Timo Pelz, VP of marketing at Clever, says his first AI push failed because permission without structure created activity but not transformation.
Clever moved from a mandate to a journey, using its existing monthly all-marketing meeting as the drumbeat and letting colleagues, not the VP, present their AI work.
By August, the 25-person team had over 30 workflows in use, but only about five have reached an actual decision, so Pelz says judgment is now the key skill.
Clever's marketing chief turned scattered AI experiments into shared habit
Rolling out AI across a marketing team is a human and structural problem more than a tooling one. At Clever, budget and permission produced scattered activity, so Timo Pelz rebuilt adoption around an existing monthly meeting and peer teaching until AI became routine. He now argues that adoption is only a milestone, and the real work is turning cheap production into better decisions.
Takeaways
Permission without structure creates activity, not transformation. Pelz first gave teams budget and freedom to try AI tools. An inventory in November 2025 found scattered experimentation, basic prompting and very uneven fluency. He concluded the gap was about context and structure rather than tooling.
Overwhelm, not skepticism, is the real blocker. The implicit message that AI changes everything and must be solved now carries little useful information and puts marketers off. Pelz demystified it with live trivia on how long technologies took to reach 100m users and set the bar at fluency rather than expertise.
Use existing rhythms instead of tiger teams. Clever ran adoption through its monthly all-marketing meeting, framed as a journey rather than a mandate. Pelz says tiger teams let people offload the problem, and that many companies reverting to mandates is unhelpful. Figuera agrees embraced AI beats imposed AI.
Peers teaching peers makes adoption compound. Competitive intelligence was the first proof, with a recurring AI-assisted pulse cutting build time about 75% while humans still interpret it. Engagement shifted when colleagues, not the VP, presented their work. By August the roughly 25-person team ran more than 30 workflows.
Adoption breeds duplication that needs a shared portfolio. Spread-out adoption produced four overlapping takes on competitor vendor evaluations and two audience libraries. Pelz created a central marketing AI portfolio document open to the whole team. Figuera warns that without a shared approach, AI agents get siloed across a company.
Judgment is now the scarce marketing skill. Only about five workflows have reached an actual decision. Most point at production, which AI makes exponentially cheap. Pelz wants to move from efficiency to effectiveness heading into Q4 2026, since the C-suite holds marketing leaders accountable for value captured.
Advice for marketing leaders rolling out AI
Check where you are proactive and where you are reactive on AI.
Expect to miss the right operating model on day one. Try something, observe, learn, add structure and try again, which Pelz says is not failure.
Ignore people on LinkedIn promising a quick AI playbook.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
Standardise how the team uses AI rather than letting everyone go their own way and burn tokens at their own rate, as Figuera suggests.
Treat adoption as a milestone and value capture as the real work.