A working scorecard for finance leaders contemplating taking a multi-entity group off Xero onto an AI-native ledger, built from what four vendors actually disclose, not what their marketing pages claim. Scored across the eight dimensions that decide the move: entity depth, statutory compliance, stack scope, operational reach, sector fit, agent governance, indicative cost, and the migration path itself.
The cost of an AI accounting system sits above a single Xero subscription — which flatters Xero in any sticker-price comparison. However, run the real cost, current-state vs. future-state, with your own numbers.
Try The Shadow Ledger →Plenty has been said lately about Xero raising prices, pushing through format and functionality changes not everyone wanted, and leaning on the very accounting partners who built its ecosystem. The business rocketed out of New Zealand in 2006 as cloud accounting's global pathfinder, when the industry was collectively at T-minus-3 on the launch pad. Xero has since gone stratospheric, but maybe it's now orbiting a little too close to the sun.
I'm not here to add to that pile-on.
Agency DNA was founded in late 2007 on one idea: to be the source of knowledge on accounting software for creative and professional-services agencies. We looked hard at Xero from day one, were genuinely impressed, and adopted it ourselves immediately. We are certified experts.
But Xero was never built for everyone, and it has real limits, particularly for multi-entity businesses like much of our client base, who need fast-turnaround job (or project) estimates, complex client billing needs, work-in-progress tracking, and intercompany automation across markets. Most of what we do is to find the right next platform for clients who've already outgrown Xero, or QuickBooks before it. That's work we've now spent the better part of two decades doing, assessing the functional, commercial, and cultural fit of a dozen-plus ERP and planning platforms.
Now AI is lighting a whole new generation of accounting software on the launch pad, all somewhere between T-minus-4 and T-minus-1 by our reckoning, and we've looked closely at them. What follows is our first round of assessment.
Search for "Xero alternative for multi-entity" and almost everything you find is a vendor's own comparison page, an affiliate blog paid per click, or a listicle that treats "AI-powered" as a feature rather than a claim to verify. None of it is written from the seat that actually owns the decision: a finance director who has to sign off on a general ledger a machine can post to, across statutory entities in more than one country, and live with the audit trail for the next seven years.
This framework scores Light, DualEntry, Rillet, and Campfire, four of the leading AI-native platforms positioned intentionally or otherwise to take a growing, multi-entity business off Xero, against eight dimensions that decide whether that move works in practice. Ratings are drawn from each vendor's own product pages, funding announcements, and independent trade coverage as of September 2026, read the way a practitioner reads them: for what is actually disclosed, not what's implied.
Xero and QuickBooks run one company per file. The moment a group needs consolidated numbers across more than one legal entity, these are the questions that separate a real platform migration from a more expensive spreadsheet problem.
| Dimension | Light | DualEntry | Rillet | Campfire |
|---|---|---|---|---|
| Multi-entity depth | Native multi-book GL; "ten entities like one company." | Unlimited entities; real-time consolidation, AI-assisted eliminations. | Automated intercompany eliminations; consolidated & entity-view reporting. | Unlimited entities; single sign-on consolidation across subsidiaries. |
| Geographic & statutory coverage | 20+ countries live, Europe-native; built for statutory complexity, not bolted on. | 240+ countries, 180+ currencies claimed; VAT/GST automated. Depth per jurisdiction unverified. | Published native VAT/GST reporting for seven countries (UK, Australia, Canada, South Africa, Switzerland, Lithuania, India). | 180+ currencies, "global compliance" claimed; no published jurisdiction list. |
| Whole-stack vs. GL-only | GL + procure-to-pay (procurement, AP, bill pay), cards, expenses, and subscription management natively. | GL, AR/AP, cash, tax, plus purchase-order/procurement (three-way matching). No inventory module. | GL, AR/AP, bank rec, revenue recognition, close. Now marketed as an "AI-native ERP," though functional scope has no procurement, cards, or bill pay. | GL, revenue recognition, reporting, close. No inventory. |
| Operational scope | Apps marketplace (FP&A, payroll, timesheet, mileage); not an MRP/inventory system. | Live procurement/purchase-order module (three-way matching, approval workflows); no inventory module. | No inventory or procurement modules disclosed. | No inventory or procurement modules disclosed. |
| Sector fit | AI/SaaS, manufacturing, services, marketplaces, ecommerce — six-plus named verticals. | SaaS, family offices, PE-driven roll-ups with lean finance teams. | Subscription/SaaS businesses; deep revenue-recognition tooling; PE portfolios, accounting firms. | Growth-stage software/SaaS/fintech (Replit, PostHog, Klarity); some healthcare. |
