Three quarters of the orders Shopify attributed to AI last quarter came from product categories outside its top 100, which is the opposite of what a search box does.
Search rewards popularity and advertising spend, and the long tail loses. Jess Hertz's argument is that an agent buying on a shopper's behalf inverts that, because it is matching a described need rather than a keyword — and that this is where the platform's smallest merchants get found.
"We power over 14% of the US e-commerce market. And what's incredible about that is you can sit at this kind of wide aperture of commerce and you can see how competition kind of shakes out in real time."
Hertz became Shopify's chief operating officer last October after a career that ran through the Obama administration, the Biden-Harris transition and the job of White House Staff Secretary, and her first internal message told the company it had drifted from its core.
The full interview is covered here so you can skip it. 30 minutes of audio, 16 minutes of reading.
Here are the 13 lessons that matter.
👤 Guest: Jess Hertz, COO of Shopify, previously White House Staff Secretary in the Biden administration
🎙️ Host: Bob Safian, who hosts Rapid Response and was formerly editor of Fast Company
📰 Published: 14 September 2026 on YouTube (Rapid Response)
🔴 YouTube | 🟣 Apple Podcasts | ⏱️ 30 min | ✅ Time saved: 14 min
Key Takeaways
AI-attributed orders are landing in the long tail rather than the head of the catalog
75% of them last quarter came from categories outside the top 100
Agents reward precision and personalization where search rewarded popularity or spend
Shopify's own internal agent lives in Slack and has taken 15M actions in five months
It works only in public channels, which she calls a feature rather than a bug
Headcount has been flat for more than eight quarters while revenue grows 34%
The drift her first memo as COO named was the sales team's compensation plan, not strategy
Complexity is the moat: a more complicated commerce world makes the platform more valuable, not less
An agent using the platform's structured catalog converts at twice the rate of one using scraped data
The right shape for an employee is no longer a T but an X — several spikes, learned fast
Merchants who reach $10M in GMV stay: a 97% retention rate over five years
1. A Company, Not a Country
Safian opened on what he called the elephant in the room: a Canadian company headquartered in Ottawa, with most of its customers in the US, operating through a trade war between the two.
Hertz's first answer was a personal one — she was about to move to Toronto, having finally decided to make the pilgrimage, after being based in Washington
On where the tension actually sits: she said the feeling is more outside the company than inside it, and that Shopify thinks of itself as a company rather than a country
"But the tariffs are the big elephant in the room," she said, framing them as one more moment of uncertainty in a company that has operated through many
The example she gave is an operational one rather than a political one: "Even when the first kind of tariffs came into place where we had a team spending the weekend building a duties and taxes calculator." She called it a moment to shine
Who it lands on is the part she was careful about: the macro environment affects merchants, and particularly smaller merchants who may not be able to absorb it without the right tools to adapt
Asked directly whether the Canadian question changes anything, her answer was one sentence: "Tariffs between any countries, we would be reacting the same way."
2. From the West Wing to Tech
Safian pointed to her time in the Biden White House, her legal background and her work on the Biden-Harris transition, and asked whether the current political moment makes her want to go back.
"It doesn't," she said, and then offered what she called a pet theory
The theory is that the two worlds are closer than either thinks: "I do think people underestimate how similar politics and tech are actually." She began her career working for a behavioral economist in the Obama administration
What politics teaches, in her account, is how to drive hard, develop pace, and compress enormous amounts of information in a very short period — a skill she says maps directly onto a moment when AI is about compressing information and reducing the need for human context
She thinks the connection is undersold rather than discreditable: people sometimes look at the crossover negatively, and she reads it as a positive
Safian put the conventional version to her — government is slow, tech has to drag it along — and she said her experience was the opposite
Her last White House job was Staff Secretary, which she described as the central nervous part of the West Wing. It gave her unusual breadth, most people in the role are lawyers, and "But it moves incredibly quickly"
3. The Drift Was the Comp Plan
Safian asked what she meant in her first internal message as COO, which told staff the company had drifted from its core.
The answer is narrower and more concrete than the memo sounded: "So that drift was in the context of our incentive compensation system for our sales team."
