Stop Measuring Agentic AI by Time Saved — Measure It by Business Outcomes

Most marketing leaders are measuring agentic AI with the wrong ruler. In a sponsored interview with Adweek ahead of the Agentic World conference, Rafa Flores — chief product and growth officer at Treasure AI — makes the case for shifting the entire measurement framework away from efficiency and toward business outcomes.

The Biggest Misconception

Flores argues that most marketers conflate agentic AI with generative AI: you prompt it, it produces something, done. His definition is narrower and more demanding. A useful agent must understand a signal in context, make a decision against a clear business objective, take action, and learn from it. Without those four elements, it's not agentic — it's a glorified autocomplete.

What Human-in-the-Loop Actually Means

Many teams interpret "human in the loop" as approval workflows. Flores reframes it as judgment: What am I approving here? What do I want the audience to infer? If you can't answer those questions from the output, the process isn't working — and no amount of approval clicks changes that.

The Right Metric: Business Outcomes

• Wrong question: Is this saving us time?
• Right question: Did the brand level up? Did we generate revenue? Is our forecast improving?
• As Flores puts it: "You're not going to report to the board on time saved."

What Changes for CMOs

Flores predicts the CMO role will expand to absorb product, growth, and operations responsibilities as AI handles more execution work. The same pressure is coming for CROs, COOs, and CFOs.

Note: This content is sponsored by Treasure AI.

🔗 Read the full interview on Adweek