In partnership with
Friday. Made it.
Somewhere in your agency relationship there is now a list of what AI may and may not touch. If you did not write it, somebody wrote it for you, and you will meet it as a delay.
Margo Waldrop's piece in The Drum is the clearest account I have read of where those lines are landing, and the range is wider than I expected.
Michael Sturrock at the DMA: "One member described a client instruction that effectively meant no technology containing AI at all, even AI assistants embedded in standard office software." That is not a policy about generative video. That is a policy that bans autocomplete.
The split that matters is production. Jay Pattisall at Forrester has the numbers: 58% of agency executives use generative AI for ideation, dropping to 30% for image or video and 29% for copy.
"We're absolutely seeing a divide between production and non-production use cases with the latter being more common," he says. And legal worry is not the top reason. Reliability and accuracy rank slightly higher.
So brands are answering by naming vendors. Several will only allow Adobe for production work involving AI, because of the indemnification it carries. "Third-party indemnification from the tech provider is how brands and agencies mitigate the risk," Pattisall says.
Behind that sits the US Copyright Office standard on human authorship, which turns the amount of human input from a craft argument into a question of whether you own the thing.
Three examples in the piece draw the line in three different places. REI pulled out of a Meta AI personalization tool in June after it altered a vendor's image and produced a bicycle with two sets of handlebars. Aerie pledged never to use AI to generate people or bodies. Almond Breeze made the absence the campaign, running the Jonas Brothers under "No AI Needed."
Three brands, three boundaries, none of them about whether the technology is any good.
If nobody at your company has written this down, your creative options are being set by whoever on your agency's team feels most nervous, one brief at a time. When did you last read your own list next to what your team has open?
Today’s sponsor:
NVIDIA's Founder Says Farmers Should Absolutely Use AI
“If I were a farmer, I would absolutely use AI.”
That’s Jensen Huang, founder and CEO of NVIDIA.
And he's pointing to one of AI’s biggest untapped opportunities: Farming. It’s an industry facing mounting pressure to produce more with less and it’s still massively under-automated.
DIT AgTech brings AI, nutrition automation, and real-time data to livestock production, helping ranchers boost productivity and get more from every animal.
And it’s already proven in one of the world's toughest livestock environments:
500+ units deployed
370,000 head of livestock on the platform
Up to 55% higher daily weight gain
Now expanding into the U.S. and Brazil, DIT AgTech is targeting a 300M+ head cattle market. And the biggest barrier to adoption? Gone. Ranchers get the technology for free when they sign up for a three-year nutrition plan.
DIT AgTech can scale adoption faster, which means more data and more recurring revenue.
Invest before this early-stage opportunity gets harder to access.
𝘐𝘯 𝘮𝘢𝘬𝘪𝘯𝘨 𝘢𝘯 𝘪𝘯𝘷𝘦𝘴𝘵𝘮𝘦𝘯𝘵 𝘥𝘦𝘤𝘪𝘴𝘪𝘰𝘯, 𝘪𝘯𝘷𝘦𝘴𝘵𝘰𝘳𝘴 𝘮𝘶𝘴𝘵 𝘳𝘦𝘭𝘺 𝘰𝘯 𝘵𝘩𝘦𝘪𝘳 𝘰𝘸𝘯 𝘦𝘹𝘢𝘮𝘪𝘯𝘢𝘵𝘪𝘰𝘯 𝘰𝘧 𝘵𝘩𝘦 𝘪𝘴𝘴𝘶𝘦𝘳 𝘢𝘯𝘥 𝘵𝘩𝘦 𝘵𝘦𝘳𝘮𝘴 𝘰𝘧 𝘵𝘩𝘦 𝘰𝘧𝘧𝘦𝘳𝘪𝘯𝘨, 𝘪𝘯𝘤𝘭𝘶𝘥𝘪𝘯𝘨 𝘵𝘩𝘦 𝘮𝘦𝘳𝘪𝘵𝘴 𝘢𝘯𝘥 𝘳𝘪𝘴𝘬𝘴 𝘪𝘯𝘷𝘰𝘭𝘷𝘦𝘥. 𝘋𝘐𝘛 𝘈𝘨𝘛𝘦𝘤𝘩 𝘩𝘢𝘴 𝘧𝘪𝘭𝘦𝘥 𝘢 𝘍𝘰𝘳𝘮 𝘊 𝘸𝘪𝘵𝘩 𝘵𝘩𝘦 𝘚𝘦𝘤𝘶𝘳𝘪𝘵𝘪𝘦𝘴 𝘢𝘯𝘥 𝘌𝘹𝘤𝘩𝘢𝘯𝘨𝘦 𝘊𝘰𝘮𝘮𝘪𝘴𝘴𝘪𝘰𝘯 𝘪𝘯 𝘤𝘰𝘯𝘯𝘦𝘤𝘵𝘪𝘰𝘯 𝘸𝘪𝘵𝘩 𝘪𝘵𝘴 𝘰𝘧𝘧𝘦𝘳𝘪𝘯𝘨, 𝘢 𝘤𝘰𝘱𝘺 𝘰𝘧 𝘸𝘩𝘪𝘤𝘩 𝘮𝘢𝘺 𝘣𝘦 𝘰𝘣𝘵𝘢𝘪𝘯𝘦𝘥 𝘩𝘦𝘳𝘦: https://bit.ly/4bzuWCi
TALK ON THE MARKET
Sarah Kreps, director of the Tech Policy Institute at Cornell, on AI-generated political advertising, reported by AFP:
The concern might be not that people believe the ads but the ads and general AI slop content risk people doubting everything
Her colleague on the other side of the argument, Isabel Linzer of the Center for Democracy & Technology, points out these ads open the door to creativity with "no special effects editors required." Both things are true, and only one of them turns up in the pitch deck. The risk isn't that somebody believes an AI ad. It's that they stop believing yours.
