Where Buying Decisions Now Get Made
Key Takeaways
AI is becoming a decision layer, not only another channel.
Trust now decides whether AI recommendations survive buyer validation.
Journey maps must include assistants, evidence, and human review.
The customer journey map most teams use is out of date. The single most important step in a purchase, the moment of comparison and recommendation, is moving into AI assistants (chat tools like ChatGPT or Claude that answer questions and suggest options), and most maps leave that step out. If we keep spending against the old map, we will keep shaping the inputs to a decision while ignoring where the decision is actually made.
What changed
A touchpoint is any moment where we can reach a buyer: an ad, a search result, a web page, an email, a demo. The old map treats each one the same way. Every touchpoint hands the buyer information, and the buyer does the deciding over time. An AI assistant works differently. When a buyer asks it to compare options and recommend one, the tool gathers, filters, compares, and recommends in a single step. It does the deciding work that used to spread across many touchpoints. That is why it now matters more than any single ad or page.
Why it has so much leverage
Influence at an ordinary touchpoint is partial, because it is one input among many the buyer will weigh later. Influence at the AI step is close to decisive, because that step does the weighing. The figure we have points the same way. Brands named inside an AI answer earn about 35 percent more clicks than brands ranked just below but not mentioned. That number probably understates the effect, because it misses buyers who act on the recommendation without ever clicking. The 35 percent figure is not sourced in the original draft, so we should confirm where it comes from before it appears in front of the board.
The decision now moves across three layers
The first layer is our familiar informational touchpoints: ads, content, email, and website. These are still real and still worth funding, but they increasingly feed the decision rather than make it.
The second layer is the AI step, where comparison and recommendation now concentrate. We do not control it, and we largely cannot see it.
The third layer is human validation, where buyers double-check the AI's answer against a brand they recognize or a person they trust. Two figures describe this layer. About 69 percent of business-to-business buyers (companies selling to other companies) verify AI insights with a sales rep. And reported trust in AI answers fell from 82 percent to 54 percent on helpfulness in a single year. Both figures are also unsourced in the draft and should be confirmed before use.
One contradiction is worth stating plainly. AI is doing more of the deciding, yet buyer trust in it is falling and most buyers still verify with a human. That tension is exactly why the third layer matters and why we cannot treat the AI step as the whole story.
What this means for where we spend
Most maps and most budgets cover only the first layer, because it is familiar and easy to control. But the decision moved largely to the second layer, and the check moved to the third. If we keep pouring effort into the first layer alone, we shape the inputs while ignoring where the choice is made and confirmed.
Each layer needs a different kind of work. Informational touchpoints respond to campaigns, the work we already know. The AI step cannot be bought like an ad. We influence it indirectly through machine legibility (content structured so an AI can read and represent us accurately) and through third-party consensus (independent sources that back up what we claim). That is upstream, structural work, not a campaign. Human validation is won by brand and proof, because a buyer checking an AI's recommendation is looking for a name they already trust.
Recommended actions
Redraw the journey map to include the AI step and the human validation step alongside the informational touchpoints we already track.
Move some budget from informational touchpoints toward influencing the AI step and winning human validation.
Make our content machine-readable and corroborate it with independent third-party sources, so the AI represents us accurately when it compares options.
Invest in brand and proof, so a buyer who doubts the AI is relieved to recognize our name.
Test our category's real buyer questions in the major AI assistants on a regular schedule, and track how we are represented over time.