AIToday
Large Language ModelsAI Business & IndustryOpenAI BlogPublished: Sep 4, 2026, 06:01 JST

Legora's AI agent completes financial review in minutes

Send on LINE
Legora's AI agent completes financial review in minutes

3 Key Points

  1. What happened

    Legora used GPT-6 Astra to review 41 financial documents in a single run, completing the work in minutes instead of an evening or days. The agent found all four planted errors, including a £500,000 gap in a revenue note.

  2. Why it matters

    Legora reports GPT-6 Astra improved performance by nearly 40% over the previous model on this financial-statement tie-out workflow. The agent also completed about 50 more checks, improving accuracy and completeness while keeping the human expert in charge of final decisions.

  3. What to watch

    Legora is expanding beyond legal work into audit, tax, compliance, and risk. The success of this approach hinges on how well the human-in-the-loop model scales to these new workflows.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

Legora's deployment of GPT-6 Astra tackles a specific pain point in legal and financial work: the painstaking reconciliation of financial statements. This task, previously consuming evenings or days, is now compressed into minutes, with the AI agent checking balances and flagging discrepancies. The improvement in performance, nearly 40% for this workflow, stands out against a modest 3% average gain across all tasks, suggesting that the benefit is most pronounced for tasks requiring large-scale document ingestion and complex cross-referencing.

The company's strategy is to extend its agentic operating system beyond legal work into audit, tax, compliance, and risk. The results highlight the value of a human-in-the-loop model, where the AI handles the exhaustive comparison and the human expert retains the final judgment. This approach could define how professional services firms integrate AI, potentially shifting the focus from automation to augmentation of expert work.

Looking ahead, the broader impact hinges on whether this performance holds across other document-heavy workflows and industries. The nearly 40% improvement is promising, but the modest average gain across all tasks suggests that returns vary. Firms will need to identify where AI can deliver the most value and how to retain the trust and accountability that human experts provide.

FAQ
What is financial-statement tie-out?
It's a workflow where every figure in draft accounts is checked against trial balances, a consolidation schedule, and the previous year's accounts until they agree. It normally takes an entire evening or days.
How much faster was the GPT-6 Astra agent compared to the previous model?
It improved performance by nearly 40% on the financial-statement workflow. Across all tasks in the Legora Benchmark for Agentic Reasoning, the improvement averaged about 3%.
Who remains responsible for the final decisions in this process?
The legal expert. The agent handles the exhaustive comparison, but the expert reviews each result and makes the judgment call.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleCrowdStrike launches Falcon Guardian to limit AI agent access