How Retail Media Teams Are Actually Using AI in 2026

One of the most interesting sections in the Skai x Stratably – State of Retail Media 2026 report wasn’t about budgets or channels.

It was about how AI is actually being used inside retail media teams today.

And the reality is far more measured than the hype.

Who were the respondents?

The survey was shaped by practitioners, not observers:

• >80% of respondents manages Business Revenue of > $50M

• Senior marketers, retail media leads, and performance owners

• In-house brand teams, agencies, and commerce-focused organizations

• Decision-makers directly responsible for media strategy and execution

• A strong skew toward teams actively running Amazon Ads, DSP, and AMC

In short — people with real money, real pressure, and real accountability.

AI usage today: recommendation > autonomy

The numbers make the pattern clear:

• ~65–70% of respondents use AI primarily for recommendations & insights

• ~20–25% use AI with human approval workflows

• <10% are comfortable with full, hands-off automation

AI is helping decide what to look at — not what to blindly change.

Human-in-the-loop is the dominant operating model

Most teams prefer:

• AI to surface patterns

• AI to suggest actions

• Humans to approve execution

This isn’t lack of sophistication — it’s risk management.

Trust, not capability, is the bottleneck

Respondents consistently cited hesitation around:

• Black-box decisions

• Limited explainability

• Unclear downstream impact

Until AI can clearly answer why a change is being made, autonomy remains selective.

Approval layers are seen as protection, not friction

Interestingly, approval steps aren’t viewed as slowing teams down.

They’re viewed as:

• Guardrails

• Accountability

• A way to build confidence over time

Autonomy is expected to be earned progressively, not flipped on overnight.

The real demand: fewer decisions, not zero decisions

What teams actually want:

• Less noise

• Better prioritization

• Clear “do this next” guidance

• Automation within boundaries

Not removal from the loop — reduction of cognitive load.

What this signals for the future of AI in retail media

The market is moving in stages:

Recommendation → Approval → Partial automation → Selective autonomy

And most serious advertisers are intentionally in the middle.

Credit where it’s due — thanks to Skai & Stratably for contributing this perspective to the broader retail media community. Reports like this help the entire ecosystem mature.