Consumer AI has an economics problem that traditional software does not have in quite the same way.
With conventional software, you build the product, maintain the infrastructure, and then adding another user can be relatively inexpensive. Scale has historically been one of the beautiful things about software.
AI changes that equation.
Every conversation requires computation. Every prompt needs to be processed. Every response has to be generated. The more people use these platforms—and the more complex those interactions become—the more infrastructure they need to support them.
Growth does not just create opportunity.
It creates a bigger bill.
So if hundreds of millions of consumers eventually spend significant portions of their day inside AI assistants, there is an obvious business question: how do you monetize all that attention?
We have seen one answer work at massive consumer scale before.
Advertising.
Search engines did it. Social platforms did it. They built products people used constantly, aggregated enormous audiences, and then created advertising businesses around that attention.
AI platforms could follow a similar path.
That does not mean every company will.
Anthropic, for example, has publicly said Claude will remain ad-free, relying instead on subscriptions and enterprise revenue. Whether every AI company maintains its current business model over the next five or ten years remains to be seen.
But if you are planning media several years into the future, ignoring AI as a potential advertising channel would be shortsighted.
And that brings us to the bigger problem for marketers.
How are you going to measure it?
The old digital advertising playbook was built around clicks.
Someone searched. They clicked an ad. They visited the site. Maybe they converted.
AI does not necessarily work like that.
Imagine someone asking an AI assistant which running shoes they should buy, what software their company should use, where they should travel, or which financial product fits their needs.
A brand could influence that decision without generating the traditional click trail marketers have relied on for years.
The consumer may receive a recommendation, continue researching elsewhere, visit a store, search the brand later, or purchase through another channel entirely.
Now attribution gets complicated.
The AI platform may report one version of performance. Search may claim another. Retail media might claim the sale. Paid social could have touched the same customer days earlier.
We have seen this movie before.
Another powerful platform enters the media mix, creates its own reporting environment, and asks advertisers to trust the numbers inside its walls.
That is why marketers should think about AI advertising measurement before the budgets become meaningful.
You will need measurement that can account for impressions and influence without depending on individual-level tracking or last-click attribution. You will need to understand incrementality, carryover, cross-channel effects, and whether AI exposure actually changed business outcomes.
The interesting question is not simply whether AI assistants become ad platforms.
Some will. Some may not.
The important question is whether your measurement strategy will be ready if consumer attention moves there.
Because wherever attention goes, advertising usually follows.
And wherever advertising goes, attribution problems are never far behind.