The answer up front: Growth breaks companies by separating decisions from the context that made them smart. The fix is systems that carry context forward: written decisions, one source of truth, and tools with memory. Atlassian’s 2025 State of Teams research found a quarter of teams spend about 25 percent of their workweek just hunting for information that already exists somewhere.
Every great restaurant has one location first.
Location one is magic. The founder is in the kitchen, tasting everything, adjusting on the fly. She knows the Tuesday regulars, why the salmon dish got pulled, which supplier fudges on tomatoes. Then success arrives: locations two through five. And somewhere around location four, a customer eats a mediocre version of the famous dish and shrugs.
Nothing was copied wrong, exactly. The recipes traveled. What didn’t travel was the why behind them: the thousand small judgment calls that lived in the founder’s head. That’s context, and losing it is the quiet tax on growth.
What actually gets lost when a company scales?
Not information. Companies scaling up produce more documents, more dashboards, more Slack channels than ever. What gets lost is the connective tissue:
- Why we decided. The decision survives (“we don’t discount in Q4”), the reasoning dies. New people either follow rules they don’t understand or re-fight settled battles.
- What we already learned. Team two cheerfully repeats the expensive mistake team one made eight months ago, because the lesson lived in someone’s memory and someone left.
- Where the truth is. Ask three departments a factual question, get three answers. Everyone’s right according to their own spreadsheet.
Atlassian put numbers on the symptom in its 2025 State of Teams research (12,000 knowledge workers, 200 Fortune 1000 executives): a quarter of teams spend roughly 25 percent of their workweek searching for information, difficulty finding information ranked as the number one barrier to moving fast, and about half of workers said teams had unknowingly duplicated each other’s work. For the teams affected, that’s a day a week, per person, spent looking for things the company already knows. That’s not a tooling gap. That’s context leaking.
Why do the usual fixes fall short?
The standard prescription is “document everything,” and it fails predictably: documentation written for its own sake goes stale the week after it’s written, and nobody trusts stale documents, so nobody reads them, so nobody updates them. A death spiral with a wiki attached.
The fix isn’t more writing. It’s writing decisions down at the moment they’re made, in one place, with the why attached. One sentence of reasoning captured on decision day beats three pages of documentation written for the auditors. Small habit, compounding return: a year in, your company can answer “why do we do it this way?” without archaeology.
Where does AI change this story?
This is where it gets interesting, because most AI tools have the same disease scaling companies do: no memory. A chatbot is brilliant for a prompt or two, then forgets your entire conversation ever happened. Fast, cheap, and fine for one-off questions. But it holds only so much context, and it starts from zero every session — which makes it another location-four employee who never met the founder.
The AI that actually helps a scaling company is the kind with your context built in: systems connected to your real data, carrying memory across sessions, getting sharper with every project instead of starting over. That’s the difference between hiring a stream of talented temps and growing a staff that’s been with you for years. We build our company on the second kind, which is exactly why we’re opinionated about it.
What should a leader do before the next growth spurt?
Three moves, none requiring a budget line:
- Start a decision log this week. One shared doc: date, decision, owner, one sentence of why. Ten minutes a week, and it becomes the most-read document in the company within a quarter.
- Crown one source of truth per number. Revenue, active customers, pipeline: each gets exactly one home. Every other copy is a view, not a truth.
- Exit-proof the founders' heads. List the five judgment calls only you or your earliest people can make. Write down how you make them. That’s the recipe book location five will cook from.
Do those three and growth stops erasing what made you good. Then, when you’re ready to give that context a workforce that never forgets it, you know where to find us.
One short read like this, twice a month: what’s changing in AI, why it matters, and what to actually do about it. Get the Insights below.
Quick questions
What is “context” in a business sense? The reasoning behind your decisions: why the rule exists, what you learned when it didn’t, and where the real numbers live. It’s what separates copying your playbook from understanding it.
We’re only 15 people. Is this premature? It’s cheapest right now. Context habits started at 15 people scale for free; the same habits retrofitted at 150 cost a re-org and a consultant.
Can’t we just use a chatbot for company knowledge? Off-the-shelf chat tools forget between sessions and don’t know your data. They help individuals in the moment; they don’t accumulate company memory. That takes systems with your context built in.
Source: Atlassian, “2025 State of Teams” (survey of 12,000 knowledge workers and 200 Fortune 1000 executives) — ~25% of the workweek spent searching for information among a quarter of teams; information-finding ranked the #1 barrier to speed; ~half of workers report unknowing duplicate work.
