How to push back on AI mandates with data, not opinions
The AI mandate arrives in your inbox with no definition of done, no resourcing attached, and no acknowledgment of how the work actually gets built. Your job is to execute it anyway.
That situation has a structural problem underneath it. According to a National Bureau of Economic Research working paper surveying roughly 6,000 senior executives across four countries, 69% of CEOs, CFOs, and senior leaders use AI less than one hour a week. Twenty-eight percent never use it at all. These are the same people setting the automation agenda for teams who use it daily.
Daily use is where you learn that a 15-step workflow takes three weeks to build, breaks three days after rollout, and requires human verification on every output. An hour a week of use is where you learn that AI is simply “fast.” The gap between those two experiences is where bad mandates come from.
The right response is not to push back. It is to educate upward with numbers, a capacity cap, and a plan your leadership can actually defend to whoever is pressuring them.
The numbers that frame the problem
Three data points worth having in your back pocket before you walk into any stakeholder conversation about AI automation:
- EY surveyed 15,000 employees and 1,500 employers across 29 countries and found 88% use AI every day, but only 5% use it to genuinely change how their work gets done. The other 83% are running searches and summaries.
- WalkMe’s State of Digital Adoption 2025 report found 78% of employees use AI tools their company never gave them, 51% receive conflicting guidance on when to use AI at all, and 60% say it often takes longer to figure out how to use AI than to do the task by hand.
- Validity’s State of CRM Data Report 2026, with 500 survey respondents, found that nearly 60% of C-suite respondents and 52% of SVP/VPs said they feel pressure to implement AI tools now while knowing the data inputs underneath those tools are not ready. Senior leaders also reported acting on AI recommendations they later suspected were wrong at roughly twice the rate of individual contributors.
Unapproved tooling, conflicting guidance, and employee learning time that nobody logged add up to work that happened and was never counted. You cannot defend a budget for work nobody wrote down.
️ Start with a 30-day hour tally
The fastest way to lose the executive conversation is to argue about whether AI automation works. Bring hours instead.
Have each person on your marketing team track AI time for 30 days across four categories, drawn from Upwork’s AI-Enhanced Work Models research:
- Time spent checking and fixing outputs
- Time spent learning tools (building custom GPTs, writing skills, configuring workflows)
- Time spent running AI workflows
- Time spent copyediting and fact-checking long-form outputs
Make it explicit that nobody is being evaluated. Anonymous submissions are better if you can arrange them, because you are counting totals, not grading individuals.
Then the stakeholder meeting has three concrete points:
- “We spent 60 hours on AI workflows last month. Here is what each of those hours was for.”
- “Here is the brand and organic work those 60 hours would have covered, and what that work’s estimated ROI is on a six-month horizon when it’s properly maintained.”
- “Here is what we want to keep running, what pre-built solutions or outside help we want to buy instead, and what we need to stop to make room for marketing work that is currently slipping.”
An executive who might wave off a vague ask like “we need more time for content reoptimization” can engage directly with a 60-hour number. That number gives them something concrete to take to the boardroom.
Set a capacity cap before you run any experiments
Experimenting is worth doing. But a cap is what makes experimenting survivable when you are still working out the details.
Start with 10% of your marketing team’s available capacity, measured over four weeks. A three-person team gets roughly six person-days per month, enough for one active experiment. A 20-person team gets 40 person-days, enough for two or three experiments with named owners.
The cap scales with company stage:
- Pre-PMF companies: 5 to 10%, because workflows still change too quickly to justify large automation builds.
- Scaling companies with repeatable acquisition motions: 10 to 15%.
- Mature companies: 5 to 10% inside functional teams, with dedicated automation capacity funded separately.
You can raise the cap to 15%, but only after a completed pilot has delivered measured gains in time, quality, or revenue. Use 20% as a temporary ceiling for a fixed rollout sprint only. If experiments delay other planned marketing work, cut the capacity budget by 5%.
Eight rules for every AI experiment
Apply these rules to every automation experiment your team runs, regardless of scope:
- Name one owner. Not a group. One person who runs the experiment and reports what happened, including time loss and time saved.
- Set a kill date before you start. Two weeks handles most tool and workflow evaluations. If the answer is still unclear when the date arrives, the answer is no, or not yet.
- Budget for tokens and cap the spend. Token spend is the first AI cost that shows up on an invoice, so set the expectation early.
- Budget for time and record it. Tokens are the small cost. Your team’s hours are the large one, and hours show up nowhere on a bill.
- Write down what winning looks like in one sentence. “Cuts brief prep from three hours to one” can be tested. “Helps us move faster” cannot, which is why failed experiments never officially end.
- Turn “use more AI” into a defined outcome. Ask leadership to choose the primary goal: lower cost, faster delivery, higher quality, or more output. A generic mandate cannot be evaluated.
- Present options rather than objections. Bring three choices: preserve the current workflow, run a bounded pilot, or fund a full implementation. Show the required hours, expected gain, risk, and displaced work for each.
- Run new workflows in shadow mode. Keep the existing process running for two cycles while the AI workflow operates alongside it. Compare elapsed time, correction time, output quality, and failures before replacing anything.
Anything past the cap is a project, not an experiment, and projects get a budget line like everything else. The cap also gives your stakeholders something concrete: it converts “the team is doing AI automation stuff” into a number they can approve, defend, and show to whoever is pressuring them.
Name what the cap protects
Write down what the other 80% to 90% of your team’s time is for in the same document where you define the cap. The work that builds organic brand authority is the work that gets quietly dropped when a small team redirects capacity to AI builds.
That slow work includes publishing enough depth on a topic to become the go-to source, earning mentions on the sites AI answers actually pull from, showing up in the Reddit and YouTube threads your buyers read, and keeping your entries current on review sites that get cited in AI responses. None of it produces a visible result in the same week it gets done, which makes it easy to deprioritize under pressure.
A three-person team that redirects 20% of its week to an AI build has stopped doing something. Naming what that something is, before the capacity gets spent, is the clearest argument you have for keeping the cap where it belongs.
If you do one thing this week: Put the 30-day hour tally on the calendar and tell your team it starts Monday. A month from now, you will have the only data that can move this conversation toward protecting the marketing work that actually drives growth.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