Want to see case studies of AI-powered campaigns?
Try our sister newsletter:
NUMBERS MATTER
164. Political ads in the US this year generated or enhanced with AI, costing just under $80 million, counted by the Wesleyan Media Project and reported by AFP. They call it an undercount, because no federal rule requires anyone to say an ad was made with AI.
Eighty million dollars of AI creative running in public with no label on any of it. Whatever line your brand draws this quarter gets drawn in that market, next to that.
DAILY PROMPT
Our AI Change Management Plan for Marketing Teams, which is what a red line looks like once it has to survive contact with a team:
Build a change management plan for rolling out [AI TOOL/CAPABILITY] across our [TEAM SIZE]-person marketing team. Work from the files I attached (org chart, current tool stack, workflow docs, last engagement survey, any vendor security documentation) rather than assumptions, and research the web for current vendor admin controls, licensing tiers, and published adoption benchmarks. Cite what you find and list every assumption you had to make.
## Context
- Current state: [DESCRIBE CURRENT WORKFLOWS, TOOLS, PAIN POINTS]
- Target state: [DESCRIBE DESIRED FUTURE STATE WITH AI]
- Timeline: [IMPLEMENTATION TIMELINE, e.g., 90 days]
- Key stakeholders: [LIST ROLES: CMO, content managers, analysts, etc.]
- Primary resistance drivers: [IDENTIFY LIKELY OBJECTIONS: job security, learning curve, quality concerns]
- Governance constraints: [DATA, SECURITY, AND LEGAL REQUIREMENTS]
## Deliverables Required
### 1. Change Impact Assessment
Analyze how this change affects each role in the attached org chart. Separate tasks that get automated, tasks that get augmented, and tasks that stay human. Quantify time reclaimed per role per week and name the new skills required.
### 2. Stakeholder Communication Strategy
Develop messaging for each audience:
- Executive leadership (focus on ROI and competitive advantage)
- Team members (focus on skill enhancement and career growth)
- Individual contributors (focus on workflow improvements and support)
Include 3-4 key messages for each group with channels, sequencing, and the specific objection each message answers.
### 3. Training and Enablement Plan
Create a phased enablement approach:
- Weeks 1-2: foundations, governance rules, use case mapping
- Weeks 3-4: supervised practice on live workflows from our own catalog
- Week 5 onward: delegation of multi-step work, evaluation habits, internal playbook building
Specify formats, owners, and a success metric for each phase.
### 4. Risk Mitigation Strategy
Identify 5-7 risks (quality variance, data handling, skill gaps, resistance, integration, over-reliance) with mitigation tactics, owners, and contingency triggers.
### 5. Success Metrics and Monitoring
Define leading and lagging indicators:
- Adoption metrics (active use, depth of use, delegated task volume)
- Performance metrics (cycle time, output volume, quality, rework)
- Sentiment metrics (confidence, workload perception, retention risk)
If I attach usage or performance exports, run the analysis and set baselines from the real numbers. Include a 90-day monitoring dashboard structure.
### 6. Quick Wins Strategy
Identify 2-3 low-risk, high-impact use cases that show value within 2-3 weeks, with the evidence each one produces for the wider rollout.
## Output Format
One executive summary page, then detailed sections. Use tables for the stakeholder and risk analysis and a dated milestone timeline. Close with the five questions you need answered to sharpen the plan, then revise once I respond.Most AI policies fail in the same place. They say what is banned and nothing about what happens next. This one forces the other half: who loses which task, what the new skill is, and what evidence the first quick win has to produce.
CMO CORNER
Great leaders don't broadcast urgency. They reduce it.
There are two ways to write an AI policy. One arrives as a warning and makes everybody quietly stop mentioning what they use. The other says where the line is and why, and then people tell you the truth about their workflow.
You find out which one you wrote the first time somebody uses something they should not have. Whether you hear it from them or from legal is the whole measurement.
We provide consulting services for CMOs on AI strategy for marketing teams through our Teamless offering.
TOOL UPGRADE
The safest place for AI this quarter is the work no customer ever sees. Three of those:
Reclaim — best for teams drowning in meetings who want the calendar defended automatically, from $20 a month per user. Try Reclaim here →
Wispr Flow — best for anyone who thinks faster than they type, free to start, Pro at $15 a month. Try Wispr Flow here →
Hynote — best for turning a call into notes and actions without another bot joining it. Try Hynote here →
We have 300+ curated, filtered, and categorized AI tools in our directory. Browse to find the best ones that help your marketing team.
→ We have 300+ curated, filtered, and categorized AI tools in our directory. Browse to find the best ones that help your marketing team.
LEARN THIS
If this week convinced you the policy has to exist: AI Center of Excellence Framework for Marketing. It covers the operating model rather than the rules, which is the part that decides whether rules get followed. Who owns the decision, how a tool gets approved, what happens when somebody wants an exception.
Read it before you write the memo. The memo is the easy part.
→ We publish workshops, courses, guides, manuals, and more on AI marketing for free. Browse our learning resources for free.