Case Study: Shopify's AI-First Memo and What It Means
In April 2025, Shopify CEO Tobi Lütke published an internal memo that quickly became one of the most discussed statements on AI at work. It said reflexive AI use was now a baseline expectation, asked teams to show why AI could not do a job before requesting more headcount and said AI use would be part of performance reviews. Here is what the memo said, why it mattered and what a small business can sensibly take from it.
Key takeaways
- In April 2025, Shopify CEO Tobi Lütke published a memo saying reflexive AI use is a baseline expectation for employees.
- Teams were asked to show why AI cannot do a job before asking for more headcount, and AI use would be part of performance reviews.
- For small businesses, the useful idea is making AI a normal part of how work gets done, with training and time to learn.
- Be careful with the harder parts: setting fair expectations, measuring results rather than tool use and treating staff with respect.
Shopify is a commerce platform that lets businesses of all sizes run online stores and sell in person. Because so many small businesses use it, and because the memo was unusually direct, it drew attention well beyond the company. Many owners and managers read it as a sign of where workplace expectations were heading.
This case study sticks to the widely reported content of the memo and does not speculate about its effects on Shopify's staff numbers or results. Company policies change, so check recent sources for Shopify's current approach. The aim is to turn a big-company policy into practical, fair choices for a small team. This is general information, not legal or HR advice.
Background: why Shopify spoke up
By 2025, AI assistants had become part of daily work for many people, but adoption inside companies was uneven. Some employees used AI for drafting, research, coding and analysis every day. Others had tried it once, been unimpressed and gone back to old habits. For leaders, that gap raised a question: should AI use be optional, encouraged or expected?
Shopify's leadership chose a clear answer. Rather than leaving AI to individual curiosity, the memo set out expectations for everyone. Publishing it, or having it shared publicly, also signaled to the wider tech and business world how seriously the company took the shift.
What the memo said
The widely reported core of the memo made three points:
- Reflexive AI use is a baseline expectation. Employees were expected to use AI as a normal, automatic part of their work, not as an occasional experiment.
- Prove AI cannot do it before adding headcount. Teams asking for more people were expected to show why AI could not do the job first.
- AI use in performance reviews. How people used AI would become part of how their performance was assessed.
None of these ideas is entirely new on its own. Many businesses already ask whether software could replace a manual process, and many already expect staff to keep their skills current. What made the memo stand out was the combination, the plain language and the fact that it came directly from the chief executive of a well-known company.
Each point changes something different. The first changes daily habits. The second changes how the company grows. The third changes how people are judged. Together they turned AI from a tool into a management policy.
Reactions: why the memo drew attention
The memo was widely shared and debated. Supporters saw it as a clear, honest signal that would speed up learning and stop teams from adding people for work that tools could help with. Others had concerns that are worth taking seriously, even for a small business:
- Hiring and jobs. Linking headcount requests to AI raised worries about fewer opportunities, especially for entry-level roles.
- Measuring the wrong thing. Rating people on AI use could reward visible tool use rather than better work.
- Uneven starting points. Not everyone has the same confidence, training or type of work, so a single expectation can feel unfair.
- Quality and risk. Pressure to use AI everywhere can push people to rely on it for tasks where errors are costly.
Common mistake
Copying a large company's AI policy word for word. A business with thousands of employees, internal tools and HR teams can absorb a bold policy differently from a team of eight. Take the useful idea and adapt it to your size, your people and your work.
Results: what the Shopify memo shows
Without inventing outcomes, the memo itself shows several things clearly.
What the approach gets right
- A clear signal that removes uncertainty
- Encourages everyone to build AI skills
- Makes leaders question default hiring habits
- Treats AI as part of how work is done
What needs care
- Fairness for people in different roles
- Reviews that reward results, not tool use
- Concerns about jobs and entry-level hiring
- Quality and data risks from overuse
The memo shows that leadership expectations shape adoption. When people know AI use is expected and supported, they are more likely to invest time in learning it. It also shows that how you link AI to jobs and reviews determines whether staff feel empowered or threatened.
