Stop Answering the Same Question Twice
Every role has a higher-value version of itself. Most people never get to live in it.
The gap between what people should be spending time on and what they actually spend time on is usually significant. Most leaders don’t know how wide it is until they go and ask.
So I went and asked. Structured conversations across the team, working through where time goes, where the friction is, what slows people down. The same thing kept coming up: people spending a huge chunk of their week answering the same questions. Not just customer-facing teams — across the board. Different people asking, same questions landing, same effort going into answering them.
Gaps we’d been living with rather than fixing — because until recently, there wasn’t a better option. Now there is.
The job description just changed
When we hired most of these people, answering questions was the job. And honestly, it worked fine. Someone asks something, you find the answer, you send it. You’re good at it, people like working with you, everyone’s happy.
The best companies figured out years ago that documenting those answers was smarter than repeating them. But even then, the cost of not doing it was manageable. Humans could muddle through with tribal knowledge and “ask Sarah, she knows.”
What’s changed is AI agents can now handle a huge amount of this work. But only if the knowledge is written down, accurate, and findable. The stuff that lives in someone’s head? An agent can’t reach it.
That’s the forcing function. Documentation went from “we should probably do this” to “nothing works without it.” The question is no longer “how do I answer this?” It’s “how do I make sure I never have to answer this again?”
Every repeated answer you give without documenting it is a choice to answer the same question twice. That’s the trap. It feels faster in the moment — and in the moment, it is. The cost compounds slowly. That’s what makes it easy to ignore.
What documentation actually unlocks
If you document well, things change. New team members get up to speed faster. People find answers without interrupting someone else. The same question gets the same correct answer regardless of who’s handling it. And when AI agents are involved, the quality of your output is a direct reflection of the quality of your documentation. Patchy source material means the agent fails — confidently and at scale.
Worth saying though: documented answers still need judgement. Not every question is simple enough to answer straight from a doc. The goal isn’t to remove humans from the loop, it’s to remove them from the parts that don’t need them so they can focus on the parts that do.
Someone has to own it — and maintain it
Documentation systems don’t fail because people are lazy. They fail because ownership is assigned once and never revisited. We’ve learned this at Stora more than once.
The principle that works: one source of truth per domain. Not one giant wiki nobody trusts. One place — Notion, Intercom, your engineering docs, wherever — that is the authoritative answer for that area. The tool matters less than the clarity about which tool owns what, and who is responsible for keeping it accurate.
Each domain needs a named owner. Not collective responsibility — that’s the same as no responsibility. And to be clear: an owner doesn’t mean one person writes everything. It means one person is accountable for the health of the system — setting the standard, making sure contributions happen, and catching what falls through the cracks.
But ownership isn’t just about creation. It’s about maintenance. Documentation updates need to happen at the point of change. When something ships, gets decided, or breaks, the update is part of that work, not a follow-up task that gets deprioritised. It either happens in the moment or it quietly doesn’t happen at all.
Your search data will tell you where to focus. Most of the value sits in a fraction of the articles. Look at what people are searching for and not finding. Look at questions that keep landing on the same topics you’ve already documented — that tells you the doc is either wrong or invisible.
And when someone finds an outdated answer, fixing it is part of their job. Not optional. The moment people find one wrong answer and move on, trust erodes. Once trust goes, contribution goes with it.
The people who get this are the ones worth keeping
I keep coming back to this idea of what makes someone indispensable. Documenting well is one of the clearest signals. It means asking whether what you just figured out is worth capturing so the next person doesn’t have to figure it out again. It costs a few minutes. It saves everyone hours.
The people who don’t — who answer and move on, answer and move on — aren’t bad. They’re just doing what the job used to require.
The job just changed.
