Insights on trustworthy company knowledge
Practical thinking on document quality, AI answers and keeping what your company has written true, current and owned.
10 articles
AI readiness checklist: 8 signs your company's knowledge is ready for an AI assistant
AI doesn't create bad information, it scales it. Eight checks that show whether your documents are ready for an assistant.
Read articleHow to find duplicate files in Google Drive
Google Drive has no one-click duplicate finder. Four manual methods, where they break down, and what it takes to stop copies coming back.
The fastest way to find duplicate files in Google Drive is to combine three built-in tools: the search bar with filters (name, type, owner and date modified), the storage view sorted by size, and a sort by name to catch near-identical titles such as "Report" and "Report (1)". Google Drive has no one-click duplicate finder, so these manual passes, or a dedicated tool, are what most teams rely on today.
The anatomy of a documentation audit: what to look for
Most audits end as a 400-row spreadsheet. The four things a useful audit looks for, and how to keep its findings from decaying.
A useful documentation audit looks for four things: gaps, where a process has no documented answer; contradictions, where two documents disagree; outdated content; and documents with no reachable owner. It agrees what "current" means before anyone opens a file, ranks what it finds by how many people each issue reaches, and gives every finding a named owner and a next step.
Your AI pilot worked. Why did it break at scale?
The model didn't get worse. The rollout reached documents the pilot never touched.
When an AI pilot works and then starts giving wrong answers after a company-wide rollout, the model usually hasn't changed. The rollout has. It extended the assistant into documents the pilot never touched: other departments' policies, old process pages and content nobody owns. The fix is to check those documents before each new phase of the rollout, not after the complaints start.
Why does AI give wrong answers? Often it is the documents
The model rarely invents your policy. It repeats the wrong version of it. Four document failures behind most wrong answers, and how to find them.
AI assistants give wrong answers from company documents when those documents contradict each other, are out of date, name people who have left, or leave a question unanswered. The assistant retrieves whatever matches the question best and states it as fact. Fixing the documents, rather than the model, is what makes the answers reliable.
Conflicting procedures: how to find documents that disagree
Two procedures, two answers, both official. How conflicting information builds up, where it hides, and a method to find it before your AI assistant does.
Conflicting procedures are two or more live documents that give different instructions for the same task. One says expenses must be submitted within 30 days, another says 60. Both look official, and nothing on either page says which is current. To find them, list the questions where a wrong answer costs the most, collect every document that answers each one, and compare the specific figures, dates and steps side by side.
Content audit: a framework for knowledge estates
An inventory tells you what you have. An audit tells you what to do with it. A framework for scoring documents and deciding their fate.
A content audit is a review of every important document in your intranet, shared drives and sites to decide what to keep, update, merge or retire. A useful framework has three parts: an inventory of what exists, a small set of criteria to score each document, and a clear decision rule for what happens next. Without the third part, an audit ends as a spreadsheet nobody acts on.
How to clean up Google Drive before Gemini
Gemini reads what your people can open. Six steps to sort out Google Drive before it does, in the order that removes the most wrong answers first.
To clean up Google Drive before Gemini, work in this order: find out what Gemini will be able to read, remove duplicates and old versions, review outdated files people still open, fix files whose owners have left, archive what's finished, and set simple rules so the mess doesn't return. Start with the files people open most, not the biggest folders.
What is AI grounding, and why your sources decide the answer
Grounding stops an assistant inventing answers. It doesn't stop it repeating the wrong document. What grounding is, how it works and what it needs.
AI grounding means making an AI assistant answer from a specific set of sources, such as your company's documents, rather than from what the model learned in training. Grounding reduces invented answers, but it moves the risk rather than removing it: a grounded assistant is exactly as right as the documents it reads. If those documents are outdated or disagree, grounding delivers the wrong answer, with a reference attached.
Policy change management: from approved policy to updated procedures
Approving a policy change is the easy part. Finding every document that still repeats the old version is where the risk sits.
Policy change management is the process of carrying an approved policy change through to every document that repeats it: procedures, FAQs, onboarding packs, templates and regional versions. Approving the new policy is the easy part. The risk sits in the old wording that stays live elsewhere, because people, and the AI assistants that read your drives, keep finding it.
See which of your documents are misleading your AI.
No articles match
Try a different word, or browse all topics.