Legal work combines high volume with high consequence

OpenAI has published a customer profile describing how Stockholm-based Legora uses OpenAI models and coding agents in what the company calls an agentic operating system for legal work. The platform supports document review, legal research and other workflows where large volumes of material compete with limited professional time. OpenAI has been a model provider since Legora’s early development.

Legora also uses ChatGPT Work and Codex internally. Its engineering team says coding agents let developers investigate many possible solutions in parallel before committing to one design. For product incidents, agents can search millions of log lines and pursue several hypotheses at once, sometimes reducing the path from report to working fix to around five minutes.

Coding agents broaden Legora’s product experiments

The company’s legal team built a playbook from previously signed non-disclosure agreements. The system takes a first pass at new NDAs and highlights clauses outside approved parameters. Legora reports that this reduced lawyer time spent on NDA review by more than 95 per cent, leaving professionals to focus on unusual terms and decisions that require judgement.

Across its customers, Legora cites a 90 per cent reduction in contract turnaround for one Fortune 100 organisation and an average 9.5 hours returned per lawyer each week. These figures are company-reported outcomes from specific deployments. They do not establish that every legal team will obtain the same result, particularly when contract types, playbooks and risk tolerances differ.

Reported time savings still need quality controls

Legal automation needs more than a fluent model. The playbook must be versioned, exceptions need clear ownership and every suggested edit should be traceable to source text and an approved rule. Confidential documents require strict workspace, connector and retention controls. A system should also abstain when language is ambiguous rather than forcing a clause into the nearest known category.

The profile shows why domain workflow design is as important as model selection. Legora combines OpenAI capabilities with legal context, product controls and professional review. Buyers should evaluate complete matters, not isolated prompts, measuring missed issues, unnecessary escalations, review time and whether the final work remains defensible to clients, regulators and courts.

What teams should validate before production

For implementation, teams should record the exact model, software, dataset, connector and policy versions behind every result. Representative evaluation must include normal work, boundary cases, deliberate misuse and component failure, with outcomes separated by user role, data sensitivity and consequence. A strong average should never hide a serious failure in a critical path. Changes should pass the same regression set before promotion, and operators need a documented fallback that preserves evidence rather than silently broadening access or lowering quality.

Procurement and planning should account for the human work required to review exceptions, maintain controls and investigate incidents. Published performance or efficiency claims are useful starting points, but production confidence comes from evidence collected in the organisation’s own environment. A named owner, measurable acceptance threshold, cost model and tested shutdown path turn a promising release into a service that can be governed over time. Where decisions affect safety, rights, money or regulated records, accountable professional review remains essential.

Security and operational evidence

Security review should map every identity, credential, network destination, stored artefact and administrative access path involved in the workflow. Logs need enough context to reconstruct who initiated an action, which evidence was consulted, what the system proposed and what ultimately executed. Sensitive prompts and traces should not be retained merely because they are useful for debugging. Retention, redaction and access must follow the underlying data classification, with a practical process for correcting or removing material when policy or law requires it.

Operational monitoring should measure quality and service behaviour together. Latency, throughput and availability matter, but so do unsupported claims, missed constraints, unsafe tool calls and the rate at which people overturn recommendations. Teams should rehearse dependency outages and partial failures before unattended operation. If a model, connector or external service becomes unavailable, the workflow should fail predictably, preserve its state and tell the user what remains unfinished. Clear degradation is safer than an agent improvising around a missing control.

Transparency should scale with adoption

Organisations should explain the scope and limitations of the deployed capability to the people affected by it. Users need to know when automation is involved, how to challenge an outcome and where a human can intervene. That transparency improves feedback and prevents a pilot result from being interpreted as blanket approval for broader autonomy. Expansion should follow demonstrated evidence, not novelty or pressure to match a competitor announcement. The durable advantage comes from dependable integration, careful measurement and learning that survives the next model or product update.

Finally, teams should schedule a review after real usage rather than treating launch as the end of governance. The review should compare expected benefits with observed quality, cost, incidents and user behaviour, then retire workflows that do not justify their complexity. A service that works well can expand through the same evidence-based process. A service that fails should be narrowed or stopped without losing the records needed to understand why. This feedback loop is what turns experimentation into accountable operational capability.