Competitor Profile
Where this is heading
PayFit sells payroll and light HR administration to small and mid-sized businesses in France, the United Kingdom and Spain, with its own statutory engine in each rather than a partner's. As of July 2026 it does not describe itself as a software company. It describes itself as a service: 22,000 customers, roughly 700 staff of whom 200 are payroll experts, EUR 80M ARR, profitable, and a product that now acts on payroll data rather than only displaying it.
The read: the headline is the AI, but the structural fact is the org chart. PayFit has put 200 human payroll specialists inside the commercial offer and named the combination of software, humans and AI a single service. That is a gross-margin decision as much as a product one, and it is the direction a payroll vendor takes when the last mile of accuracy cannot be closed in code. It also moves PayFit onto the bureau's ground, where the competitive question stops being feature parity and becomes cost-to-serve per payslip.
The UK gap here is hard evidence rather than inference. PayFit is HMRC-recognised for RTI and PAYE and appears by name in the paid-for table of HMRC's own recognised-software list, but it does not appear on HMRC's separate Construction Industry Scheme supplier list. Any UK small business paying construction subcontractors cannot run that payroll here, and the buyer can verify it against the government register in a minute.
On AI it is further along than most vendors its size, and further than its own public materials from six months ago imply. PayFit AI is free in every tier rather than sold as a surcharge, it takes actions rather than only answering, and 200 payroll experts sit behind it as a live escalation path inside the same conversation. What it does not have is a first-party agent-callable surface: the Open API is a conventional REST integration API on Standard and Premium plans, and the only MCP route into PayFit data is a connector built and operated by a third-party integration vendor.
01
02
How the product hangs together, at the level a product leader needs to place it.
PayFit is a Paris-headquartered payroll and HR platform for small and mid-sized businesses, founded in 2015 by Firmin Zocchetto, Ghislain de Fontenay and Florian Fournier, and operating in France, the United Kingdom and Spain. Its UK entity, PayFit Ltd, is registered at Companies House under number 11623900 and incorporated in October 2018, with a London office. The core proposition is that a business runs its own payroll in-product rather than sending it to an accountant or a bureau, and that the statutory engine underneath is PayFit's own in each country rather than a licensed third-party calculator.
The payroll surface covers salary calculation, deductions, payslip generation and statutory filing, with country-specific depth: RTI and PAYE submissions to HMRC in the UK; in France, automatic generation and transmission of the Déclaration Sociale Nominative to net-entreprises once payroll is closed; and in Spain, automatic submission of contribution files to the Social Security TGSS RED system alongside tax forms to the AEAT. PayFit's UK materials also claim automatic pension submissions to all major providers for auto-enrolment, statutory sick pay handling, P60 and P45 generation, and BACS and Faster Payments support, and put a figure of 90% on the share of manual payroll tasks the product automates.
Around payroll sits a light HR layer rather than a full HCM: leave and absence management, expense management with receipt OCR, an employee self-service portal, performance and people management, and an integrations directory. Two additions extend the surface beyond software. A November 2025 integration with the Irish benefits platform Kota puts private medical insurance enrolment, deduction and tax reconciliation inside the UK payroll run, which PayFit describes as making it the first dedicated UK payroll platform with health benefits built in. A September 2025 launch added an interactive payslip that explains individual pay lines to the employee in-product, released alongside a PayFit-commissioned Ipsos payslip-literacy barometer surveying 1,950 workers across its three markets. That barometer is the commercial argument for the employee-facing product line, and UK trade press carried its findings: 73% of UK workers say they fully understand their payslip against 54% in Spain and 29% in France, but 24% struggle with tax deductions, 19% with pension contributions, and 28% of under-35s contact HR or payroll every month with payslip queries against a 13% UK average. PayFit is pointing its product at that monthly query volume.
For integrations beyond the directory, PayFit publishes an Open API: a REST interface authenticated by API key, generated from the admin console's API access tab, with scopes controlling what each key can read. Documentation sits at developers.payfit.io and states that further endpoints are still being added. Access is not universal across the customer base; PayFit's own documentation places the Open API on the Standard and Premium plans.
The customer base is small-business-weighted, spanning businesses from their first employee up to several hundred, and PayFit's UK site puts the count at 22,000+ small and medium businesses. It sells direct to HR generalists, office managers and owners rather than exclusively through an accountant channel. Ownership is venture-backed and the capital position is settled: the most recent publicly confirmed valuation is EUR 1.82B (USD 2.1B), set by a EUR 254M Series E led by General Atlantic with Eurazeo, Bpifrance and Accel participating, and no subsequent round, secondary or valuation revision is evidenced. Profitability therefore arrives without fresh capital behind it, which is the context for the EUR 150M ARR target.
