Keh, who previously founded OpenPay (subscription and billing infrastructure, acquired by Airwallex), held engineering leadership roles at X the Moonshot Factory (working on AI drug discovery and computer vision), was an early engineer at Cresta, and started machine learning at Box. He holds BS and MS degrees from Duke in Computer Science and Electrical Engineering. He is also a Venture Partner at Bessemer. That is a substantial technical background, and it matters for evaluating T:0's ambitious architectural claims because someone with Lance's history has actually built and shipped machine learning systems at scale rather than just talking about them.

The company sits inside Airwallex. That context matters for several reasons. Airwallex is a serious global fintech with $1.3B in annualized revenue as of March 2026, growing 74% year-on-year, $287B in annualized transaction volume, and licences in the US, UK, EU, Australia, and multiple APAC markets. They have distribution advantages that a standalone startup does not, particularly among cross-border operators and businesses that already use Airwallex for payments. The Airwallex ownership also creates a specific structural question about independence that I address separately later in this post.

Why the ownership matters for the comparison

Puzzle is now, functionally, an independent platform whose founder just left to run product at a different company. T:0 is a subsidiary of a $11B fintech whose founder is still running it. These are not equivalent situations, and they affect the answer to "which platform will still be actively developed in three years" in ways worth taking seriously without over-reading.

Puzzle's continued independent operation post-acquisition is genuine, and the team is intact. But absent founder-level vision, product roadmap velocity may slow or shift. T:0 has a founder actively building it and Airwallex resources behind it, but Airwallex's own strategic priorities can shift, and a subsidiary product's roadmap depends on parent-company decisions that customers do not control.

Neither situation is bad. Both are risks worth pricing into a multi-year platform decision.

Pricing: The Most Important Practical Difference

Pricing is where the platforms diverge most sharply, and it is the single most important factor for most startups making the decision. Compare the tiers directly.

Puzzle pricing (verified 11 September 2026)

Annual billing saves roughly 15-20%. All plans include unlimited users. The Complete tier carries a "50% faster close or your money back" guarantee, though it applies only to single-entity companies with all-API integrations that onboard within seven days.

Working with a Certified Puzzle Advisor for monthly bookkeeping is separate from the subscription. Advisor pricing varies but typically ranges $1,000-$4,500 per month depending on complexity, on top of the platform subscription.

T:0 pricing (verified 11 September 2026)

T:0 packages the CPA into the plan rather than requiring a separate advisor engagement. That is the structural difference that determines whether you compare T:0's $199 tier to Puzzle Complete at $120 or T:0's $599 tier to Puzzle Complete plus advisor at $120 plus $1,000+.

The honest read of the pricing comparison

The two pricing comparisons that matter for real decisions:

• Platform only: Puzzle Complete at $120/mo vs T:0 Platform at $199/mo. Puzzle is 40% cheaper at this tier. What Puzzle Complete gives you: AI-native ledger with 98% claimed auto-categorisation, real-time dashboards, revenue recognition, deep native integrations, three years of platform maturity, established Certified Advisor ecosystem. What T:0 Platform gives you additionally: the Scout agent that auto-generates a custom chart of accounts from your business description, the CFO agent for conversational Q&A on cash and runway, the four-agent architecture, and the hybrid deterministic-rules approach. For most startups, Puzzle Complete is the better value at this comparison. T:0's differentiation at the platform tier is genuine but not obviously worth 66% more.

• Platform plus CPA: Puzzle Complete plus Certified Advisor at $120 + $1,500-$2,500 = $1,620-$2,620/mo vs T:0 Managed Bookkeeping at $599/mo. T:0 is dramatically cheaper at this tier, roughly one-third the cost. This is where T:0's packaging strategy becomes commercially significant.

