Every company that has been around for a while accumulates an impressive amount of material. Presentations, campaigns, research, planning documents, brand manuals, reports, logo versions, briefs, studies, work that got approved, work that didn’t, decisions made in meetings nobody quite remembers the date of. In larger companies, that archive tends to be spread across servers, drives, platforms, suppliers and, above all, people.
At first glance, it looks like the company holds a great deal of knowledge about its own brand. But storing information and preserving memory are different things.
A file records what was done. Memory helps you understand why it was done that way.
That difference, which for a long time seemed to be merely a matter of organization, is starting to take on a far greater strategic importance.
Brands lose context all the time, even when they don’t notice.
Picture a relatively simple decision: for years, a company avoids a certain kind of image in its communication. The choice may have come from an old piece of research, from a cultural issue, from a bad experience with consumers or simply from a very well-conducted discussion among people who knew the brand deeply.
Time passes, some people leave the company, others take over their roles, and the decision survives only as a rule: “we don’t use that kind of image”.
At some point, someone naturally asks why.
If nobody can answer, there are two equally bad possibilities. The rule keeps being followed without anyone knowing whether it still makes sense, or it gets abandoned without anyone knowing why it existed.
This kind of loss happens constantly, because companies are much better at documenting decisions than at documenting the reasoning that led to them.
The problem rarely shows up dramatically. It builds up little by little: when a new team starts re-litigating questions that were already settled, when a supplier receives an incomplete brief, when a campaign looks correct but strangely distant from the brand, or when every new project requires someone to explain all over again who that company is.
There’s a hidden cost in that: the cost of starting over.
Brand knowledge tends to live in people more than in manuals.
Brand books matter, and so do tone-of-voice guides, design systems and strategy documents. They can record colors, typography, principles, positioning, language and many of the elements needed to stay coherent over time.
But brands don’t live on those rules alone.
There is an enormous amount of knowledge that rarely makes it into the manual. The story of why a particular campaign worked, the subject that has to be handled carefully, the argument a certain audience rejects, the solution that looks good in a presentation but has already failed twice in practice, the concession made for a specific commercial reason, the tension between two departments, or the context that explains why an apparently odd choice still makes sense.
It’s that accumulated knowledge that lets someone experienced look at a piece of work and say “that doesn’t sound like us”, even without being able to point immediately to which rule was broken.
That may be where an important part of a brand’s memory lives.
It isn’t just a set of previous answers; it’s also the record of the decisions, the exceptions, the attempts, the consequences and the lessons that brought that company this far.
Artificial intelligence makes this conversation even more important.
The more companies bring artificial intelligence and agents into their content, service, personalization and decision processes, the greater the need to give those systems context. Recent research has begun to point to a risk called agentic brand drift: the possibility that autonomous systems make thousands of apparently efficient decisions and, over time, gradually pull the brand’s expression away from its original identity. The curious part is that this drift can happen while the operational metrics stay healthy, which is exactly why it can be so hard to notice.
Teaching an artificial intelligence that a brand is “close, innovative and human” solves very little. Those words fit thousands of companies.
What actually sets a brand apart shows up in the choices it makes when there are tensions. When being fast conflicts with being careful, which one weighs more? When there’s a commercial opportunity that goes against one of the company’s convictions, what is the decision? When humor and reputation collide, how far does that brand go? When efficiency and customer experience point in different directions, which one prevails?
It’s in the answers to those questions that an identity stops being a collection of adjectives and starts working as a decision system.
Which is why we may have to widen what we call brand management.
The next brand book may need to explain less about how the brand looks and more about how it thinks.
For decades, most brand governance concentrated on visual and verbal consistency. How to apply the logo, which colors to use, how to write, which images to choose, which expressions to avoid.
All of that is still necessary, but it isn’t enough in an environment where systems, platforms, internal teams, agencies and artificial intelligences can produce thousands of interactions on a company’s behalf.
In that scenario, it may be necessary to record the reasons as well.
Not just “this is how we do it”, but “this is how we do it because”.
Not to turn every decision into documentary bureaucracy, but to preserve what normally disappears when people, suppliers and structures change.
A company that manages to keep that kind of knowledge doesn’t necessarily have to repeat the past. Quite the opposite: it is in a much better position to disagree with it.
When we know why a decision was made, we can assess whether that context still exists. When we know only the rule, we’re left with two far less intelligent alternatives: obey it or break it.
Memory isn’t there to keep a brand still.
There’s an obvious risk in this conversation, which is turning history into excessive reverence and starting to justify any decision with “we’ve always done it this way”.
That isn’t memory, it’s complacency.
Good organizational memory should produce exactly the opposite, because it lets you tell what was structural from what was circumstantial. Some choices deserve to stay, others have already served their purpose, and others need to be completely revisited, but that assessment gets better when we know where they came from.
Perhaps one of the most useful questions for companies adopting AI, changing structures or accelerating production isn’t only “what do our systems need to know about our brand?”, but “what does our own company still know about it?”.
Because there may be hundreds of gigabytes of files, impeccably organized campaigns and an entire library of presentations.
And, even so, very little memory.