| Governance on agent-posted entries | SOC 1/2 Type II, GDPR, stated audit trails; "agents work inside your controls." Approval-threshold mechanics not independently verified. | SOC 2, GAAP/IFRS compliant. No public detail on approval gates for autonomous postings. | Dedicated security page states "0 auto-posts without approval" and "100% of AI actions logged." Reinforced by an EY partnership built around the same controls. | 1,200+ granular permissions; named approval workflows; SOC 1/2, SOX-ready trails. Public pages now confirm an approval gate for agent-posted accruals specifically; full autonomy-rate disclosure still not published. |
| Migration path off Xero | Migration mechanics described (balances and open items preserved, live in weeks); no dedicated named Xero campaign found. | Dedicated Xero page promises 24-hour CPA-led migration, implementation always included. Separate $6M "Rescue Fund" covers ERP contract buyouts generally, but doesn't name Xero specifically. | Published "Best Xero Alternative" content targeting Xero users at ~$15M ARR/10+ entities; no formal migration program disclosed. | "Implementation and Migration" offered as a general service, not Xero-specific. |
| Indicative annual cost | ~£26K/yr starting; £26K–£111K total project cost. Single-source estimate (ERP Research, a disclosed paid Light sponsor), no independent corroboration found. | ~£18K–£59K/yr, tiered by entity count (3/20/unlimited); implementation included. Single-source, quote-based estimate (ERP Scorecard). | ~£12K–£25K/yr (Vendr median ~£21K). Best-evidenced of the four — real buyer-transaction data, cross-confirmed by two independent trackers. | ~£15K–£30K/yr — ERP Scorecard's own estimate, explicitly "peer-anchored" to Rillet's Vendr data, not independently observed for Campfire. |
Every one of those figures sits above a single Xero subscription, which makes the sticker-price comparison misleading. A multi-entity group isn't paying for one Xero subscription; it's paying for several, plus a consolidation add-on, plus every hour a finance team spends stitching the group numbers together by hand. The Shadow Ledger is the companion calculator that runs that fuller comparison against your own numbers.
The agentic ERP built multi-entity-first — European statutory depth is the real differentiator, though its Xero-migration claim doesn't hold up to scrutiny (see Watch for).
The volume play — unlimited entities, 240+ countries, 180+ currencies claimed, built for groups adding subsidiaries faster than they can add finance headcount.
The revenue-recognition specialist — the deepest deferred-revenue and waterfall automation of the four, now valued at $1B two years after emerging from stealth.
The governance leader — 1,200-plus granular permissions and named approval workflows make it the only one of the four to disclose real control mechanics rather than a compliance-badge list.
These sit next to the category rather than in it: overlays on an existing ledger, single-entity plays, or ambitions not yet backed by customer depth. Worth tracking, not yet worth scoring against the seven dimensions above.
| Name | What it actually is |
|---|---|
| Doss | Operations layer: inventory, procurement, orders that partners with Rillet and Campfire rather than replacing the ledger. Relevant once a product or inventory business needs to pair an AI-native GL with real operations software. |
| Numeric | Close-management automation laid over an existing ERP. A workflow overlay, not a ledger replacement. This doesn't compete for the Xero-exit decision itself. |
| Puzzle | AI-native ledger aimed at the startup and QuickBooks segment. Single-entity focus; not a multi-entity contender at this stage. |
| Truewind | AI close automation layered on top of Sage Intacct rather than a standalone platform. |
| Everest | States a "full business OS" ambition, built by a former SAP team. No proven customer depth in public material as of this writing. |
Every one of the four uses the language of governance: audit trails, SOC certifications, permissions. None of the four public product pages fully answers what actually happens the moment an agent decides to post. That gap is exactly where the vendor conversation needs to start, and it's the qualifying line Agency DNA runs in every one of these conversations.
Can an agent post to the general ledger without a human touching it, and if so, at what value or entry-type threshold does it stop and wait?
Is every agent action attributable to a named, timestamped audit-log entry that's reversible and visible to your external auditor without a vendor support ticket?
Who is the documented human owner of record when an agent's posting is wrong, and what's the process for correcting it once it's already fed a reported number?
This page is free, and stays free. We believe it's the fastest way to get the state of this category right. Two things go further than a scorecard can.
Everything on this page, plus the evidence behind every rating, real vendor-published migration stories alongside independently verified ones as they surface, and a fresh pass each quarter as the category moves. We're building this for anyone who has to defend the recommendation in a board pack.
I run your actual entity count, close hours, and tooling spend through the Shadow Ledger model myself, score the move against Agency DNA's Automation Opportunity methodology and hand you a written recommendation. There isn't a spreadsheet you have to interpret alone. This one carries a modest one-off fee, confirmed with you before anything's committed.
Leave your email and which one you're after. I'll reply personally within one business day. The report's free; a Shadow Ledger reading carries a modest one-off fee, confirmed when we speak.
Thanks, that's in. I'll be in touch within one business day.
Something went wrong on that submission, please email chris.lever@agencydna.com directly instead.