But wait, the comparison is not apples-to-apples. Puzzle Advisors are independent CPAs (or advisor firms) with their own relationships, judgment, and business models. The T:0 Managed Bookkeeping CPA sits inside Airwallex on Airwallex's payroll, handling multiple T:0 clients as part of their job. Whether these two things are equivalent depends on what you actually need from a bookkeeper. If you need a stable relationship with a specific person who knows your business over multiple years and provides judgment beyond bookkeeping (fundraising prep, complex transactions, tax strategy coordination, board reporting help), an independent Certified Advisor delivers that better. If you need CPA-signed monthly close and standard bookkeeping executed reliably at low cost, T:0 Managed Bookkeeping delivers that better and at 60-75% less than the Puzzle-plus-advisor equivalent.

What Airwallex just told you about bookkeeping

The most important thing about T:0's pricing is that Airwallex has published, in dollar terms, what they think an AI-native ledger without a human is worth ($199) and what they think an AI-native ledger with a human CPA signing the close is worth ($599). That is a 3x price premium for adding the human, on the same platform. It is Airwallex telling you, publicly, that the AI handles the feed and the rules, but the human still signs the close, reviews the reconciliations, and takes professional responsibility. This is not a criticism of T:0. It is a data point about where the industry believes automation stops and human judgment starts.

The $2,000-$8,000 per month figure that T:0 explicitly targets as the SMB spend on part-time bookkeeping plus fractional CFO plus tax prep is a vendor claim, not third-party research, and worth reading as positioning. What it tells you is where T:0 believes it competes: not against $50/month DIY tools but against the full stack of services a growing startup buys.

Architecture: How the Platforms Work Under the Hood

Beyond pricing, the platforms take different technical approaches to the same underlying problem. Both categorise transactions using AI. Both maintain accrual books. Both connect to modern fintech stacks. But they do it differently, and the differences matter for reliability.

Puzzle's architecture

Puzzle uses AI for categorisation with what the company describes as "governed automation" where all AI actions are traceable, reviewable, and editable. The reported auto-categorisation accuracy is up to 98%, particularly when the account is connected via API to major fintech tools rather than through statement imports. Users can manually recategorise transactions, and the system learns from corrections. On the Complete tier and above, an AI Accuracy Review feature flags vendor and category inconsistencies proactively, catching errors before month-end.

The AI is not autonomous. It drafts, and a human (either the founder or an advisor) reviews and approves. That is the intentional design pattern. Puzzle has been consistent about this from launch: the AI works fastest when the underlying accounts are on API integrations, the human retains authority over categorisation decisions, and the platform's job is to make the human's work an order of magnitude faster rather than to eliminate it.

Puzzle also offers "AI Credits" as a resource metric (25 lifetime credits on lower tiers, monthly allowances on Complete and Scale) where certain AI actions consume credits. That is worth knowing if you plan heavy AI-driven workflows.

T:0's architecture (vendor claims)

T:0 describes a four-agent architecture where each agent has a specific role: the Accountant handles daily categorisation, reconciliation, and audit trail; the Scout reads your business model description and builds a custom chart of accounts with metrics; the CFO handles conversational Q&A on cash, burn, runway, and revenue; and the Integrator connects to banks, payroll, cards, and billing tools.

The specific architectural claim that differentiates T:0 from other AI-native platforms is the hybrid model. T:0 describes it this way: the AI (LLM) interprets business context and writes explicit accounting rules. A deterministic engine then applies those rules to transactions. The LLM does not categorise each transaction directly. Every entry is traceable and, per T:0, GAAP/IFRS-consistent.

This is genuinely interesting because it addresses the reproducibility problem that has been the main critique of LLM-only categorisation approaches. If the rules are explicit and the engine is deterministic, the same transaction gets the same categorisation every time, and you can inspect the rules directly. If the rules are wrong, you fix the rules once rather than fighting the LLM every close.

The important caveat is that no independent test of this architecture has been published as of 11 September 2026. Nobody has stress-tested it against the transaction patterns that historically break rule engines: a bundled payout containing sales, refunds, fees, and tips; a partial refund crossing a period boundary; a complex intercompany transaction; an inventory write-off; an FX gain on multi-currency operations; an equity comp true-up. T:0 launched 10 days ago, and the practitioner review layer that normally forms around a platform launch has not formed yet. Attribute the hybrid architecture as a vendor claim, take it seriously because Lance's technical background makes it plausible, but do not treat it as proven.

