We recently surveyed 5,000 knowledge workers on AI, and 36% of them said their company still stays silent on AI (26%), or bans it completely (10%).

That’s crazy to me. We have ample data that AI improves productivity across roles – up to 10 hours saved per week for AI experts, according to our research. Banning AI today feels like banning the internet in the late 90s – it might feel smartly risk-averse, but putting it off it asks workers to work with one hand tied behind their back.

But I actually don’t think the bans are the biggest problem. To me, the bigger opportunity is the 26% of companies who aren’t saying anything.

According to our research, “staying silent” is directly correlated with your employees having poor AI proficiency. 35% of AI skeptics and 32% of novices work in a “silent” company, compared to 7% of experts.

So you might think no one needs you to talk about AI in a meeting – but you’re wrong. You need to be doing it in every meeting.

You might think, “Every meeting, Greg? What the hell am I going to say?”

It starts with figuring out what you believe about AI … and then talking about it until people are sick of hearing it. Believe me, the Section team is sick of me talking about AI. But 2 years after I first used ChatGPT, everyone uses AI every day at Section.

The key – sitting down and writing your AI manifesto.

“Our company has an AI policy, isn’t that enough?”

No. You need both, because they serve different purposes and different audiences.

Your AI policy is a risk management plan. It outlines the boundaries of what your team should and shouldn’t do with AI, the risks to be aware of, how to protect personal and company data, etc. It’s the playbook that minimizes the risks associated with your investment in an unproven technology.

The AI manifesto is a "call to action” plan. This document is your POV as a leader on AI’s place in the company. If done right, it serves as a rallying cry to get your people behind your investment and strategy. This is not a set of rules – it’s your philosophy on AI and how it will shape the kind of work you do.

The manifesto is critical because every single member of your team has a set of cultural beliefs about how their personal value ties to their work. For most of them, it looks like “don’t cheat, don’t cut corners, come up with unique ideas that show your strategic thinking.”

AI use is in direct conflict with those traditional beliefs about work. So if you want your team to feel safe in their AI experimentation, you need to put their safety net in writing.

The components of a good AI manifesto

Your manifesto should be short but detailed. You’re aiming for something that’s easy to scan and doesn’t leave room for doubt. Here are the things you need to answer:

1. The role AI will play in your internal operations

What are your expectations for how employees will optimize their internal workflows with AI, and what are the boundaries? For example, maybe you expect AI to heavily automate marketing tasks, but you don’t want it generating letters from the CEO.

2.  The role AI will play in delivering your product or service

If you’re accelerating your product with AI, outline how you want consumers to experience the AI. Which parts of the customer experience are you willing to automate and which parts need to continue to be handled by people?

For both of these scenarios, I would recommend starting with no boundaries – or at least testing them first. Don’t make boundaries just to have them, see how far you can get without them.

2. What are the cultural norms and values around working with AI?

How are you encouraging AI use? When does an employee need to disclose that they’ve worked with AI on a deliverable? Are people rewarded or celebrated for their AI wins?

At Section, we have team Slack channels for discussing AI news, wins, and experiences. And we have a section dedicated to shouting out AI in our weekly all hands meeting. We do demos of new workflows and custom GPTs whenever a team member has one.

3. How do expectations change now that AI is part of everyday work?

If you have mandatory AI use cases, outline them here. If employees are expected to figure out their own function specific use cases, lay that out too. (I think you should do both – prescribe some common use cases everyone benefits from and then let employees run with it.)

To give you an example, I expect any employee coming to me with a proposal to have run it through AI first. If they haven’t, I don’t want to look at it yet.

This section should also acknowledge the risks. AI hallucinates, it makes mistakes, and you need to tolerate some of that if you want your team to use it.

My rule is: No one can make AI the scapegoat. Not the employee who turned in unverified work, and not the company that wants innovation at no cost.

4. What’s the AI budget?

Outline how much reimbursement you’ll provide for new AI tools, outside of your company-deployed LLM – such as transcription tools, audio/video tools, writing tools etc.

At Section, we’ll cover the cost of a trial up to $500 (and anything above $500 requires manager approval). We require the employee to test out the tool for 90 days, then make the case for an annual subscription.

Add in any other AI related expenses you’d be okay with covering here too  – such as attending AI conferences or events, books, or other learning resources.

5. Who’s in charge?

Make it clear what the governance of your AI looks like. Do you have a head of AI who is in charge of everything? Is it managed on a department-to-department basis? Make sure people know who to go to with questions.

We created an AI Manifesto Template to help you get started, you can download it for free.

Creating an AI manifesto that encourages AI adoption

This AI manifesto shouldn’t just alleviate doubts, it should build trust. Your team needs to be confident that they have the full picture and that they can take some good faith risks that won’t land them in hot water.

Our team has 100% AI adoption, and every employee is Practitioner level and above. Here are the things that made a difference when we drafted our manifesto:

1. We didn’t create it in a vacuum. Create a small tiger team to draft this document with. Find the early adopters and enthusiasts and ask them to form this team. If you’re a team of 100, pick 5 people to work on it with you that come from various teams and backgrounds.
   
   
2. It’s a living document. Our AI manifesto was last updated in October, and we’ll continue to update it throughout the year. AI is changing, your work will change, and so will your expectations. So plan to update this document annually at the very minimum.
   
   
3. Be honest and focus on the opportunity. In many functions, adopting AI will completely transform if not eliminate roles on some of your teams. So say that, because people are already thinking it. You’ll come across far more credible if you’re willing to concede some tough points. But make it clear that this change can and should be employee-led – no one is having AI pushed on them, they’re being equipped with the tools to redefine their roles.

And once you’ve nailed down a company-level manifesto, encourage your team leads to create team-level manifestos too that align with the org-wide policies and expectations. If you're a manager, you should not end Q1 without your AI manifesto.

P.S. You can read Section’s AI Manifesto for inspiration, and if you want me to review yours, drop me a note at greg@sectionschool.com and we can arrange a call.