Her interest in the subject is academic as well as operational — she worked for a behavioral economics professor in law school, and says how you design compensation plans and systems really does affect outcomes
One of the first things she did as COO was read the sales compensation plan and conclude it was not aligned with the business: "It's not actually aligned with the merchant outcome and our business model of Shopify, which is we only succeed when our merchants succeed."
What the fix was about is where the alignment shows up: "It's actually how we compensate people and put our money where our mouth is," rather than in products, actions or things the company says
4. No Reining In the CEO
Safian said the two of them had been talking before the recording about Tobi Lütke, who has called himself a chaos engine, and asked how much of her job is reining that in.
"Oh no. There is no reining in," she said
She rejects the premise about herself as well: "I think I'm a chaos person perhaps masked as an order person"
On AI specifically she says Lütke has been incredible at steering the company, and the example she reaches for is his March 2025 memo, which made using AI a baseline expectation for every employee
The interesting question to her is not the memo but the implementation: what a baseline expectation actually means, and how you encode it in a company from both a systems perspective and a cultural one
5. River Lives in Slack
The answer to that question is a tool, and it is the most concrete thing in the interview.
"And we have this amazing tool called River. And River is basically an agent that lives within our Slack," she said
How it works, and the constraint that defines it: "And you describe a task in a public channel. It all has to be out in the open in public. So that's a feature, not a bug. And it can read code. It can run tests."
The scale is the part worth writing down: "And it lives in about 10,000 channels in the company and you know about 90% of employees have used it and it's run 275,000 sessions and it's taken 15 million actions in 5 months"
Her theory of why it matters is about information rather than labor: "I do really think the velocity that a company moves at can be judged based on how quickly information is moving within the company." River reads everything in public channels at all times, which drops the transaction cost of finding things out and leaves the company relying less on human context
The second-order effect she names is compounding — "river gets better with every interaction"
6. Where the Growth Comes From
Safian noted Shopify had just posted what he believed was a fifth consecutive quarter of growth above 30%, and asked how much of that is taking share from Amazon or someone else and how much is genuinely AI.
Hertz split it in two: a core business she calls incredibly durable, and AI expanding the upside on top of it
The number she uses for durability is a retention number dressed as a revenue number: almost 90% of quarterly revenue comes from merchants who have been on the platform more than a year. "That growth compounds and it really shows how strong the business actually is," she said
On the agentic side she was explicit that the base is small: agentic GMV is small, and growing
The early evidence is product-level. Sidekick is helping merchants sell more, and merchants who are part of catalog are starting to make more money
"Sidekick is our AI powered commerce agent that lives within Shopify," she said, for listeners who do not follow the company
Catalog is the product layer for agentic commerce — "So that's what makes products understandable, discoverable, purchasable by AI"
The conversion number is the one that makes the case: when a buyer arrives from an agentic channel, the conversion rate is twice as high where the agent is using Shopify's catalog data rather than scraped data
7. Complexity Is the Moat
Safian's sharpest question was the platform-risk one: Shopify is building an infrastructure layer for AI commerce, with agentic storefront integrations into ChatGPT, Gemini and Copilot, so how much does she worry those platforms build the layer themselves?
"What I would say is I think we're past that you know SaaS apocalypse moment," she said, and Shopify has come out the other end of it
The reframe is the answer: "complexity is both the challenge and the moat for Shopify." A complex world helps the company
"We've always been a unified operating system for commerce," she said — an aggregator across an online store, in-person retail and now agentic commerce, with buyers arriving through all of them
The argument is that this is hard, and hard is the point: merchants cannot manage that complexity alone
What is scarce in an AI world, in her telling, is not the model but the plumbing: "I think what's even more scarce in this AI world is that infrastructure that you're talking about, right? And building that complete stack for the merchant."
Safian pushed back from the merchant's side. The standard claim is that AI lets small independent businesses compete better; his counter is that it also makes the whole thing more complicated, so they will not be competing on their own at all
Hertz agreed with the description and not the worry: the complexity of a commerce transaction will always exist, the infrastructure around it will keep compounding, and Shopify absorbs it on the merchant's behalf
8. Merchant of Record
Safian's follow-up was whether merchants who need help managing an agent sitting between them and their customer end up feeling trapped.