The honest architectural comparison

Puzzle has three years of production usage validating its AI categorisation approach at 98% claimed accuracy. That is a track record. It is not a guarantee, and users have reported difficulty overriding AI decisions in some cases, but the platform has been stress-tested in the field for long enough that its actual behaviour is knowable.

T:0's architecture is more ambitious on paper. If the hybrid deterministic-rules approach works as described, it would represent a genuine improvement over pure LLM categorisation for reliability and auditability. But "if it works as described" is doing a lot of work in that sentence. Ten days into a public launch, the market simply does not know yet. Give it six months.

For decisions today, the practical implication is this. If you value proven, if you have volume that would stress-test unproven architecture, or if you cannot tolerate categorisation surprises during a fundraising close, Puzzle's three years of maturity is a real advantage. If you have room to be an early adopter, if you value auditability and reproducibility highly, or if T:0's architectural claims resonate with technical judgment you trust, T:0's approach may prove out and reward early adoption.

What Each Platform Actually Includes

Feature comparison at the platform level, based on public documentation from both companies as of 11 September 2026.

Both platforms include

• AI-driven transaction categorisation from connected accounts

• Real-time balance sheet, income statement, and cash flow

• Cash and accrual bookkeeping (Puzzle maintains both simultaneously; T:0 defaults to accrual)

• Revenue recognition workflows

• Deep integration with Stripe, Mercury, Brex, Ramp, and Gusto

• Investor-ready reporting

• Startup-specific dashboards (burn, runway, ARR)

• SOC 2 compliance and standard security infrastructure

• US-only operation (both explicitly)

Puzzle-specific features

• AI Accuracy Review (Complete and above) that proactively flags vendor and category inconsistencies before month-end

• Certified Advisor ecosystem with independent CPAs on Puzzle's public directory

• 50% faster close guarantee on Complete tier for qualifying single-entity companies

• Free tier with meaningful functionality until $20K monthly transaction volume

• AI Credit system for tracking AI action usage

• Three years of feature velocity with continuous shipping

T:0-specific features (vendor claims where noted)

• Four-agent architecture (Accountant, Scout, CFO, Integrator) as the framing structure

• Scout agent that auto-generates a custom chart of accounts from your business model description

• CFO agent for conversational Q&A on cash, burn, runway, and revenue

• Hybrid deterministic-rules architecture (vendor claim, untested publicly)

• 12,000+ integrations (vendor claim, not independently verified against integration directories that publish counts)

• Bundled CPA option at $599/month tier

• Bundled tax filing at $749/month tier for federal plus one state

• Airwallex payments integration likely to deepen over time given corporate ownership

The features neither has (worth naming)

• Deep e-commerce integrations (Shopify, Amazon, WooCommerce, BigCommerce)

• Inventory management and COGS handling

• Multi-state sales tax nexus tracking and filing

• Multi-currency operations (both are US-only)

• Full audit-ready trail with GAAP disclosures for M&A-scale diligence

• Robust support for complex holding-company structures

If you are running an e-commerce business, neither platform is the right answer today. My Amboras QuickBooks integration analysis covers what to do about e-commerce specifically, and my analysis of why the AI accounting wave is US-first explains why the ideal-fit customer profile is so narrow.

Ideal Customer Profile for Each Platform

Both platforms target US-based VC-backed SaaS startups. Within that segment, they fit slightly different specific profiles.

