January 12, 2024

How to set an AI policy for your company

hero image for blog post

I’ve gotten a ton of requests for this post – everyone seems to realize they need an AI policy for their business, but they don’t know where to start. 

Today, I’ll walk you through an exercise to determine your company’s stance on AI and codify your AI operating principles. Plus, I’ll share our AI use policy at Section. If you have your own policy, I’d love to see it – shoot me a note at taylor@sectionschool.com

Want to join me for my next workshop, AI for Strategic Decision-Making, on Jan. 24? Get 25% off with code STRATEGY at checkout. Enroll >


The four parts to a great AI policy

  • Your company stance on AI 
  • Operating norms for how you use AI
  • Clear policies and rules for AI usage
  • Real-life use cases from inside your org

Part 1: Your company stance on AI 

Your company stance is a philosophical (but tangible) reflection on how you feel about AI. It helps your employees understand the potential and risks of AI for your business, and how you’re thinking about it at the highest level. 

Reflection questions to write this section

1. What are the threats and opportunities that AI poses to your business?

2. How will AI impact your customers and your category? 

3. How will you treat employees who use AI day-to-day (e.g. promote them)?

4. How important is it to you that your employees are using AI to improve their productivity (must-have, nice-to-have) and why?

Section’s company stance on AI

1. The “AI class” is forming. In the next two years, we’ll see the knowledge workforce start to separate into two classes: people who leverage AI in their day-to-day work, and people who don’t. 

2. The AI class will be the new high performers. In the next five years, we’ll start to see the “AI class” get promoted, rewarded, and compensated at a higher rate than the non-AI class.

3. The AI class will be created bottoms-up (through individuals), not top-down (through company leadership). Organizations will move into the AI class as their employees start using it – not by appointing a head of AI. 

4.  It’s our job at Section to get ourselves and our students into the AI class. The best thing for our organization is to figure out ways to use AI to increase productivity, efficiency, and quality of our work. The best thing for our students is to make them AI-proficient, so that they have more job opportunities, raises, and rewarding work in their futures. 

5. To get students into the AI class, we have to question everything about what we do. We won’t get ourselves or our students into the AI class unless we’re AI-proficient in every area of the business. This will mean changing how we make, market, sell, and deliver our product.

Part 2: Operating norms for how you use AI 

Your operating norms help employees understand how they should be using AI day-to-day (if they should be using it at all!). These aren’t rules and regulations – that’s the next section – but rather a set of principles that guide internal behavior.

Reflection questions to write this section

1. What do you want people to use AI for?

2. What do you NOT want them to use it for?

3. How rigorous and/or fast do you want employees to be when they launch AI tests / try out use cases?

4. What is your stance on using company data in AI prompts?

5. How do you want employees to talk to each other about their AI usage?

Section’s operating norms on AI 

1. Use AI in accordance with our operating principles: 

  • Live and breathe the data – use evidence to identify AI use cases 
  • Get to V1 fast – test your AI use case quickly 
  • Work high-low – look for opportunities to use AI in strategic AND tactical work
  • Live in reality – quickly and honestly evaluate the efficacy of AI tests, and decide whether to continue

2. Use AI strategically: We don’t use AI for AI’s sake. We’re looking for big cognitive gains (high brain power work) and efficiency gains (high manpower work). 

3. Use AI boldly / accept risk: We prioritize experimentation over concerns about internal information in prompts – so we’re okay with employees referencing Section-specific information in their prompts. 

4. Talk openly about the use of AI: We all need to learn from each other, so we talk openly about how we use AI, where it works, and where it doesn’t. 

Part 3: Rules and policies

Your rules and policies should be tangible, specific guidance on how employees should use AI. We recommend breaking this into two sections – rules for using AI day-to-day, and red / green / yellow data policies.

Reflection questions to write this section

  • Which AI platforms do you suggest or require employees use?
  • Are there any AI platforms employees should NOT use?
  • Do you pay for any AI platforms as a company benefit? Will you reimburse employees for AI tools or platforms?
  • How will AI usage be tied to performance reviews?
  • What are your “stop / pause / go” rules around using data in AI prompts? 

Section’s AI rules and policies 

Using AI day-to-day

1. We have a company ChatGPT-4 and Claude account available for use for all employees.

2. If you need your own account to allow for more custom usage (including custom instructions), explain your use case to your manager and we’ll reimburse you for that account as well. 

3. We value fast experimentation. We’ll reward employees for testing new optimizations with AI, even if they don’t work well. We don’t have patience for dismissing a suggestion to use AI without trying it.

Internal Data Policies

Part 4: Real-life use cases

The best way to get your team using AI is to talk openly and regularly about how you’re all using it. At Section, we’ve collected an ongoing list of use cases from each department to serve as inspiration. 

(Some of) Section’s AI use cases

  • [Marketing] AI populates weekly “Week Ahead” email template with upcoming courses
  • [Product] Upload decks to Claude and get back experiment and feature briefs
  • [Creative] Built a custom GPT to answer questions from our CEO about his video production and tech set up
  • [Operations] Develop and brainstorm discussion questions for courses
  • [Sales] Draft outbound sales email templates
  • [Leadership] Upload our board deck and ask for feedback and potential questions as a board member
  • [Finance] Built a custom GPT to review all our faculty contracts and assess for specific criteria

Want to learn to use AI for strategic-decision making? Join my next workshop, AI for Strategic Decision-Making, on January 24. Get 25% with code STRATEGY at checkout. Enroll >

Greg Shove
Taylor Malmsheimer, Head of Strategy