Most forward thinking companies want to make AI a part of their woodwork. For example, it is clear from Shopify CEO Tobi Lutke’s post on X and Shopify’s Q4, 2024 Investor Overview, that they aspire to make AI a natural, reflexive part of their work. For this, they need to not just adopt new tools; but evolve skills and processes. Embedding AI requires more than mandates; it demands inspiration, shared learning, and integrating AI thinking into core operations. According to recent report on AI in the Workplace from McKinsey, only 1% of companies believe they are at maturity, and 70% of such digital transformations fail. Based on my learnings on bringing change, I believe that there are three key pillars and 9 specific steps to drive AI adoption. This article go through them using Shopify as an example.

1. Make the "Why" for AI Obvious and Compelling

People embrace change when the reason is clear and connects to a larger purpose. For Shopify, tie the "Why" for AI to the mission and growth aspirations.  

a. Explicitly Link AI to Merchant Success: Continuously communicate how AI tools directly help achieve Shopify’s core mission: lowering the complexity curve for entrepreneurs and making their platform the best place to build a business. Show how AI helps solve their problems.  

b. Showcase Tangible Wins (Low-Hanging Fruit):  Actively find and celebrate internal examples where teams used AI to achieve significant (10x or even 100x) improvements – whether in product development, merchant support, or operational efficiency. This builds trust and demonstrates immediate value.  

c. Connect AI to Market Leadership: Use business context (like market share growth and GMV expansion ) to illustrate how AI is crucial for staying ahead in the e-commerce landscape. Frame AI adoption as essential for continued success and innovation.  

2. Inspire Adoption Through Leadership and Shared Learning

Mandates alone don't create lasting change; inspiration and a supportive culture do. Foster an environment where learning and experimenting with AI feels exciting and collaborative.  

a. Lead by Example (Walk the Talk): All leaders, including the executive team, and at every level in the org chart, must visibly use AI as a thought partner, researcher, or tool in their own work and share their experiences, including challenges. The following statement in Tobi's article on X is powerful. As explained in the HBR - Culture Change That Sticks, role-modeling by senior leaders (not just the CEO) is a powerful driver.

b. Amplify Peer-to-Peer Knowledge Sharing: Encourage and recognize ("the first followers" ) those already sharing useful prompts and AI use cases on internal platforms. Facilitate spaces where teams can share wins and lessons learned, building collective skill and trust. It is important that they not only share learnings, but also connect learning to impact, that is believable.

c. Frame AI as Skill Development: Position AI proficiency not just as a job requirement, but as a vital skill for personal growth that aligns directly with Shopify’s core values of being a "Constant Learner" and "Thriving on Change". Emphasize the opportunity to augment their craft. In Colgate-Palmolive, for creating data fluency, they created a learning academy, custom contextual learning suggestions per department, external credentialling that motivated people on board and it become a movement. See related story from the Data Chief Podcast.

3. Integrate AI Thinking Deeply into Workflows

For AI use to become reflexive, it needs to be woven into the fabric of their daily operations, product development cycles, and resource allocation.  

a. Embed AI in Early-Stage Exploration: Make AI exploration a standard, expected part of the prototype phase in every Shopify project. Use AI to dramatically accelerate learning and information creation from the start, fueling faster innovation cycles.

b. Build AI Proficiency into Performance Development: Include constructive questions about AI usage and skill development in performance and peer reviews. Focus on the learning process and effective application, not just ticking a box, providing feedback to overcome initial hurdles.  

c. Encourage AI-Augmented Headcount Requests: In addition to using head count approval gates and demanding why something cannot be done with AI, which can appear negative, a positive approach may be to support requests that have clear measurable business outcomes with AI amplification built into it.

Bonus: Understand the Law of Diffusion

The Law of Diffusion of Innovation is a time-tested idea from the 1960s. Human nature hasn't changed much. It is well-worth your time to understand it for bringing any change, including AI Adoption, to complex human systems.