Puzzle fits best when

• You are a Delaware C-corp or LLC with a US EIN and US bank accounts

• Your fintech stack is Mercury, Brex, or Ramp for cards, Stripe for payments, and Gusto or Rippling for payroll

• Your revenue is subscription or usage-based SaaS, or a mix that includes both

• You are pre-Series A through mid-scale (up to roughly $50M ARR)

• You want to work with an independent CPA or Certified Puzzle Advisor rather than a platform-employed bookkeeper

• You value founder-friendly self-service with strong UI and platform maturity

• You have engineering resources or a technical founder comfortable with API-driven integrations

• You want to preserve your relationship with an existing advisor while modernising the platform

T:0 fits best when

• You are a Delaware C-corp or LLC with a US EIN, and specifically one considering or already using Airwallex for payments

• You want platform-plus-bookkeeper bundled in a single monthly bill rather than managing two vendors

• You value CPA-signed close at a lower total cost than platform-plus-advisor separately

• You have complex international payments needs that Airwallex specifically serves

• You are willing to be an early adopter on a 10-day-old platform with no independent reviews

• You value the specific architectural approach (hybrid deterministic rules, four-agent framing) over platform maturity

• You want to bundle tax filing into the same relationship at $749/month tier

• Your ideal outcome is one vendor handling everything rather than multiple specialists

Neither fits well when

• You run e-commerce operations (Shopify, Amazon, WooCommerce, Etsy)

• You have multi-currency or non-US operations at material scale

• You need multi-state sales tax nexus tracking and filing

• You have inventory management and COGS handling needs

• You have historical books that need cleanup before the platform can add value (this is where my catch-up cleanup service specifically fits, before you land on any AI-native platform)

• You are running a holding company structure with multiple entities requiring consolidation

• You need audit-ready GAAP disclosures for M&A or public offering diligence

The Scenarios Where Each Wins

Practical scenarios rather than abstract feature checklists.

Scenario 1: Pre-seed founder, sub-$500K raised, minimal transactions

Winner: Puzzle Starter (free tier). Puzzle's free tier until $20K monthly transactions is genuinely useful and covers this stage completely. T:0's platform tier at $199/mo is real money for a pre-seed team, and the additional features are not yet justified by the operational complexity of the business. Revisit when you cross Series A.

Scenario 2: Series A SaaS startup, $5M raised, $50K MRR, scaling team

Winner: depends on how you want to work. Puzzle Complete at $120/mo plus a Certified Puzzle Advisor at $2,000/mo totals $2,120/mo and gets you platform maturity, independent CPA relationship, and someone who will help with fundraising prep across future rounds. T:0 Managed Bookkeeping at $599/mo gets you platform plus a bundled CPA at less than one-third the cost. If the independent relationship matters to you, Puzzle wins. If cost matters more, T:0 wins. Both are defensible answers, and neither is universally correct.

Scenario 3: Series B SaaS company, complex revenue recognition, multiple entities

Winner: Puzzle Scale plus advisor at this stage, with T:0 CFO Services as the closest alternative. Puzzle's three years of platform maturity handling complex revenue recognition is genuinely valuable at this scale, and a senior Certified Advisor brings judgment that a T:0 platform CPA is unlikely to replicate. T:0's CFO Services tier is priced custom and could compete, but there is no public information on what it delivers at that tier yet, and taking that risk at Series B is generally not a good bet.

Scenario 4: Startup considering Airwallex for cross-border payments

Winner: T:0 likely wins here because the platform-payments integration is the specific reason Airwallex built T:0. If your operations are naturally aligned with Airwallex's cross-border strengths, T:0 gains a bundle advantage that Puzzle cannot match. The one caveat is that T:0 is US-only at launch despite Airwallex's global infrastructure, so this synergy is prospective rather than immediate.

Scenario 5: Startup preparing for M&A or IPO diligence

Winner: Puzzle plus experienced advisor, at least for now. The AI-native platforms in general are relatively new to producing books that survive M&A diligence, and Puzzle's three-year track record beats T:0's 10 days for this specific need. Additionally, working with an independent Certified Advisor who has taken clients through diligence before is more valuable than working with a platform-employed bookkeeper who may or may not have that experience. Neither answer is ideal for this scenario, and QuickBooks Online plus a specialised CPA still wins for the most demanding diligence-preparation cases.