Lessons for a small business
1. Make AI use normal, and support it
You can borrow the spirit of "reflexive AI use" without the pressure. Give your team approved tools on business plans, set aside time to learn, share prompts that work and celebrate useful experiments. Expectations without training are just stress.
2. Set realistic, role-based expectations
A bookkeeper, a warehouse packer and a marketing assistant will use AI very differently, and some roles will use it rarely. Agree with each person where AI can help in their role and where it does not fit. Write it down.
I run a small [type of business] with these roles: [list roles]. For each role, suggest: - 3 tasks where an AI assistant could realistically save time - 1 task where AI should NOT be used, and why - a simple 30-minute exercise to practice Keep it practical for people who are new to AI.
3. Ask "could a tool help?" before hiring, but think it through
Before adding a role, it is sensible to ask whether better tools, automation or AI could cover part of the work. But be honest about the answer. Some work needs people, and overloading a small team with "AI will handle it" can lead to burnout and mistakes. Use the question to design better roles, not to avoid hiring when you genuinely need help.
4. Judge results, not tool use
If you bring AI into performance conversations, focus on outcomes: quality, speed, customer results and good judgment about when not to use AI. Avoid counting prompts or logins. Someone who uses AI rarely but wisely may be doing better work than someone who uses it for everything.
5. Be fair and transparent with staff
Explain why you are making changes, what support you will provide and how decisions about roles and reviews will be made. Give people time to adjust. Employment rules vary by country and state, so check the official guidance or a qualified adviser before you change job descriptions, review criteria or contracts.
Pro tip
Include data rules in any AI expectation. Staff should use approved business accounts, check whether the plan excludes data from training and never paste customer or employee personal information into personal AI accounts. Pressure to use AI without these rules creates risk.
Example
Imagine a 10-person online homeware store inspired by the memo. Instead of announcing that AI use will be rated, the owner gives everyone a business AI account, runs a one-hour training session and asks each person to find one task in their role where AI could help. In monthly check-ins they share what worked. Reviews stay focused on results. When the team needs another customer service person, they first improve their reply templates with AI, then hire anyway because volume keeps growing. Everyone understands the reasoning.
Your AI adoption action plan
- Choose approved AI tools on business plans with suitable data controls
- Write a short AI policy covering approved tools and data rules
- Run a practical training session for the whole team
- Agree role-based AI expectations with each person
- Add "could a tool help?" to your hiring checklist, honestly applied
- Keep performance reviews focused on outcomes and judgment
- Explain changes openly and give people time to adapt
Frequently asked questions
What did Shopify's AI memo say?
According to widely reported coverage of the April 2025 memo, CEO Tobi Lütke said reflexive AI use is a baseline expectation, teams should show why AI cannot do a job before asking for more headcount and AI use would be part of performance reviews.
Should a small business make AI use mandatory?
A softer approach usually works better in a small team: provide tools and training, set role-based expectations and share wins. Mandating use without support can damage morale and push people to use AI where it does not fit.
Is it fair to include AI use in performance reviews?
It can be, if you assess results and good judgment rather than how often someone uses a tool, and if everyone has had access to training. Check employment rules and your existing contracts before changing review criteria.
Does this memo mean AI will replace small business staff?
Not necessarily. In most small businesses, AI changes parts of jobs, especially writing, research and admin, rather than replacing whole roles. Customer relationships, hands-on work and judgment still need people.
Related reading
The bottom line
Shopify's AI-first memo is a clear example of leadership setting firm expectations for AI at work. For a small business, the useful lesson is to make AI a normal, supported part of how work gets done, while keeping expectations fair, reviews focused on results and staff fully informed. Copy the intent, not the exact policy. Rules on employment and data vary and change, so check official guidance or a qualified adviser before changing roles or reviews.