"I don't think they'll feel trapped," she said
The evidence she offers is a standard rather than a promise: the universal commerce protocol, which she describes as the open commerce standard Shopify developed with Google and the transaction layer of commerce
Under it, the merchant "continues to be the merchant of record"
Her framing of the intent: the products and partnerships are not about disintermediating the merchant, but about letting a merchant run one business that works everywhere commerce happens
9. Tokens, Headcount, ROI
Safian brought in a conversation he had a few months earlier with Uber's COO, Andrew McDonald, who he said ignited a meme about the hidden costs of AI — money spent on tokens with no clear return — and noted that Hertz's team can still access tokens mostly without restriction.
"Yes, I think we see a very healthy ROI," she said
She drew the distinction the meme is really about: "Now, look, the idea of adoption is never going to be the same thing as impact." The company is working on ways to measure impact
The financial proof she gives is a ratio rather than a token bill: "headcount for Shopify has been flat for more than eight quarters and while revenue is growing 34%"
On cost control, the policy is deliberately loose: "We have a pretty open and flexible view and we want engineers to be able to use the best models for the best circumstance."
"Obviously, we have thoughtful speed bumps built in along the way," she added, with constant monitoring
Her own summary of what all of it is for is productivity — more revenue with the same number of people
10. T-Shaped to X-Shaped
Safian asked her to unpack a line of hers about the ideal employee having changed shape.
The old model was the T: "So, you have lots of breadth, but you still have this one vertical that you go deep in, which is, you know, your craft, right?" Hers would have been legal or regulatory
"AI is really pushing towards what I'm calling X-shaped people," she said — several spikes of expertise, acquired much faster, with the ability to absorb complexity and learn new areas
Her examples are modest by design: being a seven-out-of-ten designer if you want to be, or in her case doing her own data work
The part she thinks is underrated is not the individual but the fit between them: "It's actually that team combination that becomes the accelerant to incredible amounts of work that you couldn't even, you know, imagine." The work is finding where people's edges are and how they fit together
Safian's reading was that this sounds like a generalist, until the team part — which makes them generalist specialists
Hertz took the correction and extended it: "I think a little bit about Tetris maybe it's not that you're a pure generalist and a pure athlete." She still has specialized expertise; what changed is the speed — "I can learn different areas much faster because I have different tools available to me"
11. AI Rewards the Long Tail
Safian asked what Shopify is seeing and doing for small merchants exposed to tariffs, trade wars and oil price spikes.
Hertz gave the platform's share of the market first: "We power over 14% of the US e-commerce market." What that buys, in her words, is a wide aperture on commerce and a live view of how competition shakes out
The first thing she sees from there is a collapsing barrier to entry: "The barrier to entry because of AI has reduced, right? So the ability to start your own business and have your first sale, that time has just compressed significantly."
The figure behind it: "last year it was about 10% of new shops made their first sale within a single day of signing up which is just incredible"
The second is the shift in what gets found. Search tended to reward popularity or spend; agents, she says, are rewarding precision and personalization, and "75% of AI attributed orders last quarter came from categories outside of the top 100, which I find particularly interesting"
Her explanation is that "AI can understand the actual need and not just a keyword"
She made it concrete with her own summer: her eldest daughter was going to a ranch camp in Wyoming for the first time, and the packing list included a trunk that would fit under her bunk. She did not know the specifications, used AI to search, and found a specialist merchant in Wisconsin that has been making them for 40 years
That merchant, she said, is exactly the kind AI can help people discover
12. Nobody Outgrows Shopify
Safian raised the classic trap for a service business: the pull upmarket toward bigger customers, and the risk that your best small customers outgrow you and leave for a platform built for them.