Scenario 6: Existing Puzzle user hearing about T:0 for the first time

Winner: stay on Puzzle until you have a specific reason to move, and then evaluate. Platform switching is expensive in time and disruption. The Accrual acquisition changed Puzzle's corporate structure but not the platform for direct users. T:0 is genuinely interesting but 10 days old. Give it 6 months, watch for independent reviews, and re-evaluate. If Puzzle's roadmap noticeably slows under new ownership dynamics, then T:0 becomes a stronger case at that point.

Scenario 7: Existing QuickBooks Online user evaluating both

Winner: Puzzle for most cases, with T:0 as a defensible alternative if the bundled CPA is specifically what you want. Post the August 2026 QBO price increases (41% on Plus, 70% on Advanced), migrating to an AI-native platform is more economically justified than it was six months ago. My complete Puzzle vs QuickBooks Online comparison walks through that migration case, and everything in it applies today. T:0 becomes the alternative when you want to bundle the CPA in from day one.

What Existing Puzzle Users Should Know

If you are currently on Puzzle and evaluating whether the T:0 launch changes anything for you, the honest answer is: probably not, and not yet.

Puzzle continues operating independently for direct startup and small business customers per the Accrual acquisition announcement. Your subscription, platform access, integrations, workflows, and support continue as before. Sasha Orloff's departure to Accrual creates uncertainty about long-term product direction, but the immediate operational reality is unchanged.

T:0 exists now as an option. It did not exist for you 10 days ago. But it is 10 days old, and no independent reviews or hands-on tests have been published. Making a platform switch based on a 10-day-old alternative is generally a bad decision because the switching cost is high and the downside risk of the new platform not maturing as expected is real.

The reasonable posture for existing Puzzle users:

1. Continue using Puzzle as normal for the next quarter at minimum

2. Watch T:0's early customer signal as it accumulates over the next 3-6 months

3. Watch Puzzle's product velocity under post-acquisition dynamics

4. Re-evaluate at year-end 2026 when both platforms will have clearer trajectories

5. Talk to your Certified Puzzle Advisor about your specific situation because the answer varies by business

For deeper context on how the Accrual acquisition specifically affects Certified Advisor relationships and the wider Puzzle ecosystem, my Accrual acquires Puzzle analysis covers those angles in detail.

What T:0 Prospects Should Know About Puzzle

If you are evaluating T:0 as your first AI-native accounting platform and are unfamiliar with Puzzle, the fair thing to do is give Puzzle serious consideration as your alternative because it has attributes T:0 does not.

Puzzle has three years of platform maturity and 7,000+ customers who have stress-tested it in production. That track record is not a nice-to-have. It is the difference between deploying a platform that works today and deploying one that might work well in a year. For a growing startup where finance function reliability directly affects fundraising credibility, that difference matters.

Puzzle's Certified Advisor ecosystem includes independent CPAs whose businesses depend on serving Puzzle clients well over multi-year relationships. Those advisors are not employed by Puzzle. They are independent professionals who have chosen to specialise in the platform. That structural independence brings judgment, relationship depth, and long-term continuity that T:0's platform-employed bookkeeper model does not naturally provide.

Puzzle's Complete tier at $120/month is 40% cheaper than T:0's platform-only tier at $199/month. For founders who are comfortable running their own books (or working with a fractional bookkeeper they already have), the cost difference at this comparison is meaningful.

None of this makes T:0 wrong. It makes Puzzle a legitimate alternative that deserves serious evaluation rather than dismissal. If after evaluating both you conclude T:0 fits your specific situation better, that is a valid conclusion. If you conclude Puzzle fits better, that is also valid. What is not valid is choosing T:0 without evaluating Puzzle, because the two most commonly-heard reasons for not evaluating Puzzle (I never heard of it before, and I heard the founder just left) are both bad reasons for skipping a serious platform comparison.

For a comprehensive introduction to Puzzle's platform capabilities, my complete Puzzle guide for startups walks through the platform in depth.

Migration Considerations

If you decide to switch from one to the other, or from QuickBooks Online to either, migration is a real project. Both platforms handle migration well but with different specifics.

