AI Strategy

What the Big Companies Got Right About AI (and What Small Businesses Should Steal From Them)

Shopify, BCG, Moderna, and Duolingo are all running the same experiment. The patterns in their wins — and one notable failure — are exactly what a small business needs to know.

The companies and what they're actually doing

Shopify went the boldest route. CEO Tobi Lütke made AI usage a baseline expectation across the company. Before any team can request more headcount, they have to demonstrate why AI can't do the work instead. AI usage is now folded into performance and peer review. The signal is loud: AI fluency is a job requirement, not a bonus.

BCG reported that nearly 90% of their consultants are using AI, with about half using it daily. The interesting part isn't the percentage — it's how they got there. AI usage was paired with training teams, local champions, and proprietary tools their consultants actually wanted to use. Adoption followed enablement, not the other way around.

Moderna is probably the most impressive enterprise case. They rolled out ChatGPT Enterprise to thousands of employees and ended up with more than 750 custom GPTs built across legal, clinical, manufacturing, commercial, and research functions. The legal team alone reported 100% adoption. Every function built its own use cases. IT provided the platform; the business built the tools.

Chipotle is using an AI hiring assistant to cut application-to-start time from 12 days to four, and to raise application completion rates from roughly 50% to 85%. Separately, they promoted 23,000 employees internally in 2024, with 85% of restaurant management roles filled from within. AI accelerated the hiring funnel; culture and visible career paths handled retention.

Duolingo is the cautionary tale. Their "AI-first" memo pushed AI into hiring, performance reviews, and resource allocation. Then they walked back the part about measuring AI use in performance reviews, after employees pushed back. When you measure AI usage instead of AI outcomes, your people will give you usage.

The pattern underneath all of it

Translating to a small business

You don't have 550,000 employees to train. You don't have a Chief AI Officer. But the patterns translate cleanly, and in some ways small businesses can execute them faster and better than the giants.

Make AI fluency a baseline expectation — and pair it with real enablement. Not a one-hour onboarding video. Real, ongoing, hands-on practice. A small business with three or four people can run a 30-minute AI working session every two weeks. The point isn't the training itself — it's the signal that this is part of the work.

Redesign one workflow at a time, not the whole business. Pick one — customer intake, proposal generation, lead nurture, invoicing follow-up — and redesign that one workflow around AI from end to end. The mistake to avoid is treating AI as a tool you sprinkle on top of existing work.

Measure outputs, not activity

Don't track whether you or your team are "using AI." Track whether the work is getting better, faster, or higher quality. Concretely: how many proposals are you getting out per week? How fast are you responding to leads? How many customer issues are resolved on first contact?

Use AI to speed up hiring, but invest in the human stuff for retention. People stay when they can see a future. AI will not retain your people. You will.