What AI Actually Recommends for Lifecycle Marketing
A reviewed exploratory study of how leading AI models choose lifecycle engagement platforms across realistic buyer situations.
Published August 2, 2026 · Reviewed August 1, 2026
432
Total responses
417 accepted
24
Buyer scenarios
6 market segments
3
Models
Anthropic · OpenAI · Google
$6.82
Model cost
Pilot execution
The headline
Visibility is not the same as fit.
Braze received 44.2% of valid unaided recommendations, more than twice the next platform. It is the study’s strongest spontaneous AI default.
After models saw balanced, neutral profiles for scenario-appropriate candidates, Adobe Journey Optimizer was selected in 52.6% of its eligible opportunities—despite only 6.3% unaided recall.
The gap suggests that AI recommendation visibility contains at least two different signals: whether a platform comes to mind, and whether it looks like the best fit once the buying context is explicit.
Unaided recommendation
What models recall
No vendor names or profiles were shown. Denominator: 206 accepted responses.
Equal-exposure preference
What models choose
Rates use eligible scenario-specific opportunities, not a shared denominator.
How to read this note
It measures AI behavior
The results describe model recommendations under fixed conditions. They do not measure product quality, market share, or customer outcomes.
The modes answer different questions
Open mode measures unaided recall. Controlled mode measures preference after balanced exposure. Their rankings should not be combined.
This is a pilot, not a final market ranking
Three repetitions reveal useful patterns, but they are not sufficient for narrow statistical claims or small rank-order differences.