Migrating from QuickBooks Online to Puzzle

Puzzle has a mature QuickBooks Online migration process refined over three years. My complete QuickBooks Online to Puzzle migration guide walks through it step by step. Standard turnaround is 3-5 business days per Puzzle's official documentation. Free migration is included through Puzzle's team at getstarted@puzzle.io. Three migration options exist based on the level of historical detail you need preserved. No downtime is required with parallel running during transition.

Migrating from QuickBooks Online to T:0

T:0's migration process is publicly less-documented because the platform is 10 days old. What is known: T:0 targets QuickBooks Online migrations as a specific use case in their launch messaging, the Managed Bookkeeping tier includes onboarding, and the Scout agent's job is to read your business model and build a custom chart of accounts that would incorporate mapping from your existing QBO structure. Specific timeline, cost, and process details would need to come from T:0 directly during your evaluation.

Migrating from Puzzle to T:0

This is possible but not straightforward. Both platforms accept QBO imports. Neither has published a Puzzle-to-T:0 migration path specifically because the use case did not exist before 1 September 2026. The realistic approach is to treat Puzzle as your source system, export chart of accounts, transaction history, and balances, and import to T:0 through their standard onboarding process. Expect friction, expect a specific migration project, and expect it to take longer than the equivalent QBO-to-T:0 migration.

Migrating from T:0 to Puzzle

Same conceptual friction in reverse, plus the additional problem that T:0 is 10 days old and export capabilities may or may not match what a migration would require. If you deploy T:0 today and decide six months from now to move to Puzzle instead, expect some pain. This is a real risk of being an early T:0 adopter that Puzzle adopters do not carry to the same degree.

The safest migration posture in general

For any platform migration, do it at the start of your fiscal year rather than mid-year. Run parallel for at least one month. Preserve historical data outside the platform being replaced. Document your chart of accounts before migration. Do not migrate during fundraising, audit, or major business events. Work with an accountant who has done the specific migration before.

The Bottom Line

T:0 and Puzzle are both credible AI-native accounting platforms for US-based VC-backed startups. Neither is universally better than the other. The right choice depends on specific factors about your business, your team, and how you want to work.

For most pre-seed to early Series A startups, Puzzle wins today for three reasons: three years of platform maturity, meaningful independence in advisor selection, and lower cost at equivalent tiers. The Accrual acquisition changed Puzzle's corporate structure but not the platform for direct users, and the immediate operational reality for existing and new Puzzle users is unchanged.

For startups that specifically value bundled CPA-plus-platform at a lower total cost than Puzzle-plus-advisor, or that have Airwallex payments alignment, or that want to be early adopters of an ambitious architectural approach, T:0 is a defensible alternative. The 10-day-old status is a real caveat that should not be dismissed, but Lance Co Ting Keh's technical background and Airwallex's resources give T:0 a realistic chance of maturing into a genuinely differentiated platform over the next 12-24 months.

Neither platform serves e-commerce operations, cross-border with material multi-currency needs, or multi-state sales tax filing well as of 11 September 2026. If those describe your business, both are wrong answers today and QuickBooks Online plus specialised tooling remains the practical choice.

For existing Puzzle users, staying put is the correct default. Watch T:0's trajectory over the next 3-6 months. Watch Puzzle's product velocity under post-acquisition dynamics. Re-evaluate at year-end 2026 when both platforms will have clearer trajectories.

For existing QuickBooks Online users on Plus or Advanced tiers absorbing the August 2026 price increases (41% and 70% respectively), the migration case to an AI-native platform is stronger now than it was six months ago. Puzzle Complete at $120/mo remains cheaper than QBO Plus at $140/mo post-increase, and the AI capabilities have compounded three years of development. T:0 at $199/mo Platform tier is more expensive than QBO Plus but bundles capabilities QBO does not offer natively. Both are legitimate migration destinations depending on how you want to work.

