AI news ·
Knowing where not to use AI is a marketing advantage, according to Zendesk
Automating every marketing function without strategy fails, warns Zendesk's VP of Marketing. Winning teams deliberately withhold AI where human judgment matters most.
Automation without strategy is becoming a visible failure point as AI floods the marketing sector. Emma Acton, VP of Marketing at Zendesk, warned at Zendesk Showcase London that companies automating every function will lose their way. The winners, she said, will be the ones who know exactly where to withhold AI and let humans take over.
The data paradox: too much input, too little clarity
Marketing teams are drowning in customer data pulled from disconnected systems. Acton said the real problem is not a lack of information but a lack of cohesion. "Your data sources have got to be taught to each other in order to form that picture of the customer," she explained. Without that connection, teams get fragmentary views that make confident decisions nearly impossible.
Adding to the strain, budget holders now demand clear ROI instead of activity tallies. Acton observed that ROI visibility in most organizations falls into three buckets: "very good, very bad, or none at all. There's a middle ground." That middle ground typically contains partial figures, inconsistent tracking, and an incomplete grasp of customer journeys. When measurement is shaky, every tech investment becomes a gamble.
The human-in-the-loop advantage
Acton argued that smart AI deployment means placing automation where it fits and humans where they matter. "It is still the human in the loop, you are allowing AI to be in the areas that need to be automated and AI agents for those tasks," she said. "But you're then able to put humans onto those more valuable, kind of almost white glove approaches." This framing turns the AI for Marketing conversation on its head: the skill is not plugging AI into everything, but picking the moments that require a person's judgment.
For marketing professionals, the lesson is sharper than a generic "use AI responsibly" message. It points to organizational design. Teams that spread AI too thin risk losing the signals that show which touchpoints actually move revenue. Those that deliberately preserve human-led interactions in critical stages can measure what matters more accurately.
Putting AI-first into practice at Zendesk
Zendesk is applying this thinking internally. Acton described an environment where staff test AI in contained experiments, filter out what doesn't work, and double down on proven approaches-all within defined safety parameters. Leadership backs the process explicitly. "Tom, our CEO, and the executive team, are incredibly supportive of that," she said. The goal is a practical adoption cycle, not a one-off transformation sprint.
The company's outcome-based pricing model reflects the same logic. Instead of charging for AI use by volume, Zendesk ties cost to resolved outcomes. For buyers, this shifts the focus from output metrics to whether AI actually solved a customer's problem. It's a direct response to the vendor noise Acton cited, where competing AI claims make it hard to tell innovation from clutter. This strategic approach aligns with broader AI for Executives & Strategy priorities: proven results over buzzwords.
Why this matters for marketing professionals
The main takeaway is not to slow AI adoption. It's to audit your automation map ruthlessly. Identify the data connections that are still missing, then pinpoint the customer moments where a human hand lifts trust or revenue. Companies that ignore this step will burn budget on AI that clouds rather than clarifies performance. In a market stuffed with vendor promises, the marketer who says "we don't use AI here, on purpose" may have the sharpest competitive edge.