Internal Rules

The company’s General Counsel conducted an internal audit and uncovered inefficiencies in how the legal team handled incoming contracts — particularly those received from external parties. Although the team followed well-defined internal policies, the manual nature of the review process led to delays, inconsistencies, and repeated re-checks.

While the problem was clear, it wasn’t obvious whether automation could handle such nuanced, company-specific rules — especially without compromising security or control.

During our discovery phase, we mapped out the team’s workflow in detail and proposed an AI-powered solution that mirrors their internal compliance logic and runs fully within their secure infrastructure (on premise). In this case, employees interact with the system via Google Chat: they simply send a contract, and the bot returns comments in the document along with a summary in chat.

The solution is channel-agnostic and can be integrated into any environment — Slack, Microsoft Teams, internal portals, email, or CRM systems. Google Chat was chosen here for convenience, but the architecture supports full flexibility.

The main challenge wasn’t just technical — it was structural. The legal team was already working under detailed internal rules and procedures, but applying those manually across a high volume of non-standardized, externally sourced contracts created bottlenecks, inconsistency, and slow contract turnaround time.

Manual review was especially time-consuming in fast-paced business scenarios, and although the General Counsel had a clear understanding of the risks and inefficiencies, it wasn’t obvious whether automation could actually handle the complexity of internal policies — or if such logic could even be reliably translated into AI.

We needed to ensure the output was not only accurate, but also practical for end users — helping legal reviewers save time, reduce rework, and trust the results. The solution also had to be secure, easy to maintain, and fully embedded in the team’s everyday tools.

The full implementation took just under a month:

– 2 weeks: digitizing internal contract policies, training a private AI model, and integrating the bot with Google Chat
– 1 week: testing the system with real-life use cases
– 1 week: fine-tuning the model based on legal team feedback

The bot returns a summary of all major issues directly in the chat, and also inserts comments directly into the original contract document. This makes it significantly easier for legal professionals to review — everything is highlighted in context, with no need for cross-referencing outputs.

The entire solution runs on-premise, requiring no external APIs or cloud services, which ensures full compliance with internal security standards.