The AI accounting market has entered a phase where multiple credible platforms exist and platform choice deserves genuine analysis rather than defaulting to whichever platform got mentioned first. Both T:0 and Puzzle have earned their place in the evaluation set for the segment they serve. Neither should be dismissed. Both should be pressure-tested against your specific business needs rather than adopted based on marketing.

Ready to make the T:0 vs Puzzle decision for your business?

Most SaaS founders we talk to about this decision came in expecting a clear-cut answer and left understanding it depends on their specific situation. That is not a diplomatic answer. It is what the honest analysis actually shows. The right platform for you depends on how you want to work with a CPA, how much platform maturity matters relative to bundled cost, whether Airwallex payments alignment adds value for your business, and where you are in the growth curve.

At Catch Up Clean Up, we serve US-based startups on Puzzle, on QuickBooks Online, on Xero, and we are actively evaluating T:0 for the moment client demand justifies deeper adoption. We are a Certified Puzzle Advisor, and our platform-agnostic advisory approach means our recommendations are based on your specific business needs rather than platform advocacy.

What you get:

• A 30-minute call to discuss the platform decision for your specific business

• Objective evaluation across Puzzle, T:0, QuickBooks Online, and Xero based on your situation

• Honest platform recommendation with clear rationale, not a sales pitch

• Migration support if switching makes sense (from QBO to Puzzle, QBO to T:0, or otherwise)

• Ongoing Puzzle-based bookkeeping if that fits your business

• QuickBooks Online expertise if that remains the right platform for you

• Xero expertise if international operations point that way

• Ongoing monitoring of both Puzzle and T:0 developments as the market matures

• Certified Puzzle Advisor perspective combined with broader accounting expertise

• No pressure to choose the platform I have certification in if it is not the right answer for you

Book a free consultation and we will help you make the platform decision honestly.

Frequently Asked Questions

What is the main difference between T:0 and Puzzle?

T:0 packages the platform and a CPA bookkeeper together in a single monthly bill starting at $599/mo, while Puzzle sells the platform alone starting at $30/mo and requires engaging a Certified Puzzle Advisor separately for bookkeeping services. T:0 is 10 days old at time of writing; Puzzle has three years of platform maturity.

Which is cheaper, T:0 or Puzzle?

At the platform-only tier, Puzzle Complete at $120/mo is 40% cheaper than T:0 Platform at $199/mo. At the platform-plus-CPA tier, T:0 Managed Bookkeeping at $599/mo is dramatically cheaper than Puzzle Complete plus a Certified Advisor at typically $1,620-$2,620/mo total.

Does the Accrual acquisition of Puzzle affect direct users?

No immediate operational change for direct Puzzle users. Puzzle continues operating independently for startups and small businesses. Accrual acquired the accounting-firm-focused technology and business. Sasha Orloff's departure to Accrual creates uncertainty about long-term product direction but does not affect the platform functioning today.

Is T:0 owned by Airwallex?

Yes. T:0 is a subsidiary of Airwallex, the cross-border payments fintech valued at $11B after their $320M Series H in June 2026. This corporate structure matters because your books and your payments processor are the same company under T:0, which raises independence questions worth considering.

Which platform is better for e-commerce?

Neither, as of 11 September 2026. Both are built for VC-backed SaaS startups. Neither publicly names Shopify, Amazon, or e-commerce-specific integrations. For e-commerce operations, QuickBooks Online plus specialised tooling (A2X, Link My Books) remains the practical choice.

Can I migrate from Puzzle to T:0 or vice versa?

Migration is possible but not straightforward. Neither platform has published a specific Puzzle-to-T:0 or T:0-to-Puzzle migration path because T:0 is 10 days old and the use case did not exist before. Expect migration to require exporting from the source platform and importing through the destination's standard onboarding process, with more friction than a QBO-to-either migration.

Should existing Puzzle users switch to T:0?

Not based on the launch news alone. Puzzle continues operating for startups today as it did before both announcements. Making platform decisions based on a 10-day-old alternative rather than current problems typically produces worse outcomes than staying with a working platform. Watch T:0's trajectory over the next 3-6 months before re-evaluating.