Like a lot of people who ended up building things on the internet, my interest in technology started with a simple belief: that moving something online would make it better. What had been managed on paper or locked inside a local system could suddenly be connected, fast, and available to anyone. The internet felt like a promise — that the friction and limits of the physical world didn't have to apply anymore. Speed and scale were finally within reach, even for small businesses. And the reason I cared wasn't the technology itself. It was the people on the other side of it. The business owner who wouldn't have to chase down a paper order form. The customer who could see their account and make a change without calling anyone. I got into this work because I genuinely believed that putting things online made people's lives better.
My early work took me deep into the world of ERP software — specifically NetSuite. I found my way into that community in my early 20s, picking up work through an online bulletin board where NetSuite users and implementers connected. I was also doing sales advocacy: writing proposals on behalf of startups to get them a pricing break for their first couple of years, with the expectation that they'd grow into the platform. The pitch was always the same — give this business a chance to run online, and they'll scale. It was exactly the kind of work that matched my belief in what the internet was for. In practice, it was also where I first started to see the gap between what technology was supposed to do for people and what it actually did.
In every NetSuite ERP implementation I worked on between 2007 and 2014, I watched the same drama play out in two acts. In the first act, a business would spend months and significant money bending the software to match how they already worked — custom scripts, custom workflows, consultants layering complexity on top of complexity. In the second act, a different business would take the opposite approach: they would essentially reinvent their processes — and sometimes even their teams — to match how the software was built to run the business online.
Both approaches had real costs. Both created real friction. And in both cases, the same person was conspicuously absent from the conversation: the customer. The end purchaser. The person whose order was supposed to come out the other side of all that complexity, on time, as promised.
That observation — made in my 20s, working with small and medium-sized businesses in the NetSuite community — became the founding principle behind everything I've built since. Technology in commerce should service people from end-to-end: the people growing the business and the people buying from it.
The Missing Person in Every Automation
The solutions I was helping to design and build in those years were mostly created to solve gaps that didn't consider how people actually performed their job at work. The people who implemented ERP software were mostly financial experts and financial consultants — genuinely skilled people solving a genuinely hard problem: making sure the business could be run accurately and repeatably inside the platform. That's not easy. It still isn't.
But the result was that once the financial accounting layer and streamlined transactions were in place, the people in the business still often needed more software in addition to the software they had just bought — and more design thinking from consultants like myself — to actually use it to do their jobs the way they needed to do them online. The solutions and integrations I was involved in building weren't luxuries. They were what made it possible for the rest of the business to function at all.
And the end customer — the B2B purchaser, the retailer waiting on materials, the DTC customer, the subscriber — was still missing from the conversation. Not because anyone intended to leave them out, but because by the time you'd solved the financial layer and then built the operational layer on top of it, there was rarely time, budget, or attention left for the experience of the person on the other end of the order.
This was my reason for wanting to work increasingly more on the website — the online store — as part of the e-commerce implementation. I wanted to create a good experience for the customer that also worked well for the business. When that was done well and the business was strong, growth would come quickly.
What good technology and AI should do is absorb the work that humans have a hard time doing well. Not because humans aren't capable, but because so much of the hardest work in commerce is reactive, unplanned, and unpredictable. A FedEx service outage that's about to impact a shipment before the customer even knows it's coming. A subscriber who just booked a last-minute trip and won't be home when their order arrives. These aren't edge cases. They're Tuesday. Life doesn't run on a fixed schedule, and neither do businesses. When technology takes on that burden (the firefighting, the stress, the constant drain of solving problems that shouldn't exist), humans get back the thing that matters most: the time and attention to do the work only they can do. Build better products. Make better agreements. Explore new ideas. Invent. This is what leverage looks like when it's working. And it's the reason I've spent my career trying to build it. (I could go on about how this also must be observable, and how the systems of trust required to make it work are a topic entirely of their own. One I think about constantly and plan to write about soon.)
What Procurement Taught Me About Promises
I used to ask procurement managers, production leads, and sales operations people a simple question: What would you pay to get your purchase orders delivered on time?
The answer was almost always the same: Anything.
Not because they were dramatic. Because they understood, viscerally, what a broken delivery promise costs. A retailer waiting on materials that don't arrive disrupts production. A subscriber who doesn't receive their box on the day they expected it starts wondering whether to cancel. A B2B purchaser whose recurring order arrives late has to explain to their own customers why they're out of stock. The downstream effects of a single missed promise compound in ways that are genuinely hard to recover from.
What I saw in those years was that the hardest part of procurement wasn't the purchasing itself — it was the unplanned work. The firefighting. The hours spent solving problems that shouldn't have existed if the system had been more predictable. Procurement managers aren't bad at their jobs. The system around them is just unreliable, and unreliable systems create stress, bad outcomes, and a constant drain on the most valuable thing any operator has: their attention.
This is what I mean when I say People-First AI. Not AI that replaces the procurement manager. AI that absorbs the unplanned work so the procurement manager can do the work only a procurement manager can do: build better supplier relationships, negotiate better agreements, find better products, make better decisions.