03
Geographic footprint, and where payroll is native rather than partner-delivered.
United Kingdom: NATIVE for RTI and PAYE, confirmed against HMRC's own register rather than vendor marketing. PayFit Ltd appears by name in the paid-for software table of HMRC's recognised payroll software list, with the product named "PayFit UK". PayFit's UK site states that RTI submissions go to HMRC automatically, that pension submissions run to all major providers for auto-enrolment, and that statutory sick pay, P60 and P45 are handled in-product, with BACS and Faster Payments for the payment leg. Construction Industry Scheme support does not exist: PayFit is absent from HMRC's separate CIS commercial software suppliers list, and no PayFit product page, help-centre article or third-party CIS comparison names it as CIS-capable. Contractor and subcontractor payroll is not evidenced as a PayFit capability at all, which makes the UK footprint narrower than "native UK payroll" alone conveys.
EU/EMEA: NATIVE in France and Spain. France is the founding and dominant market, with the statutory engine generating and transmitting the DSN to net-entreprises on payroll close, and accounts for roughly 85% of revenue; the United Kingdom and Spain together make up the remaining 15%. In Spain the product carries RED authorisation and submits contribution files to the Social Security TGSS alongside tax declarations to the AEAT. Germany and Italy do not appear in PayFit's current core-market materials, which name only France, Spain and the UK, and the German office was closed in the 2023 restructuring. A December 2025 technology partnership with the French workforce-scheduling vendor Skello pushes rota and time data into PayFit as payroll variables automatically, a France-first integration aimed at shift-heavy sectors and claimed to save roughly 15 hours a month across combined scheduling and payroll administration.
North America: NOT SERVED. No US or Canadian entity, office, customer or statutory payroll capability is evidenced. PayFit's operations are confirmed concentrated in France, the UK and Spain. This is scope PayFit has not attempted rather than a gap it is closing.
South Africa: NOT SERVED. No South African presence, native filing capability or partner arrangement is evidenced. The product appears in South African software directories as a searchable listing, not as an operating local payroll provider.
04
The buyer is an HR generalist, office manager or the business owner directly, at a company without a dedicated payroll specialist, and PayFit's stated range runs from a business's first employee up to several hundred. The pitch is that payroll stops needing an expert, and since July 2026 that pitch has a second half: where the software cannot close a question, a PayFit payroll expert does, inside the same product. PayFit frames the combination of platform, 200 payroll specialists and PayFit AI as one end-to-end service rather than a licence with support attached, and its stated ambition is to become the single service through which a manager handles everything about their people.
That ambition has a declared scope extension attached. PayFit states it will widen the service by the end of 2026 into health insurance, IT equipment and freelancer management, areas that sit outside payroll and outside the HR admin layer the product covers today. The Kota health-benefits integration is the first of those three already in market in the UK.
Pricing is not published. PayFit's pricing page carries no tiers, rates or base fees, and routes every visitor to a demo booking or a trial sign-up, so no rate card can be sourced from the vendor directly. What can be sourced is packaging structure: the Open API is limited to the Standard and Premium plans, while PayFit AI is stated to be included in all packages at no extra cost.
That AI packaging decision is the commercially interesting one. Pricing the assistant at zero across every tier removes AI from the upsell conversation entirely and converts it into a retention and cost-to-serve instrument, which is consistent with a vendor whose margin story now depends on human specialists it has to pay for.
05
PayFit AI is a shipped, in-product payroll and HR agent, and the distinction between an assistant and an agent matters here because PayFit crossed it. It began as a question-answering tool for HR teams in 2024 and was opened to employees in January 2026, reaching 220,000 employees with more than 500,000 conversations logged since launch. From July 2026 it also acts: PayFit states it detects payroll inconsistencies, generates HR reports on demand, answers employee questions and enters payroll variables directly. PayFit's product page describes it handling reports, variable pay and leave requests, flagging anomalies before payroll closes, and letting employees file holiday requests and check payslips directly, with the user retaining a validation step. PayFit states a further extension by the end of 2026 covering sick leave, contracts and attestations.
The grounding is the part competitors should read closely. Answers are contextualised to the individual employee's contract type, seniority, leave balance, payslip and applicable collective agreement, drawing on more than 1,000 help-centre articles written and maintained by PayFit's own payroll specialists, and administrators can upload company documents so responses reflect internal policy. This is a curated first-party knowledge base maintained by the same people who run the payroll, not a generic model pointed at a product manual.
The escalation design is the structural signal. When a situation requires human judgement, PayFit states a payroll expert joins the same conversation with full context already loaded. That is the seam where the AI product and the 200-person expert bench meet, and it is what "AI-augmented service" means in practice: PayFit is adding both software and humans, with the AI deployed to raise the number of payslips each specialist can carry rather than to remove the specialist.
