How Companies Can Really Get Value from GenAI

What the Shopify memo says about making GenAI stick

Recently, Shopify’s CEO, Tobi Lütke, made waves with a memo that sets a clear standard for AI adoption. Lütke made it clear to his employees: using AI isn't a choice—it's mandatory. Here are a few ways he wants to encourage AI use:

• Everyone should be using AI a lot and sharing what they’ve learned through internal channels

• All prototyping must be done with GenAI

• Performance reviews will measure whether employees use GenAI tools

• Before asking for new headcount, you will need to prove that AI can’t do this job

This memo highlights a growing theme from recent conversations with funds: investors are increasingly looking for ways to help portfolio companies leverage Generative AI (GenAI) to boost actual financial results.

I’ve seen three approaches to getting a company to adopt and get value from GenAI. Two of them work, and one of them doesn’t. The Shopify approach uses elements from all three.

What Doesn’t Work: Dabbling for Dollars

Many companies want to jump on the GenAI bandwagon but don’t want to mandate anything from the top. They give broad access to platforms like ChatGPT Enterprise and encourage teams to go crazy. Initially, this seems to work: analysts build numerous custom GPT tools to summarize common documents, finance departments streamline tedious processes, and many peoples’ workdays feels a little lighter.

However, this kind of bottoms-up experimentation often doesn't lead to financial impacts. In the best case scenario, you get highly diffuse productivity boosts—a few minutes saved here, an hour saved there—which rarely accumulate into meaningful cost savings or significant revenue growth. You also get limited sharing of successful ideas. The analyst does not have an easy way to show other analysts the custom GPT she built. Some employees might get an extra coffee but that’s all.

And that’s the best case. The more likely situation is that a small number of people use GenAI, but most people just keep doing what they’ve been doing and ignore it. Occasionally, people hear about a great new tool or custom GPT, but they don’t understand what it is or how they can get involved. Maybe they try it once, decide the output isn’t perfect and go back to the old way.

Companies taking this approach often say that “GenAI is embedded in the organization” or that “everyone is using GenAI” with cool examples in pockets. But, they haven’t made the fundamental change needed to get results.

From reading the Shopify memo, I suspect they started with this approach. They are trying to move away from dabbling by making usage mandatory and providing a platform for sharing what works. However, I think Shopify might be even more successful if they focused the GenAI effort on some specific high value use cases.

What Works: The Carrot Approach

A more effective strategy than dabbling is aligning GenAI projects with the company's top strategic goals. Instead of treating AI as an isolated innovation exercise, investors and executives should first ask: "What are our company's 3-5 most important business objectives for the next year?" (Not AI objectives – just overall business objectives.)

Once you've identified these priorities, think about which ones GenAI can accelerate. Maybe GenAI fits perfectly into several strategic goals or maybe even none. But where there's alignment, you’ve found the right place to use GenAI because if it works, you’ve just helped solve once of the business’s most pressing problems. Presumably, that will lead to a meaningful increase in profit and attract volunteers who want to help.

Examples of company objectives where GenAI is a good idea include:

• Improve sales productivity: AI tools can rapidly identify and qualify prospects at scale to grow the funnel. It can help sales reps create custom messages and materials faster than ever

• Reducing Costs: Automating tasks in customer support, IT help desk, or marketing can lead to substantial cost savings

• Product Development: New GenAI-enabled features could make the product more valuable to customers, and GenAI can make developers more efficient in prototyping and building new offerings

Focusing AI efforts on critical strategic goals is far more impactful than random experimentation. People are more willing to try AI when it’s mandated directly.

Shopify is mandating GenAI for the third one of these – product development – by requiring that employees create GenAI prototypes. I hope that they are also applying GenAI strategically for their biggest challenges rather than trying to apply it to everyone on everything.

What Works: The Stick Approach

Alternatively, there's a tougher route: the "sink or swim" method. Choose a department where AI should already be making a difference but isn't yet. Be direct with the leadership: "Here are AI tools to make you more productive. However, your staffing budget is frozen, and the business is set to grow 10% this year—figure out how to use these tools to make it work."

This approach pushes teams out of their comfort zones, forcing rapid adaptation and integration of AI tools. Shopify is doing this by forcing groups to prove that AI can’t do a job before adding headcount. Of the various initiatives they are pursuing, this is the one I believe is most likely to have a near-term financial impact.

Time to Choose Your GenAI Strategy

If you're an investor or executive, now is the moment to decide how you'll strategically push GenAI integration to get real results. Shopify's example shows that effective AI adoption isn't about casually experimenting—it's about clear expectations, strategic alignment, and decisive leadership. Companies that embrace this focused, deliberate approach will reap substantial financial benefits, while others risk falling behind.