01 the business
Online pharmacy
70 staff, selling in two countries
Rules change in both countries the pharmacy sells in. Newsletters pile up unread. Most updates do not apply, but the ones that do get spotted late, and linking a new rule to the affected internal procedure is slow.
02 before
The problem
Rules change in every country you work in, and newsletters pile up unread in someone's inbox. Most of the updates do not apply to you, but the ones that do get spotted late. Linking a new rule to the internal policy it affects is slow, careful work.
- Updates are read irregularly.
- Relevant changes are spotted late.
- Nobody tracks which policies need updating.
03 the build
What the AI does
- Step 1: Monitors the regulator sites, gazettes, law firm bulletins and newsletters you choose, in each country you operate in.
- Step 2: Drops items that do not apply to your sector, size or locations using rules you set together with us.
- Step 3: Summarises each remaining change in plain English with the effective date and the original source link.
- Step 4: Matches each change to the internal policies or procedures in SharePoint or Notion that it may affect.
- Step 5: Sends a weekly digest to Outlook or Teams and opens a task for each policy that may need an update.
paused: waiting for a person to approve
How it was put together
- Monitored chosen regulator sites, gazettes and bulletins in each country.
- Dropped items that do not apply by sector, size and location.
- Summarised each remaining change with its effective date and source link.
- Matched changes to internal policies and opened a task for each one that may need updating.
04 human in the loop
Where a person stays in charge
Your compliance lead or outside counsel reviews the digest and decides which changes need action. No policy is edited until the policy owner approves the change.
- The compliance lead or outside counsel reviews the digest and decides on action.
- No policy is edited without the owner's approval.
- Every summary links to the original source.
05 rollout
From kickoff to live in 4 weeks
- Weeks 1–2Choose sources and relevance rules with the compliance lead.
- Week 3Run digests in parallel with current reading.
- Week 4Make the digest the weekly routine.
06 after
What changes
You hear about relevant rule changes early and in plain words. Updates turn into tracked tasks with owners, so nothing depends on someone reading every newsletter.
Assume 150 updates a month across sources at 4 minutes each to read and judge, with 85% filtered out as irrelevant.
150 × 4 minutes = 10 hours a month of reading. With 85% filtered, about 1.5 hours is spent on the changes that matter.
What to measure
- Days between a rule's publication and internal review.
- Policies updated before effective dates.
- Compliance lead time spent reading.
- Irrelevant items filtered out.
07 evidence
What the research says
From August 2026, AI systems that talk directly to people in the EU must make clear they are AI, unless that is obvious.
EU AI Act, Article 50: transparency obligations (opens in a new tab) · 2024
NIST's framework organises AI risk work into four functions, Govern, Map, Measure and Manage, with a profile for generative AI added in July 2024.
NIST, AI Risk Management Framework 1.0 and Generative AI Profile (opens in a new tab) · 2023
08 lessons
Mistakes to avoid
- Filters too tight. Review dropped items now and then.
- Summaries without sources.
- Treating summaries as advice. Counsel decides.
09 questions
Common questions
- Does this replace our lawyers?
- No. It helps them and your compliance lead spend time on the right changes.
- Can it track AI rules too?
- Yes. Rules such as the EU AI Act's transparency duties from August 2026 can be on the watch list.