Back to Blogs
Summary: Corporate AI training that only teaches marketers what not to do will not help big companies compete. CMOs need sanctioned AI skunk works tied to pipeline, conversion, retention, customer experience, and cost efficiency. Governance still matters, but it should clear the runway for faster learning, not turn AI into compliance theater and cautious paralysis instead.
Overcoming Corporate Governance
“We all just completed corporate AI training that told us everything we couldn’t use,” shared a CMO from a $6B global conglomerate, who added sarcastically, “and that was fun.”
The huddle shuddered while I pondered the fate of big companies.
To be clear, I’m not anti-governance. Big companies have big risks, big brands, big legal departments, big customer promises, and big blast radiuses. A wrong email from a rogue agent is not a cute learning moment. It is a brand impression that lasts.
But corporate AI training that only teaches avoidance is not training. It is legal self-protection wearing a learning management system badge.
Congratulations, you now know 47 ways to get fired and zero ways to grow faster.
This is where the Innovator’s Dilemma comes paddling into the harbor. Incumbents are not stupid. They are often too rational. They protect the current business, the current process, the current approval chain, and the current definition of “safe” until the new thing looks dangerous, messy, and unworthy of serious resource allocation.
Then the new thing becomes the market.
Years ago, my agency produced a video game called Carrier: Fortress at Sea, which meant I learned more than expected about aircraft carriers. One fact stuck with me: a carrier needs roughly five miles to come to a complete stop.
Five miles.
That is not a steering problem. That is physics.
Large companies have the same problem. They have scale, resources, data, customers, brand equity, and distribution. They also have inertia, compliance layers, procurement drag, approval loops, and a thousand smart people trained to slow things down before something breaks.
So no, a $6B conglomerate will not turn like a speedboat.
But it can launch one.
That is the opportunity for CMOs at large companies right now. Stop waiting for corporate AI training to teach marketing how to compete. It won’t. Corporate can and should create guardrails, but guardrails are not a go-to-market operating model.
Marketing needs its own AI skunk works.
Not a random Slack channel for prompt sharing. Not a lunch-and-learn where someone shows how to summarize a PDF. Not a rogue team hiding from IT and hoping Legal does not notice.
A real skunk works. Sanctioned. Structured. Business-outcome obsessed.
Give it a new mandate: improve pipeline, conversion, retention, customer experience, and cost per opportunity. Give it permission to rebuild workflows, not just add AI sprinkles to broken ones. Give it its own training program because “don’t paste confidential information into ChatGPT” is not a curriculum.
Give it different KPIs. The point is not token usage, tool adoption, or the number of agents built. The point is whether marketing can create more business impact with the same or fewer resources.
That is the part many CMOs are still tiptoeing around. The 2026 planning conversation is not going to be, “How many more people do you need?” It is going to be, “How much more growth can you produce without more people?”
Which brings us to Snowflake.
Denise Persson, Snowflake’s CMO, is a useful counterexample to the idea that large companies are doomed to move slowly. Snowflake has scale, complexity, risk, and hundreds of marketers. And yet, they are building AI fluency across the marketing organization, not waiting for fluency to arrive by memo.
Their approach includes weekly AI skills training, function-level hackathons, an AI council, quarterly AI days, and AI goals inside every person’s quarterly priorities. This matters because AI transformation is not a software rollout. It is a behavior change program with business consequences.
Snowflake also shows why governance should clear the runway, not ground the fleet. Their centralized AI engineering team certifies agents that will scale, prevents duplicate builds, and makes sure the systems behave correctly. That is the right tension: centralize risk control, decentralize experimentation.
And they are getting results. Persson cited a 30% reduction in cost per opportunity over six months by pulling fragmented media channels together and letting the system recommend optimizations daily instead of waiting until a campaign ended to learn what failed.
That is the big-company dream: carrier-sized data, speedboat-style learning.
But here’s the catch. Big companies will not get there by treating AI as a compliance module. They will get there by redesigning the work.
The dashboard is a good example. Dashboards tell you what happened. They do not tell you why. They do not settle the sales-marketing argument over who sourced what. They do not recommend what to change before the month ends.
Talking to your data changes the job. It turns reporting into interrogation. It turns analysis into action. It turns the weekly meeting from “whose number is right?” into “what are we doing next?”
But only if the data is ready.
Bad data plus AI is not a small problem. It is bad decisions at scale. The agent does not fix your messy data estate. It amplifies it. The old Salesforce hygiene lesson is back, only this time the consequences compound faster.
So the CMO skunk works needs more than clever prompts. It needs a data mandate, a workflow mandate, a training mandate, and a governance mandate. It needs GTM engineers, curious operators, adaptable marketers, and people who can translate business pain into repeatable AI-enabled systems.
Meanwhile, smaller-company CMOs should be paying close attention.
Your advantage is not more data, more budget, or more people. Your advantage is maneuverability. You can turn faster. You can test faster. You can train the whole team faster. You can build the new operating model before the aircraft carriers finish scheduling the steering committee.
Do not waste that advantage by acting like a tiny version of a big company.
The biggest companies have scale. The smaller companies have speed. The winners will be the ones that build operating models where AI serves business outcomes, not corporate theater.
This is why I keep coming back to that old carrier game. My favorite part was the challenge of landing an F-15 fighter jet on the carrier. If you crashed, the game delivered a brutal little message:
“Congratulations, you just crashed a perfectly good $30 million airplane!”
That is what many corporate AI programs feel designed to prevent. No crashes. No mistakes. No mess.
Fair enough. Nobody wants to crash the plane.
But at some point, someone still has to learn how to land it.
So here’s the challenge for CMOs at big companies: build the AI skunk works before someone else builds around you. Give it a mandate tied to business outcomes. Give it room to test, train, certify, and scale. Use governance to clear the runway, not close the airspace.
You may not be able to stop the ship in under five miles.
But you can launch something from the deck today.
Q&A
What is the first step in creating a marketing AI skunk works?
Start with one business outcome, such as reducing cost per opportunity or improving conversion. Then build a small, sanctioned team with permission to redesign the workflow around that goal.
How should CMOs balance AI experimentation with governance?
Separate risk control from experimentation. Central teams can certify tools, data access, and scaled agents while marketing teams test practical use cases inside clear boundaries.
What should CMOs measure besides AI adoption?
Measure business impact: pipeline quality, conversion rates, customer experience, productivity, retention, speed to insight, and cost per opportunity. Tool usage is not the win.
What is the biggest risk of corporate AI training?
Training that focuses only on restrictions can create fear without capability. Marketers need to know how to use AI responsibly and how to turn it into better business outcomes.
Do smaller companies have an advantage with AI?
Yes, if they use their speed. Smaller teams can test, learn, train, and redesign workflows faster, but only if they resist copying big-company bureaucracy.