In boardrooms across retail and e-commerce, there’s a quiet hope that AI will finally fix the hard stuff: slow releases, over-stretched tech teams, and customer service that buckles at peak.
It won’t.
What it will do is force a rethink of what good technical leadership looks like. The CTO’s role is shifting to orchestrating a hybrid workforce; part human, part machine, and figuring out where automation frees people from mundane tasks instead of replacing them.
The Hype We Keep Buying Into
Every few months there’s another headline about an “AI developer” that can be handed a Jira ticket and return production-ready code. These tools exist. They’re clever. They’re also nowhere near as autonomous as the marketing suggests.
In one of our recent projects; a barcode-scanning app used to speed up warehouse operations, we tested one of these agents. It could pick up existing code, interpret the spec, and make small adjustments surprisingly well. But the quality of the outcome depended entirely on the quality of the ticket we gave it. Vague or incomplete specs sent it into endless loops of clarification questions, burning through compute credits while we waited for answers that still needed debugging.
AI has undoubtedly shortened the distance between an idea and a working prototype. But if you think it’s going to ship production-ready features without oversight, you’re in for a slow, expensive lesson.
A CTO’s Job Is Becoming Less About Code
Great CTOs have always been systems thinkers. The difference now is that they need to design systems where the humans and the AI understand each other.
That means investing as much in promptability; the skill of telling AI what you actually need, as in headcount. Ten years ago, a standout junior engineer was the one who could dig through Stack Overflow faster than anyone else. Today, the standouts are the ones who can coax an agentic dev tool into doing the right job first time.
It also means reorganising teams. Instead of growing purely by adding more developers, the CTO has to balance a mix of seasoned engineers, QA-minded testers, and AI agents that are great at repeatable tweaks but useless without supervision.
The CTO’s success will be measured less by the size of their team and more by how well they orchestrate the interfaces between people, AI, and the messy, bespoke systems that underpin retail operations.
When “AI” Is Just a Fancy If-Statement
You don’t have to look far to see why that orchestration matters. Some of the most visible examples sit in customer service.
I recently interacted with Sky’s chatbot - clearly a new AI rollout that was capable of negotiating contract renewals. It wasn’t flawless, but it could interpret intent, respond conversationally, and only passed you to a human when you pushed to cancel outright.
Compare that with the decades-old automated phone system; clunky menus, rigid responses, no reasoning. Both get marketed as “AI.” Only one delivers a modern customer experience that reduces human workload instead of frustrating the customer enough to pick up the phone.
For e-commerce brands, this difference matters. An intelligent chatbot integrated into returns or exchange flows can save a customer interaction and protect a sale. A poorly-implemented one adds friction, drives calls to the contact centre, and tanks NPS.
Choosing and implementing the right tools isn’t something you can delegate to an app marketplace listing. It’s part of the CTO’s evolving remit: understanding the difference between a reasoning engine and a rules-based workflow, and knowing where each belongs.
Efficiency Beats Replacement
Fear of replacement is usually misplaced. In fact, the people who embrace automation tend to move up, not out.
One of our own team, Dom, started as a systems admin and made a habit of automating every repetitive part of his role. Rather than working himself out of a job, he freed up time to tackle more complex challenges, and was promoted for it.
Read more about Dom’s journey here:
That’s the lesson for leaders. The aim isn’t to rip out entire teams in favour of agents. It’s to look at the most painful bottlenecks; catalog enrichment, reconciliation tasks, error-prone manual data entry, and apply AI surgically to remove friction.
CTOs who approach it this way will find they’re not shrinking their departments, they’re levelling them up.
The Real Skills Gap No One’s Talking About
AI isn’t just changing the job description of the CTO, it’s changing the skill profile of everyone who works for them.
Promptability, the ability to translate business logic into something an AI can act on, is fast becoming as valuable as knowing a framework. Meanwhile, many graduates are entering the workforce with little exposure to these tools because their universities banned them.
Leaders can’t assume their teams will naturally figure this out. If you’re serious about efficiency, you’ll need to invest in training people to work with the machines, and to know when not to.
A Future-Proof CTO Looks Different
The AI-era CTO doesn’t need to be the best developer in the room. They need to be the best conductor: balancing humans and machines, knowing when to automate and when to escalate, ensuring that every new tool introduced to the stack delivers more than it disrupts.
For retailers and e-commerce brands, that shift matters. The pressure to do more with less isn’t going away, but the winners won’t be those who chase full automation. They’ll be the ones whose leaders understand that AI’s real power lies in amplification, making good teams faster, sharper, and more consistent.
The question for every tech leader now isn’t whether AI will take their job. It’s whether they’re ready to do the new one.