EAC Briefing: AI-Assisted Briefing Preparation for DG EAC
The Challenge
DG EAC needed a faster, more reliable way to prepare cabinet-ready briefings for high-stakes meetings and engagements. Teams were working under tight deadlines, yet the information required for each briefing was scattered across multiple systems and documents. Staff had to assemble meeting requests, policy material, previous briefings, programme figures and interlocutor information, then shape it into a concise document tailored to the occasion, the audience and the official speaking. That process was not only time-consuming, but also vulnerable to inconsistency. Valuable time was spent searching for evidence, verifying whether figures and statements were current, and repeatedly reworking drafts to fit the required structure, tone and length. Reviewers often discovered missing key messages, unsupported details or untraceable figures late in the process, creating additional rounds of clarification and editing when time was already short. DG EAC’s core challenge was to reduce the manual effort and late-stage rework involved in briefing preparation, while preserving the accuracy, traceability and governance expected in a European Commission environment. The goal was to let staff focus more on policy judgement and final clearance, rather than document assembly.
Our Solution
Nesqual Tech designed and delivered EAC Briefing, an AI-assisted application that guides DG EAC staff from an initial briefing request to a sourced draft ready for review, clearance and export in the Commission’s Word templates. The solution was built as a secure workspace that structures the drafting process around the specific occasion, the official’s role, the required sections and the requested length. Instead of forcing staff to manually compile material from disconnected sources, the application retrieves relevant policy documents and attachments, reads live Qlik chart values when figures are required, and can incorporate approved, read-only Databricks development data. It also enables staff to look up interlocutors in official sources, inspect the supporting excerpts and decide what should be attached to the briefing package. Technically, the platform combines a React and TypeScript front end with a Fastify API running on Azure behind Application Gateway with WAF. The API orchestrates retrieval, drafting through Azure AI Foundry, and automated checks for missing sections, unsupported claims and length overruns. The workspace presents sources and open questions alongside the draft, maintains version history and records sign-off, ensuring transparency throughout the review process. Staff remain fully responsible for the final content and clearance, but they begin from a structured, evidence-backed draft rather than a blank page. The result is a workflow that improves drafting speed, strengthens traceability and supports higher-quality briefing preparation in a controlled government setting.
Key Results
- Produced a checked draft in a median of 84 seconds, with an estimated AI generation cost of €0.21 per briefing.
- In the October 2026 benchmark, version 154 completed all 32 requests without errors, and every draft included the requested deliverables while staying within the benchmark’s upper length limits.
- Reduced estimated AI generation cost by 7.5%, from €0.224 to €0.207 per draft, while cutting unsupported-detail flags by 60%, from 0.47 to 0.19 per draft, and improving overall quality from 3.32 to 3.39 out of 5 on the primary rubric.
Technologies Used
Have a project in mind?
Get an instant AI price estimate for it, or talk directly to our team.
One email a month on what we learn building with AI