The Right AI for the Right Job: Pega's Tara DeZao on Marketing's AI Maturity Curve

Every year at PegaWorld, Constellation Analyst Liz Miller sits down with a Pegasystems leader, Tara DeZao, for what's become an annual tradition: a candid, marketer-to-marketer conversation about where AI hype ends, and real value begins. This year's conversation cut through a lot of noise and surfaced a few hard truths every enterprise marketing leader should hear.

Outputs Are Not Outcomes

The line that anchored the entire conversation: just because an AI model answers correctly doesn't mean the business has moved forward. Too many organizations are conflating AI outputs - a generated email, a summarized report, a chatbot response - with actual business outcomes like customer lifetime value and sustainable growth.

DeZao's advice: lead with the outcome you want to achieve, then work backward to the right AI for the job. Not every problem needs a model. Some need a new messaging strategy, a partnership, or simply better alignment across teams.

The Martech Stack Is Finally Shrinking

For years, Constellation's analysts have predicted consolidation in marketing technology. This year, DeZao confirmed it's actually happening: enterprises are pursuing vendor takeout, fewer point solutions, and fewer data silos. It's a validating moment for a thesis the industry has held for a long time but rarely seen executed at scale.

It Was Never a Data Problem

Perhaps the most important observation from the conversation: what looks like a data problem or a systems problem is, more often than not, a process problem. CRM promised to solve this in the early 2000s. The Customer Data Platform promised it again a decade later. AI is now closing that workflow gap, not by replacing the systems that came before, but by finally making them work together the way marketers always intended.

Compliance as an Enabler, Not a Blocker

One of the more counterintuitive points: centralizing compliance and governance doesn't slow marketers down; it frees them up. When guardrails are built into the system rather than bolted on after a mistake, marketers in regulated industries can test and learn at a level of granularity that used to be impossible. Decision Hub's approach to centralized compliance lets teams experiment with confidence instead of asking forgiveness.

The Secret Sauce AI Still Can't Replicate

AI can absorb the technical expertise no single marketer can hold in their head, the latest trends, backend configurations, channel mechanics. What it can't replace is the experience, taste, and relationships built over a career. Marketing remains fundamentally relational, and that won't change no matter how sophisticated the tooling becomes.

Stop Talking About Speed

Perhaps the most pointed takeaway from the conversation: enterprises need to stop bragging about how fast they built an agent or launched a workflow. Velocity is not the outcome. In fact, over-indexing on speed risks giving marketing teams a false sense of failure, as if they're behind when, in reality, most organizations are still early on a long maturity curve.

The Bottom Line

AI's promise in marketing isn't about doing more, faster. It's about finally solving the process and workflow problems that CRM and CDP investments never fully closed, while keeping the judgment, taste, and guardrails that protect the brand firmly in human hands. That's the realistic, pragmatic view of AI that Constellation Research continues to advocate for, and it's exactly the conversation our analysts are having with enterprise leaders across every industry.