For most of my career, selling technology meant negotiating between what a business wanted and what technology could economically provide.
A customer could describe a perfectly sensible workflow. Information would come in from the client, move through a few systems, somebody would interpret it, another team would act on it, and the customer would ask why this could not simply happen automatically.
Very often, the answer was not that technology could not do it. The answer was that making it happen required multiple products, connectors, engineering effort, configuration, manual intervention and people who understood where one system stopped and another began.
The technology existed. The economics of making the technology usable did not always exist.
CONTEXT
Technology stored the transaction. People carried the context.
Enterprise software became powerful because it gave organizations structure. A CRM could define an opportunity stage. An ERP could define how an order moved. A ticketing system could define how an issue was classified.
But reality rarely arrived in those neat categories. A salesperson did not experience a customer as “Opportunity Stage: Qualified.” The salesperson understood that procurement was difficult, one stakeholder was supportive, the budget was probably lower than stated, and the conversation had changed after the last meeting.
The CRM stored the transaction. The person carried the context.
KNOWLEDGE
Most organizations have three kinds of knowledge
Over time, I started seeing organizational knowledge in three broad forms.
Transaction: How does something move? Which form is filled, who approves it, where does it go next, and what informal route gets something unstuck?
Decision: What is allowed to happen? Which thresholds apply, who has authority, and where does escalation begin?
Cognition: What does this actually mean?
That last category has historically been expensive.
ECONOMICS
AI reduces the cost of bringing context to the point of work
Imagine that every CRM entry, proposal, transcript, email, objection, meeting note and eventual outcome from ten years of sales activity exists somewhere. Historically, the limiting factor was not only whether the information existed. It was whether anybody could consume enough of it, compare enough of it and retrieve the relevant parts when a decision had to be made.
AI can reduce the marginal cost of some of those activities. It can help retrieve previous cases, compare documents, summarize long histories, classify information, surface contradictions and prepare a richer set of context for a person making a decision. That does not make judgment irrelevant. It changes what judgment has available to work with.
synthesize · surface
point of decision
OUTSIDE THE ORGANIZATION
The same change is happening outside the organization
A few years ago, preparing seriously for a customer meeting often meant searching, opening the first few useful results, reading long articles and trying to construct a reasonable understanding of the customer’s business before time ran out. A great deal of useful information existed. There was only so much one person could economically consume.
AI-assisted research changes that constraint. It can help a person move across more sources, compare perspectives and prepare with a depth that previously required substantially more time.
The advantage was not mystical intelligence. It was access to usable context accumulated through experience.
CONCLUSION
That is what I think is different about AI
Previous enterprise software became extraordinarily good at storing, moving and standardizing information. People remained responsible for carrying much of the context required to interpret it.
AI does not remove every technical or organizational constraint. But it can reduce the cost of specific cognitive activities around work: retrieval, comparison, synthesis, classification, drafting and exploration.
Which previously expensive acts of understanding have become cheap enough to change the way this business can operate?
Once that question is answered, another follows naturally.
What happens when an organization can economically handle more variation than before?