Contract lifecycle management software has spent two decades doing one job: giving legal and deal teams a place to store, route, and sign documents. Leah, the London company formerly known as ContractPodAi, spent September 17 arguing that job is over. It launched Leah Contracting, a product it says does not manage contract work so much as execute it, and in doing so put a specific, testable claim in front of every revenue team still treating CLM as a workflow tool: the software layer under deal execution is about to look completely different.
From software people operate to work a system runs
Leah Contracting runs the full contracting lifecycle, intake, review, negotiation, approval, signature, obligations, and renewal, through what the company calls Leah Maestro, an agentic harness that coordinates specialist agents across each stage. Those agents apply a company’s own playbooks, assess risk, draft and redline language, and execute permitted actions without a human re-keying the outcome into a separate system.
“CLM was built for a world where people operate software,” said Sarvarth Misra, Leah’s co-founder and CEO. “Leah Contracting does” the work itself, he argued, rather than presenting a queue of tasks for a person to clear. Misra was careful to draw a line between that claim and the current glut of “agentic OS” branding across enterprise software: “Agentic OS is rapidly becoming one of the most overused phrases in enterprise software,” he said. “The model brings intelligence. The harness brings control. Domain depth brings judgment.”
Advertisement
That distinction, model versus harness versus domain depth, is doing real work in the pitch. A general-purpose reasoning model can draft a clause. It cannot, on its own, guarantee an approval gate fires before a discount past a certain threshold goes out, or that an audit trail survives a renewal six months later. Leah is selling the control layer, not the model, and pricing it that way: the company has moved to consumption-based pricing tied to completed workloads rather than metering by AI token usage, a structural bet that revenue and legal teams will pay for finished contract cycles, not compute.
Why a contracting vendor’s move belongs in a revenue conversation
CLM sits closer to the deal desk than most revenue leaders give it credit for. Every enterprise contract that stalls in redlines or approval routing is pipeline that has technically closed but has not actually landed, and RevOps teams have historically had no lever over that stage beyond escalation emails. Leah’s bet, backed by SoftBank and Insight Partners and already running across 400-plus enterprise customers, is that the fix is not a better queue but an execution layer that removes the queue entirely.
Adil Karachiwala, Leah’s SVP of Commercial Strategy for procure-to-pay, framed the shift as the end of a familiar pattern of vendors adding features to old architecture. “The next leap in contracting was never going to come from more workflow,” he said. “That is the shift Leah represents.” The company backed that with scale claims: 750-plus pre-built industry templates and deployment options spanning single-tenant cloud, any region, or fully on-premise virtual machines, aimed squarely at regulated enterprises that have resisted moving contract data into shared infrastructure.
The pattern this fits
Leah’s move is not isolated. Revenue teams have spent the past year watching the same argument play out in adjacent categories, agentic layers replacing point tools rather than sitting on top of them, and RevOps buyers have started asking pointed questions about what a vendor’s agents are actually allowed to touch before signing off, as this publication has covered in the broader push for security receipts from agentic AI vendors. It also lands inside a wider reckoning over who owns the AI layer sitting across the revenue stack, a question RevOps leaders are increasingly being asked to answer directly, rather than leaving it to whichever point vendor moves fastest.
Leah is not a startup making this claim from a standing start. The company, founded in 2015 and now operating out of London, New York, Dubai, Mumbai, Sydney, Glasgow, and Singapore, is rebuilding its own installed base rather than pitching an unproven category to net-new buyers. That geographic spread matters for the regulated-industry pitch specifically: a bank or a defense contractor evaluating an agentic contracting layer is going to ask where the data sits and under which jurisdiction’s rules, and Leah’s answer, any region, single-tenant, or fully on-premise, is aimed directly at the enterprises that have sat out CLM’s earlier cloud-only generation for exactly that reason.
What it means for the revenue operations leader
Two things are worth tracking before treating Leah’s launch as proof the category has flipped. First, the control claims, approval gates, permissions, audit trails, are exactly the features that are hardest to verify from a press release and easiest to overstate; any evaluation should ask for a live walkthrough of what happens when an agent hits a case its playbook does not cover, not a demo of the cases it does. Second, consumption-based pricing tied to completed workloads sounds friendlier than token metering, but it shifts the negotiation to defining what counts as “completed,” a definition the vendor writes first, and RevOps should get that definition in writing before signing rather than discovering it at renewal.
A practical evaluation checklist follows directly from what Leah announced. Ask any CLM vendor, incumbent or challenger, to show a contract moving through an approval gate it is not supposed to clear, not one it is. Ask what happens to an in-flight negotiation if the agentic layer goes down, whether the human fallback is a real workflow or an afterthought. And ask how obligations tracking behaves a year into a multi-year agreement, since renewal management is where most CLM deployments quietly degrade back into a shared drive with better search.
None of that erases the more basic signal: a 400-customer, decade-old CLM incumbent chose to describe its own prior category as finished rather than defend it. That is a company betting its installed base will tolerate a rebuild rather than risk a startup doing it first. Revenue and legal leaders evaluating contract tooling this quarter should treat the announcement as license to ask every CLM vendor on their shortlist, including incumbents, the same question Leah just answered about itself: what, specifically, does your system do without a person, and what is still just a better inbox.
Source: Leah