Building Around What You Can Actually Fulfill
When I built QPilot, I made a deliberate choice that confused some people early on: I put a lot of the configuration surface area around fulfillment timing, not just billing. Processing windows. Lead times. Production cycles. Cut-off times.
The reason was simple. A subscription platform that only thinks about billing is a platform that makes promises the business can't keep. If your fulfillment operation needs three days of lead time and your subscription platform is scheduling orders with one day of notice, you're setting up your operators for exactly the kind of unplanned work I described above — and you're setting up your subscribers for broken promises.
People-First design at the platform level means the software should only offer what the business can actually deliver. If a delivery date isn't achievable given real fulfillment constraints, it shouldn't be an option. Removing impossible choices isn't limiting. It's honest. And honesty, in commerce, is the foundation of trust.
This same philosophy extends to the subscriber portal. The best subscriber experience isn't one with the most options. It's one where every option shown is one the business can actually fulfill. When a subscriber reschedules their next delivery, they should be able to trust that the new date is real. That trust, earned through consistent and reliable delivery, is what drives long-term retention far more than any discount or loyalty program.
The Delivery Date as a Promise
Amazon has trained an entire generation of consumers to expect a delivery date at checkout. Not a shipping date. Not a range. A date. And they fulfill on that promise with remarkable consistency.
Small and mid-sized DTC brands and B2B merchants can't match that without intelligence. The calculation is genuinely complex: processing time, carrier transit time, cut-off windows, regional variation, holiday schedules. Getting it right manually is hard. Getting it wrong, promising Tuesday and delivering Friday, erodes exactly the trust that makes subscription commerce work.
This is why I built Nextime.AI. Not as an operations optimization tool, though it is that. As a promise-keeping tool. When AI can tell a subscriber, at the moment they're placing or managing their order, "Your next box will arrive on Tuesday, March 24", and then actually deliver on that, something important happens. The subscriber stops worrying. They stop checking tracking. They stop wondering whether to cancel. They just trust.
That's People-First AI in its most concrete form. The technology is doing the hard work of calculation and prediction so that both the person running the business and the person buying from it can experience something simple: a promise kept.
The Supply Chain Verticalization Dream
There's a phrase I've heard throughout my career that I've always found compelling: supply chain verticalization. The idea that through shared data, automation, and increasingly AI and robotics, the entire chain, from raw materials to finished goods to final delivery, could become genuinely coordinated. Predictable. Reliable.
For most of my career, this felt like a dream reserved for enterprises with the resources to invest in it. The Amazons and Walmarts of the world. Not the small coffee brand on Shopify, not the specialty manufacturer selling B2B through WooCommerce.
I think that's changing. The combination of API-first platforms, AI agents, and increasingly capable automation tools means that the sophistication that used to require a dedicated supply chain team is becoming accessible to businesses of any size. The Autoship platform and the QPilot API are part of that story, connecting subscription commerce to the fulfillment and delivery intelligence that makes promises keepable.
But the bigger vision is this: as more of the supply chain becomes automated and AI-assisted, the humans in that chain get to do something different with their time. Less firefighting. Less unplanned work. Less stress from a lack of predictability. More time for the things that actually require human judgment: building better products, making better agreements, exploring new ways of doing things, inventing.
Human ingenuity benefits enormously from not having to spend most of its energy solving problems that shouldn't exist. That's not a technology argument. That's a human argument. And it's the reason I believe, genuinely, that People-First AI, done right, is one of the most important things we can build.
People-First: The Ongoing Work
We can focus more on the things that humans want to spend time doing — and also the things we are really good at.
That sentence is the simplest version of everything I've tried to build. But I want to be honest about something: People-First isn't a destination. It's not a design principle you implement once and check off. It's the ongoing work. And as technology becomes more integrated into commerce — as AI becomes more capable, more disruptive, more embedded in the decisions that affect people's businesses and lives — the challenge of keeping people first actually gets harder, not easier.
Every new capability raises the question again. The AI that can proactively reschedule a delivery can also be tuned to make cancellation harder. The automation that removes impossible options can also be configured to remove inconvenient ones. The platform that earns subscriber trust can also exploit it. The technology doesn't decide. We do.
The ERP years taught me that software can either service people from end-to-end (the people growing the business and the people buying from it) or it can demand service from them. The subscription commerce work taught me that automation can either earn trust or exploit it. The AI moment we're in now is teaching me that the stakes are higher than they've ever been, because the systems we build today will shape the defaults that millions of people experience tomorrow.
People-First isn't a constraint on what you can build. It's a commitment to keep asking who the technology is actually serving, on both ends of the transaction, every time you make a decision. In my experience, it's also the only durable competitive advantage. Subscribers who trust the system stay longer, spend more, and tell others. Operators who trust the platform invest in it, grow with it, and don't look for alternatives.
The businesses winning in subscription commerce in 2026 are the ones that figured out, as I did watching those NetSuite implementations in 2007, that the best technology services people from end-to-end: the people growing the business and the people buying from it. That work never finishes. And I wouldn't have it any other way.
Related:
Subscriptions Are Having Their AI Autoship Moment
Autoship Cloud — Shopify & WooCommerce subscription platform
QPilot API — The API powering agentic subscription commerce
Nextime.AI — AI-powered delivery date intelligence