Skip to content

AI Oversharing Report

Rolling out AI on Microsoft 365 - Copilot, ChatGPT, Claude, or an internal agent? Grant read access and we show you, site by site, exactly what it would surface to the wrong people the day you switch it on, before you switch it on.

Grant read access - the report comes back within a few working days. You keep it either way.

The promise

3+

exposures you didn't know about, or it's free

  • Site by site, before you switch anything on
  • What AI would surface, and to whom
  • Fee credited in full against the pack it leads to

The problem

Any AI you point at Microsoft 365 - Microsoft Copilot, an enterprise ChatGPT or Anthropic Claude deployment wired to your data, or an internal agent over SharePoint - grants no new access. It works with the permissions each user already has, and simply makes existing access instant and searchable. That is exactly the problem: most tenants have accumulated years of organisation-wide sharing links, stale permissions, and unlabelled sensitive content that nobody sees day to day - until a tool that searches everything at once surfaces it all at once. It is why so many teams licensed an AI assistant and never switched it on. This is the fast, low-cost way to see the exposure before you enable it: not a full remediation, but a clear reading of what the AI would reach on day one.

What's Included

  • An oversharing sweep of your Microsoft 365 - organisation-wide links, 'Everyone except external users' access, and stale permissions from people who moved or left
  • The sensitive content an AI assistant would surface that is unlabelled or misclassified, identified by site and library
  • A red/amber/green exposure report, with the highest-risk sites ranked by how much they expose and to whom
  • A short call to talk through the findings and what closing them before switch-on would involve

What This Assumes

  • Read-only access to your Microsoft 365 - SharePoint, OneDrive, and Microsoft Graph, plus Microsoft Purview where in use
  • One tenant - a very large SharePoint estate may be re-scoped at quoting

Not Included

  • Any remediation - the report shows the exposure, it does not close it (that is the AI Readiness Pack)
  • The AI assistant's own licensing, rollout, or configuration

Outcomes

  • A clear-eyed view of what AI would expose on day one, before you enable it - Copilot, ChatGPT, Claude, or an internal agent
  • The oversharing hotspots found and ranked, not discovered by an employee's prompt
  • A costed shortlist of what to fix, and in what order, to switch on safely

Who It's For

  • Teams that have licensed Microsoft Copilot, ChatGPT, Claude, or an internal AI agent and held back from rolling it out
  • IT and security leads who need the data-exposure risk quantified before switch-on, not after
  • Anyone weighing the full AI Readiness Pack who wants to see the size of the gap first

Engagement Model

A fixed-price, fast-turnaround report: grant read access and get the exposure report back within a few working days. Defined deliverables, a fixed price, no day rates. And it is guaranteed: if the report does not surface at least three exposures you did not already know about, it is free. The fee is credited in full against the AI Readiness Pack if you book within 30 days.

Pricing: from £2,500 / $3,200 - larger engagements are tailored to scope.

We govern the AI we deliver with

This work is delivered AI-accelerated, so AI tooling reads configuration and permission data from your estate along the way. It is fair to ask how that is governed - and the answer is published: which providers we use and on what commercial terms, what client data never reaches a model, that a named architect signs off every output, and that AI tooling holds no standing write access to client systems.

Read our AI policy

Related

Get started

Find out where you stand, free

Most engagements start with a no-cost posture audit against the baselines that matter to you. If a fixed-fee roadmap is the right next step, we'll say so - and if it isn't, we'll say that too.