Bringing AI into everyday engineering

Shopify is one of the companies that aggressively introduced AI tools into its engineering workflow. In this podcast, Farhan Thawar candidly explains how Shopify experimented with AI, the lessons learned, and how the company's problem-solving culture helped the transition.

"Shopify often says we aren't a swim-laned company. When we see a problem, anyone curious will step up regardless of title."

Farhan himself once fixed a Wi-Fi issue during a company event, earning the nickname "chief Wi-Fi officer," proving their culture values curiosity over formal roles.

Early adoption and a spirit of experimentation

Shopify was experimenting with Copilot before the official GitHub launch. Farhan reached out to Copilot's team, asking permission to roll the tool out company-wide with a promise to share feedback.

"They said it wasn't ready for production, but I said we just need every engineer to try it as soon as possible."

Since then, Shopify has run trials with Cursor, Cloud Code, Devon, and other agents, measuring what actually moves the needle. Cursor usage has grown outside of engineering—finance, sales, and support teams are adopting it fast.

"The strongest growth has been outside engineering. Even non-dev teams are shipping things with Cursor."

New ways of working with AI

Shopify pairs engineers with AI researchers (Anthropic and others) to explore workflows together. This "pairing" culture champions learning together, not handing off problems to tools without oversight.

"We want to be at the frontier. When something new appears, we try it with the people who built it."

It's about collaboration, not delegation: the company emphasizes solving problems together and learning through the process.

"Vibe coding" and democratized software creation

AI has empowered non-developers to build their own tools. Business teams now create dashboards that integrate Salesforce, Google Calendar, Gmail, and Slack without touching code.

"Non-developers now do their own 'vibe coding.' If it doesn't work out, they iterate. They don't fear starting over."

Unlike WYSIWYG solutions of the past, these creators rely on AI to implement behavior without manually writing every line of code.

Blurred boundaries and new challenges

As "vibe coding" grows, the line between engineering and product blurs. Product managers can now submit pull requests, but understanding the code they submit remains crucial.

"Anyone can open a PR, but you still have to understand what you wrote."

AI-generated code can also bloat, increasing review workloads and making discipline essential.

The future of SaaS, AI, and human expertise

Shopify believes anyone can create their own SaaS, but platform-level systems still need human insight.

"We're still in a human-in-the-loop phase where engineers understand architecture and can rebuild it from scratch when needed."

Rather than threatening software, AI will explode the volume of it.

"The world still needs ten or a hundred times more software, and giving everyone the tools to build it is a win."

Investing in AI usage

Shopify encourages heavy AI consumption and even highlights top token spenders on a leaderboard.

"If AI saves you 10% of your productivity time, $1,000 a month is worth it. Spend $5,000 even."

Farhan mentioned wanting to meet anyone spending $10,000 a month to understand if they're using tools wisely.

"Don't skimp on AI. If you use it intelligently, the payoff is huge."

They also recommend stronger models (01 Pro, 03 Pro, Gemini Ultra) rather than sticking with $10 base tiers.

"Using a cheap model is a missed opportunity. Try a $200 model to feel the difference."

Hiring "AI-savvy" talent

Shopify runs a massive internship program that brings in 1,000 interns annually. These interns are seen as "AI centaurs" who teach the rest of the company new ways to work.

"Interns are our secret weapon. We tell them to bring both an AI and a curious brain to work."

The company converts interns into full-time roles, making the internship itself the primary entry point.

"Internships aren't philanthropy—they're an investment in learning."

Internal tooling with AI automation

Shopify's GSD (Get Shit Done) tracker blends project info with AI-generated weekly updates, pulling PRs and Slack conversations into draft summaries.

"AI drafts weekly updates, but humans own the final version."

This lets engineers skip repetitive documentation and focus on the creative problem-solving work.

Leadership, AI, and accountability

Every director-level hire still goes through a coding interview, and AI tool usage is explicitly allowed. Leaders must critically assess AI-generated code and edit it when necessary.

"Great leaders don't run from coding—they lead teams with impact and technical curiosity."

The ability to evaluate AI output is now a leadership expectation.

Advice for teams introducing AI

Farhan says successful adoption requires modeling: leaders must use AI themselves and publicly share prompts and experiments.

"The most effective approach is for leaders to show how they use AI, build a prompt library, and share wins."

He concludes by reinforcing a culture of experimentation, transparency, and knowledge sharing.

"Don't fear failure. Try things, learn from each other, and repeat."

Conclusion

Shopify demonstrates how AI can change engineering by combining experimentation, openness, investment, and cross-generational learning. Its culture offers a playbook for any organization seeking to grow with AI. 🚀

"Nobody has the perfect answer yet. Keep experimenting, learning, and sharing."