The Prior Advantage
Every era of discovery rewards a different behaviour.
The search era rewarded visibility.
The AI era increasingly rewards understanding.
But once priors emerge, something else happens.
The playbooks diverge.
The same system creates two very different games.
And understanding which game you’re playing may become one of the most important strategic questions of the next decade.
First, Know Where You Exist
Not every organisation occupies the same place in the discovery stack.
Some exist primarily in the index.
Others live in the context window.
A few have become part of the weights themselves.
The distinction matters.
Because each position carries different opportunities and different risks.
The Index
You can be found.
But only if the system searches.
Without retrieval, you disappear.
The Context Window
You can be described.
The system can synthesise information about you from retrieved sources.
You influence the answer.
But the answer depends on the search occurring first.
The Weights
You are already understood.
The model carries assumptions about who you are, what you do, and when you’re relevant.
Retrieval may refine those assumptions.
But it doesn’t create them.
The question is no longer:
“How visible are we?”
It becomes:
“Where do we exist?”
The Diagnostic
You can test this today.
Ask several AI systems:
What is this company?
What is it best at?
Who is it for?
First, allow the systems to search.
Then ask the same questions without search.
The difference between the answers tells you something important.
If the model only knows you when it searches:
You exist in the index.
If the model can describe you through synthesis but requires retrieval:
You exist in the context window.
If the model already understands you before retrieval begins:
You have a prior.
Most organisations will discover they are earlier in this journey than they imagined.
That isn’t failure.
It’s a map.
The Prior Advantage
Priors compound.
A recommendation can change quickly.
A default can flip.
A better answer can emerge.
But priors evolve on training-cycle time.
Months.
Sometimes years.
They are built from the accumulated sediment of the corpus itself.
Which creates an asymmetry.
New entrants compete in the context window.
Incumbents often already exist in the weights.
Fresh evidence argues for change.
Existing assumptions resist it.
The prior can outvote the context window.
At least for a while.
Two Opposite Strategies
This creates two entirely different playbooks.
If You Already Have Priors
Defend them.
Your greatest risks are:
→ drift
→ contradiction
→ category confusion
→ loss of specificity
→ forgetting what made you distinctive
Your job is to reduce unnecessary variance.
Protect the truths that made the prior sharp in the first place.
Coherence becomes defence.
If You Don’t Have Priors Yet
Build them.
Your greatest risks are:
→ generic positioning
→ borrowing everyone else’s language
→ compressing into invisibility
→ chasing attention without distinction
Your job is different.
Maximise specificity.
Occupy a distinctive position in the context window.
Play the long corpus game.
Coherence becomes construction.
Distinctive and Sharp
The strongest priors are not merely familiar.
They are distinctive.
Specific enough to occupy their own region of the model’s understanding.
Consistent enough to occupy it sharply.
Coherence without specificity becomes invisible.
Specificity without coherence becomes noise.
The advantage belongs to those who can achieve both.
Say something specific.
Say it everywhere.
The Strategic Shift
For years, organisations asked:
How do we get found?
Increasingly, they ask:
How do we get recommended?
But perhaps the deeper question is:
What assumptions are future systems making about our category?
And:
Are we shaping them?
Because the future of discovery may not simply belong to those who win retrieval.
Or even those who win resolution.
It may belong to those who understand which arena they are competing in and play accordingly.
Conclusion
The same framework creates two very different strategies.
Defend coherence.
Or build distinctiveness.
Protect the prior.
Or earn one.
The internet becomes memory.
The models inherit it later.
And eventually, the organisations that thrive won’t simply be the easiest to find.
They’ll be the easiest to understand.