Published adoption and satisfaction figures are vendor-stated and worth treating as such: 65% of customers using it monthly, 82% reporting tasks are easier, 72% reporting daily time saved, and 81% saying they trust its answers. The January 2026 launch material carried a figure worth reading against the deflection story, that 77% of HR teams reported internal request volume still rising. No accuracy or error-rate figure is published for any of the acting capabilities, and PayFit's own UK compliance material states plainly that employers remain responsible for RTI errors, which places the liability for an agent-entered payroll variable on the customer.
No underlying model or AI provider is disclosed anywhere in PayFit's public materials. On agent-callable surfaces, no first-party MCP server is evidenced. The Open API is a conventional REST integration interface on Standard and Premium plans, authenticated by API key with scopes, and its documentation makes no mention of MCP or agent access. The only MCP route into PayFit data is a connector built and operated by StackOne, a third-party integration vendor, exposing 23 pre-built actions across collaborators, payslips, contracts, absences, worked time, payroll status and accounting data, of which 19 are reads and four are writes. There is no evidence PayFit co-builds or endorses it. PayFit's own marketing claims PayFit AI is "the first AI agent for payroll and HR", a claim about the in-product assistant rather than about any external protocol surface.
06
Diagnosis and the play together. Each is a live attack vector, not a general weakness.
UK Construction Industry Scheme gap. PayFit is listed by name on HMRC's recognised payroll software register for RTI and PAYE and absent from HMRC's separate CIS commercial software suppliers list, with no CIS content anywhere in its product or help materials. Any UK construction or trades SMB using PayFit has no CIS path in-product and must run a second tool or a bureau relationship alongside it. This is a register-verifiable gap rather than an inference, which makes it usable in a live evaluation: the prospect can check it themselves in a minute against the government list.
Concentration risk in a three-country footprint. Roughly 85% of revenue comes from France, with the UK and Spain sharing the remaining 15%, and Germany and Italy are absent from PayFit's current core-market materials. A UK buyer is therefore buying from a vendor whose product investment priorities are set by a market they are not in, and any multi-country SMB with employees outside those three countries has no PayFit answer at all. Vendors with committed local statutory depth across a wider EMEA footprint can contest both the UK single-country deal and any multi-country expansion case.
The services model is unpriced and untested at scale. PayFit has put 200 payroll experts inside the commercial offer and simultaneously priced its AI at zero across every tier, having reached profitability on EUR 80M ARR while targeting EUR 150M by 2028. Nearly doubling revenue while carrying a human expert bench is a gross-margin problem, and the obvious resolutions are raising price, tightening what the service actually includes, or leaning harder on automation than accuracy allows. Competitors selling to a CFO can legitimately ask what the service level is contractually, what is included versus billable, and what happens to it at 2028 revenue.
No first-party agent-callable surface. The Open API is a conventional REST interface restricted to Standard and Premium plans, with no MCP or agent access documented, and the only MCP route into PayFit data belongs to a third-party integration vendor rather than to PayFit. Buyers building their own assistant estate, or standardising on an agent layer above their HR stack, cannot reach PayFit data on the vendor's own terms. Any competitor shipping a first-party, scoped, auditable agent surface has a clean differentiation line, and the audit and permissioning question is the one that lands hardest with a payroll buyer.
Pricing opacity in a self-serve segment. PayFit publishes no rates on its pricing page and gates every tier behind a demo booking or trial sign-up, in a segment where the buyer is an owner or office manager comparison-shopping online before ever speaking to a salesperson. Vendors with transparent published per-employee pricing win the evaluations PayFit never enters, and the packaging split, where the Open API sits behind higher tiers while AI is free everywhere, is itself only discoverable after the buyer is already in a sales process.
Recent restructuring and a support-heavy service claim. PayFit cut roughly 20% of headcount in 2023, including closure of the German office, and a further approximately 14% in 2024 including 110 customer-service roles, before arriving at ~700 staff and the current profitability position. Building a service proposition on human responsiveness two years after cutting the customer-service function invites a legitimate continuity question from any prospect weighing vendor risk, and it is answerable only with contractual service levels rather than positioning.
07
What PayFit has bought, where each sits now, and what it says about direction. None of these are signals; they are standing facts about who owns what.
No acquisitions found in the five-year window.
Checked 2026-08-15. Operating cadence and status are re-checked on each profile refresh; a change of state is itself a finding, because nobody announces the day an acquisition finishes being absorbed.
08
The latest signals where PayFit is the subject. Full history on the signals page.
09
The live predictions whose evidence rests on this vendor, including the ones going against us.