What Shopify’s CEO Gets Right — And Wrong — About Workforce AI

Shopify’s CEO Tobias Lütke released an internal email he’d sent to employees with the subject, “AI usage is now a baseline expectation.” His enthusiastic missive implores employees to adopt AI tools “as a thought partner, deep researcher, critic, tutor, or pair programmer.” His ultimate conclusion? “AI will totally change Shopify, our work, and the rest of our lives.”

Lütke says the company needs to embrace AI to keep up with its explosive growth: “In a company growing 20–40% year over year, you must improve by at least that every year.” Shopify doesn’t want to hire more employees, however: While the company has grown at least 21% per year since 2022, the number of employees has declined from 11,600 in 2022 to 8,300 in 2023 to 8,100 at the end of 2024. So this is an organization dedicated to efficiency.

Reading Lütke’s memo reveals some important lessons about workforce AI, including several that you should emulate and others that you should avoid.

What Shopify Gets Right: Vision, Practice, And Community

You should take inspiration from several Shopify beliefs and practices for workforce AI:

  • Develop a vision. Lütke writes: “You’ve heard me talk about AI in weekly videos, podcasts, town halls, and [Summit, a Shopify event].” Executive leadership is crucial to workforce AI efforts: Demystifying myths (such as “AI will steal my job if I use it”), establishing the benefits to both the organization and to employees, and painting a picture of the future state are all crucial to driving adoption success. For example, leaders can position AI as an opportunity builder for employees — taking boring, predictable tasks off their plates, improving their competitiveness in the job market, and solving hard problems.
  • Empower learning on the job. Lütke writes: “Using AI well is a skill that needs to be carefully learned by using it a lot.” On-the-job experience is central to successfully learning how to use generative AI. Employees must apply AI to tasks along their daily employee journey that can benefit from the productivity-enhancing impact of genAI. They will learn from both their successes and failures.
  • Encourage social learning. Lütke writes: “Share what you learned […] Slack and Vault have lots of places where people share prompts that they developed.” Social learning is twice as important as formal (think classroom) learning. Clients report that vigorous peer-to-peer and champions programs are central to successful genAI adoption. Many employ active Slack or Teams channels dedicated to workforce AI, too.

What Shopify Gets Wrong: Expectations And Learning Styles

There are a few beliefs and practices in Shopify’s memo that you should avoid, however:

  • Be judicious, not reflexive, about using AI. Lütke writes: “Reflexive AI usage is now a baseline expectation at Shopify.” Our artificial intelligence quotient (AIQ) research shows that not all employees have the understanding, skills, and ethical awareness needed to use genAI appropriately. High-AIQ employees know when to use it and when not to use genAI for their work tasks.
  • You can’t prove a negative. Lütke writes: “Before asking for more headcount and resources, teams must demonstrate why they cannot get what they want done using AI.” Demonstrating a financial business case for genAI is actually quite challenging today. Most AI augments rather than replaces human labor at this stage of the market, and “proving” that something doesn’t exist is hard to do.
  • Learning requires a system of engagement. Lütke writes: “Learning is self-directed.” Learning workforce AI requires you to build a comprehensive learning system that combines formal learning, social learning, and on-the-job experience. Even sophisticated companies that rely on purely self-directed approaches are less successful than those that create iterated, reinforcing learning opportunities.
  • Set realistic productivity expectations. Lütke writes: “Get 100x the work done” and “What would this area look like if autonomous AI agents were already part of the team?” While agentic AI is on the rise, its promise remains mostly in the future. Most of today’s AI assistants save more modest amounts of labor time. Setting appropriate expectations and KPIs, and measuring them continuously, sets you up for success. Inflated hyberbole doesn’t.

Next Steps For You? Let’s Talk.

I’ve spoken with over 200 organizations about workforce AI. Forrester clients should reach out to schedule a guidance session with me, and I’ll help you design a successful workforce AI strategy.