Beyond the Queue: How Leah’s Agentic Harness is Rewriting Enterprise CLM

LONDON — September 17, 2026 — In the modern enterprise, resilience is rarely born from working harder; it is engineered by working differently. For those of us who study the mechanics of consistent value creation, the enterprise software stack has long presented a glaring inefficiency. We have spent the last two decades investing in systems of record that are, fundamentally, just elaborate digital queues. Human beings are still required to manually pull tasks from these queues, process them, and push them to the next stage.

Today, that paradigm officially fractures. Leah, the London-founded enterprise agentic AI company previously known as ContractPodAi, has launched Leah Contracting. Powered by its proprietary orchestration engine, Leah Maestro, the platform does not merely assist human lawyers in managing contracts—it autonomously executes the contract lifecycle. Backed by over $170 million from SoftBank and Insight Partners, the company’s latest move signals a structural transition in legal technology: the definitive end of the traditional Contract Lifecycle Management (CLM) era.

The Death of the Queue

To understand the magnitude of this shift, one must look at the legacy architecture of enterprise CLM. Incumbents like Ironclad, Icertis, and Conga built deep moats by digitizing the filing cabinet and the routing slip. When a commercial counterparty redlines a Master Services Agreement (MSA), traditional CLM software alerts a human lawyer, places the document into an inbox, and waits. AI features added in recent years—such as generative drafting or clause extraction—made the human operator faster, but the underlying requirement for human propulsion remained.

Leah Contracting fundamentally rewrites this workflow by transitioning from a queue that people work through to a system that the business runs on. Utilizing specialist autonomous agents, the platform handles intake, reviews third-party paper, benchmarks redlines against corporate playbooks, negotiates fallback clauses within approved tolerances, and extracts post-signature obligations.

“CLM was built for a world where people operate software. We helped build that category, and it has reached its limit,” said Sarvarth Misra, Co-Founder and CEO of Leah. “Putting AI on top made people faster at operating it. It did not make the software do the work. Leah Contracting does.”

This is not merely a theoretical upgrade. Traditional CLM implementations have increasingly suffered from "shelfware fatigue," where complex platforms are abandoned because they require too much manual data entry and continuous human intervention. By delegating the execution layer to autonomous agents, Leah removes the friction that has historically bottlenecked corporate legal departments.

“During my time at Comerica, I saw first-hand where traditional CLM delivered value and where it reached its limits,” said Adil Karachiwala, SVP of Commercial Strategy, P2P at Leah. “The next leap in contracting was never going to come from more workflow or another layer of tooling, but from changing how the work itself gets done. That is the shift Leah Contracting represents.”

The Harness: Solving the Autonomy vs. Governance Dilemma

The primary barrier to agentic AI in the enterprise has never been intelligence; it has been liability. Large Language Models (LLMs) are inherently probabilistic. They guess the next most likely word. Corporate contracting, however, is strictly deterministic. A hallucinated indemnification cap or an unauthorized governing law clause can expose an enterprise to catastrophic financial risk.

This tension is where Leah Maestro, the company’s "super orchestrator," proves its value. Maestro acts as an agentic harness—a deterministic control plane that bounds the probabilistic nature of the AI.

“Agentic OS is rapidly becoming one of the most overused phrases in enterprise software,” Misra said. “A model, a loop and some tool calls are not an operating system, and they are not enough to run something as consequential as contracting. An agent can do a task. A harness is what gets agents working together to run the process reliably.”

Maestro treats AI agents as bounded execution workers. It enforces deterministic state transitions, meaning an agent cannot advance an agreement to signature if compliance tests fail. It utilizes short-term memory for specific negotiation turns and long-term memory drawn from over 750 pre-built industry templates and the customer's historical precedents. Crucially, the harness enforces strict human-in-the-loop approval gates for irreversible actions or deviations that exceed predefined fallback limits.

As one Chief Information Security Officer at a global financial institution noted during a recent evaluation of agentic platforms, raw model intelligence has commoditized, and the true enterprise value now lies entirely in the orchestration and guardrail layer. Without an auditable harness that provides cryptographic telemetry and immutable audit logs, heavily regulated corporate legal departments simply will not accept systemic model autonomy.

“The model brings intelligence. The harness brings control. Domain depth brings judgment,” Misra added. “For enterprise contracting, you need all three.”

Pricing the Outcome, Not the Token

Beyond its technical architecture, Leah is challenging the foundational economics of enterprise B2B software. The industry is currently caught between two flawed monetization models: per-seat SaaS licensing and consumption-based token billing.

Per-seat licensing disincentivizes enterprise-wide adoption, as companies only buy licenses for their core legal team, leaving sales and procurement disconnected. Conversely, token-based pricing—where companies pay for the compute required to run AI models—introduces erratic, unbudgetable operating expenses. Chief Financial Officers cannot predict whether a complex, multi-turn negotiation on a cross-border vendor agreement will cost two dollars or two hundred dollars in LLM reasoning loops.

Leah is abandoning both in favor of outcome-based pricing: paying for completed workloads. Instead of billing for the software seats or the AI compute, the company prices against the finished business outcome, such as a fully processed Non-Disclosure Agreement or an audited supplier MSA.

This shift reflects a broader macroeconomic trend. As one enterprise software procurement analyst recently observed, when software transitions from a tool that assists labor to a system that replaces labor, procurement departments demand that it be priced like operational productivity rather than an IT subscription. To defend its margins under this model, Leah utilizes highly efficient small language models (SLMs) for routine extraction tasks, reserving expensive frontier models—accessed through alliances with Microsoft, Google, OpenAI, and Anthropic—only for high-stakes negotiation analysis.

The New Baseline for Enterprise Execution

The true mark of a resilient enterprise system is its ability to break down departmental silos. Contracts do not exist in a vacuum; they are the financial and operational lifeblood of a company. A contract negotiated by legal contains payment milestones that finance must recognize and vendor service-level agreements that procurement must track.

By running Leah Contracting on the Maestro harness alongside Leah Legal, Leah Procurement, and Leah Finance, the platform allows work to move across functions without stopping at the boundary of individual systems. An obligation agreed upon in a contract is automatically propagated to the procurement module for compliance tracking, surfacing potential exposures before they become liabilities.

Furthermore, the company has addressed the paramount concern of enterprise data sovereignty. Customers can deploy Leah Maestro as a single tenant on any cloud, in any region, or on a virtual machine inside their own corporate perimeter. The company guarantees that customer data, proprietary negotiation positions, and playbooks are never used to train third-party foundation models or Leah’s own proprietary systems.

In a global landscape defined by unpredictability and margin compression, the businesses that win will be those that systematically eliminate operational friction. Passive software that merely tracks human effort is no longer sufficient. By introducing a secure, deterministic harness that allows autonomous agents to execute complex, high-stakes work, Leah is not just upgrading Contract Lifecycle Management. It is establishing a new baseline for how enterprise value is consistently and securely created.