Hertz's answer is that the range is the product: the company wants to serve merchants from the kitchen table up to Kim Kardashian and Skims, and the infrastructure is designed for exactly that
The retention numbers are her evidence, and they go up with size: "You know, we see over the last five years, merchants that reached a million in GMV had a 92% retention rate and then merchants that reached 10 million had a 97% retention rate"
Safian pressed with a running brand whose chief executive he had recently met, one he expected to grow past what Shopify could do
Hertz's reply was that the company is already a Shopify merchant, that it has had great success there, and "We actually don't think that they're going to outgrow Shopify"
Her explanation for why is architectural and dates back years, to the choice to build a single unified data model: "We make it extensible. It's how a large brand can run, you know, a thousand different campus stores, 1,200 storefronts, you know, six million SKUs on one stack"
Asked again whether she worries about losing them as they grow, she inverted it: "I think it's a little bit reversed. We want to be able to help them grow."
13. Keep Identity Lightweight
Safian's last line of questioning was about a self-description of hers — a deeply unsentimental person — and whether that is an asset or a hindrance.
"It is a it is perhaps both an asset and a hindrance," she said, describing it as first-principles thinking
The company phrase she took to immediately was "keep your identity lightweight" — the idea that you can wake up smarter and better the next day, and that you do not have to get attached to ideas
Keeping things non-personal and almost irreverent is what makes it work, and she says it has let her pivot repeatedly in her career without carrying the baggage of having always done something a certain way
Safian's read was that first principles are what she is sentimental about — an anchor that makes the chaos navigable, where the route can move but the destination does not
Hertz agreed and pushed back on the word chaos itself: "I don't actually feel like we're a chaotic company and so there is a deep focus and I think it's because of this purity of mission which is really an alignment mechanism." With that anchor, tradeoffs get made without the company feeling chaotic
The company, on her account, does not feel chaotic from the inside: it feels curious, open and innovative for the right reasons, rather than attached to what an industry is doing or to internalized best practice
Asked what is at stake for Shopify now, she gave the same answer she says was true at the start: "It is to be the 100-year company and to grow entrepreneurs everywhere and I think that is the bet and we are going to win or die trying"
Bonus Insights
On leaving Washington for Toronto, she said she gets energy from being with people in person and is not too worried about not having roots in the US
Safian closed the episode with a reflection of his own on the word sentimental, arguing that Hertz is not using it in its usual emotional sense — the one tied to empathy, which he says all good leaders need — but to prioritize agility and egolessness
"The pace of change today means we don't have the luxury of being nostalgic," he said. "And for all of us, ego can be a bottleneck to learning"
His closing line: "Instead of trying to quiet the storm, sometimes our best choice is just to get comfortable with the rain"
Safian's own framing of the interview at the top was that Hertz is forthright about how chaotic the environment is, and about when it makes sense to bring order to chaos and when to let it ride
Hertz's bottom line is that the agentic shift is not a threat to the platform layer but the thing that makes it scarce, and that a company which absorbs complexity for merchants — while keeping its own headcount flat and its identity lightweight — is the one that gets to keep the long tail it just made discoverable.
Products, Companies & Tools Mentioned
Shopify (Powers over 14% of the US e-commerce market, has posted five consecutive quarters of growth above 30%, and draws almost 90% of quarterly revenue from merchants of more than a year)
Sidekick (The AI commerce agent inside Shopify, which she says is helping merchants sell more)
River (Shopify's internal agent, living in Slack across about 10,000 channels, used by around 90% of employees for 275,000 sessions and 15 million actions in five months)
Slack (Where River lives, and where the public-channel-only rule makes the company's information legible to an agent)
ChatGPT, Gemini and Copilot (The agentic storefront integrations Safian used to ask whether the big AI platforms will build the commerce layer themselves)
Google (Co-developer with Shopify of the universal commerce protocol, the open standard under which the merchant stays the merchant of record)
Skims (Her example of the top of the merchant range, against the merchant starting at a kitchen table)
Uber (Its COO, Andrew McDonald, is credited by Safian with igniting the meme about the hidden costs of AI tokens)
Books & Resources Mentioned
Reflexive AI usage is now a baseline expectation at Shopify – Tobi Lütke (The March 2025 memo Hertz calls the example of how a chief executive steers a company on AI, and the thing she had to encode into systems and culture)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: