Checkbox

What it is

Checkbox is a no-code platform that in-house legal teams run as their front door. Requests arrive from Slack, Microsoft Teams, email, web forms, and Salesforce; an AI layer classifies each one by type, urgency, and risk; a routing rule assigns it by matter type, lawyer expertise, workload, or business unit. Everything downstream — the matter record with its linked documents, tasks and approvals, generated NDAs and letters, and the dashboard the GC takes to the board — is assembled by legal ops in a drag-and-drop builder rather than by engineers. Founded in Sydney in 2016 by Evan Wong (CEO) and James Han, now headquartered in New York, Checkbox raised a US$23M Series A on 28 January 2026 led by Touring Capital, with Peak XV, Conductive Ventures, Tidal Ventures, Five V Capital, and Evisort founder Jerry Ting participating. More than 100 enterprises run it, among them SAP, PepsiCo, Hitachi, Telstra, Woolworths, Coca-Cola Europacific Partners, Macquarie Group, and Xero.

Why it shows up in Legal Ops stacks

• It is a configuration toolkit, not a prebuilt legal app. The builder handles linear, parallel, and trigger-based flows over custom matter statuses, fields, and views, and reports through five widget types — list, bar, line, pie, metric — on top of prebuilt templates. That is why the same tenant routinely carries compliance, HR, finance, and procurement intake next to legal: the engine has no opinion about which department a request came from. It is also why go-live is a project rather than a signup.
• The AI sits on top of a 2016 workflow engine, and it matters which layer you are buying. Generative triage, the legal chatbot, and document generation were added to an automation platform that predates them by most of a decade. In practice the AI is strong at classifying and routing an arriving request and is not a substitute for judgment on its substance — which is exactly why the Ivo and Ironclad integrations exist.
• The published outcomes are about deflection, not drafting. Hitachi’s director of legal operations reports 83% of routine legal and compliance requests now partially or fully automated across a 40-country footprint; Woolworths is cited at up to an 80% reduction in matter volume. Both figures are vendor-published and neither is independently audited, but they name the metric Checkbox is actually optimised for: how much never reaches a lawyer. See legal intake and matter management for the underlying model.

Pricing reality

Checkbox publishes no price. There is no pricing page on the site — the URL returns an authorization error and the sitemap carries no entry for it — and every route ends at a booked demo. Third-party software directories list an entry point around A$5,000 per year (roughly US$3,300); that number is not vendor-confirmed, and it describes the floor of a single-team deployment, not what an SAP-scale rollout across legal, compliance, and procurement costs. Read it as evidence the product does sell below enterprise price points, not as your budget line.

What you are really buying is a scoped implementation: a first workflow, a channel, a reporting set. Since the Series A, Checkbox has run an offer to build a team’s first legal intake agent at no cost — the cheapest available way to price the configuration effort before a contract exists. Use it, and time it.

Best for

Legal ops managers and GCs in enterprises above roughly 1,000 staff, where legal is one of several service functions drowning in intake and someone on the team owns process design as an actual job title. The trigger: requests land in five channels, nobody can report demand by business unit, and the same NDA question gets answered forty times a quarter. It is also the right pick when procurement, compliance, or HR want the same front door — running four departments on one configured engine is where Checkbox beats a legal-only tool on total cost.

Versus the alternatives

Onit is the enterprise legal management incumbent — pick it when matter management and outside-counsel spend control are the actual purchase and intake is a form that feeds them. Ironclad is the pick when the contract is the object of record and request capture is a stage of the CLM rather than the product. Wordsmith is the fastest-growing entrant in this segment and the sharper choice when you want prebuilt legal behaviour and review callable over MCP instead of a canvas your team assembles. Streamline AI is the purpose-built legal-intake rival; Checkbox runs a comparison page against it, and the honest reading of that page is that Checkbox trades time-to-live for range. GC AI fits when the lawyers want a better assistant and the business-side intake problem does not exist. If none of them fit, note that Checkbox also runs a page arguing against Jira: a service-desk project in the Jira or ServiceNow instance you already pay for, plus a triage rota, fixes measurement for a fortnight of admin — and that is the correct answer if you cannot name your top three request types today.

Watch-outs

• The configuration burden is the product’s main risk. A canvas that runs any department is a canvas somebody has to design, and an under-configured Checkbox is a form builder with a dashboard. Guard: have the vendor build your two highest-volume request types during evaluation and time it; if the first workflow takes more than three weeks of your calendar, put the internal FTE into the business case before signing.
• No published price means no published floor. The quote you get for legal is not the quote you get after procurement and HR are added, and cross-department expansion is the vendor’s growth motion. Guard: ask for year one and year three pricing at double the departments and triple the request volume, in the same document, before you sign year one.
• Every efficiency figure on offer is vendor-published. The 83% and the 80% are Checkbox’s numbers about Checkbox’s customers. Guard: instrument the baseline before go-live — requests per quarter by type and channel, time to first response — and gate renewal on a deflection rate you measured.
• Strong certifications, undisclosed AI data terms. Checkbox holds SOC 2 Type II, ISO/IEC 27001:2022, ISO 27017, and ISO 27018, with audit reports available on request through trust.checkbox.ai. What the public site does not state is hosting region, data residency options, or whether customer content reaches model training. Guard: get the subprocessor list, the model provider, and a no-training commitment written into the DPA before privileged material passes through the front door — non-negotiable for EU entities under GDPR.
• No MCP server and no public API reference. Webhooks and an API exist and the vendor says the site’s integration list is not exhaustive, but there is no developer documentation site and nothing MCP-callable, so Checkbox cannot currently be a step inside agents your own team builds. Guard: if that matters, request the API reference during evaluation and confirm webhook coverage for matter-created and status-changed events rather than assuming it.