Key takeaways
AI made content cheap to produce, and cheap content creates demand for far more of it.
In agentic commerce, AI agents read product content for facts, and they skip or guess when facts are missing.
The Amplience Content Operations Platform is designed to run content, assets, workflows and AI as one operation.
People set the direction and approve the results, with the platform taking on the coordination in between.
Studio One at Abbey Road is the largest purpose-built recording studio in the world. It’s where the London Symphony Orchestra recorded John Williams’s scores for Return of the Jedi and Raiders of the Lost Ark. (A different John Williams. Ours builds technology, not film scores.) Before any session starts, the orchestra tunes up. Every musician finds their note at once, and for a minute the room fills with sound that’s busy, loud and completely unresolved. Then the conductor raises the baton, and a hundred separate instruments become one piece of music.
Running commerce content feels a lot like that first minute right now. There are more channels than ever, more markets and more formats, and a growing number of AI tools, each producing something useful on its own. There’s more activity than any team has ever managed, and very little of it moves in time.
In September 2026, we brought 150+ retail leaders into that same building for Amplify: Agentic Commerce Live, to introduce what we’ve built to bring it all back in time. This is the story behind it.
How is AI changing content operations?
AI has made content cheaper to produce, and cheaper content creates demand for far more of it.
For a long time, the rules were simple. You built a store, made the content, published it and waited for customers to arrive. Content supported the sale. It explained the product and made it look good, and content teams could keep pace.
Then AI arrived and made the writing, translating and resizing dramatically cheaper. Everyone assumed the content bill would shrink.
“It’s done the opposite,” said Jeremy Straker, Co-CEO and CCO at Amplience. “When content gets cheaper to produce, you don’t make the same amount more efficiently. You discover a hundred new places you need it. Economists call it Jevons Paradox.”
Those hundred new places add up quickly. A single product in eight colorways, across 15 channels and formats, in four regions, now needs more than 480 versions of its content. Every one of them still has to be briefed, checked, approved and delivered.
So teams did what teams do. They added tools, then agents. One agent wrote descriptions, another handled translations. The tasks got faster, but someone still had to hand the work from one tool to the next, chase the approvals and make sure the right version went live.
John Williams, our Co-CEO and CTO, has a simple way of describing where that leaves teams. AI can play every instrument faster, he says, but people are still the orchestration layer holding it all together.
That coordination is where the money goes. Our CORE Report models annual content operations spend at $33M to $70M for a large-catalog retailer, and much of it goes on the work around creation, like briefing, localization, approvals and handoffs.
How are AI shopping agents changing what product content needs to do?
AI shopping agents read product content for facts rather than feelings. When the facts aren’t there, they guess or move on to someone else.
The audience for all that content was changing too. Chuck Gahun, Vice President and Principal Analyst at Forrester, has spent the past year measuring this shift toward agentic commerce, where AI agents research, compare and increasingly buy on behalf of people. He describes the difference between search and answer engines as the difference between a librarian handing you a book and an expert simply telling you the answer. For everyday shopping questions, people increasingly prefer the answer. Forrester found that four in five regular answer engine users already research products there, and three in five discover new products that way.
These new shoppers don’t behave like the old ones. Chuck’s research describes human shoppers as driven by emotion, browsing experiences designed to persuade them. AI agents are driven by logic. They retrieve what they need, work within the limits of a context window, and have no patience for gaps. When product content is incomplete, an agent either fills in the missing details itself, which risks hallucinations, or it moves to a competitor whose content is complete.
Chuck illustrated what that means with a story he says is already playing out. A head of digital commerce shelves a project to get their content ready for agents, while two competitors triple their revenue. When the team finally looks, they find agents stuck on pages they can’t read and leaving without prices, specs or return policies. They find product content on third-party channels that’s out of date. And when they ask how long it will take to fix, the answer is weeks, if not months.
The team in that story knew exactly what was wrong. What they couldn’t do was move fast enough. The problem sat in the operation behind the content, which was built for one campaign at a time and never for every product, on every channel, for an audience that never stops asking.
What is the Amplience Content Operations Platform?
The Amplience Content Operations Platform is the platform we’ve built to run that operation as one. It brings our CMS, DAM, Dynamic Media and Workforce together on one composable foundation, so content, assets, workflows, people and AI can work as a single system, from the first brief to every channel. We gave the first preview at Amplify.
We could have built another assistant. It was the obvious move, and we did think about it. But another assistant would have given teams one more tool to manage without changing who holds everything together.
The thinking that shaped the platform came from John. “Enterprise applications were designed on the assumption that the person operating the screen was also doing the work,” he said. For decades that was true. It stops being true when AI is doing much of the work. Nobody should have to sit and watch a product description generate.
So instead of building software that you operate step by step, we’ve built a platform that you direct.
How does AI content orchestration work?
AI content orchestration connects AI agents, your systems and your people in governed flows. You describe the outcome you need, and the platform coordinates the steps to get there. Working this way should feel like handing over a brief and getting finished work back to review.
Picture a Monday morning. Your autumn campaign needs to launch in France by the end of the week, across homeware and fashion. In the old way of working, that means a week of briefs, handoffs and status checks.
With the Amplience Content Operations Platform, you describe the outcome you need. It works out what that means, including the product descriptions to write, the imagery to prepare, the pages to localize and the order to do it all in. Then it gets on with the work.
While it does, you get on with the work only you can do. That might mean agreeing next season’s priorities with merchandising, deciding which markets to push hardest, or digging into what last month’s campaign taught you.
Before anything goes live, the platform also picks up on what French shoppers and their AI assistants are asking this season, like how to layer for changing weather or how to make a home feel cozy for autumn. It researches and writes a buying guide for each question, so the guides are ready to launch with the rest of the campaign.
By midweek, the work is waiting for your review. You approve what’s ready, send back what isn’t, and the campaign launches on Friday as planned. The work kept moving while your attention was on the bigger decisions.
John sums it up as direct, delegate and review. The platform is designed to take on the coordination that used to fill your week, so you can keep the decisions that need your judgment.
Can AI content automation work across a whole product catalog?
Yes. AI content automation can work across an entire catalog when it follows the same rules as your best copywriters, and The Very Group is already doing it.
The Very Group’s assortment had grown to around 150,000 products a year from about 900 suppliers, while its copy team stayed the same size. The gaps got filled with supplier copy, which covered the products without really describing them, and its previous content provider took months to deliver each product type. So the team turned to Amplience Workforce, which is available today and will be part of the new platform. The team built a flow of more than 80 steps that captures every SEO, brand and legal rule its copywriters apply.
Now its editors review each product line in about two minutes, and when the Fashion catalog launched, it was fully live within 36 hours. The Very Group also scores product copy out of 50 using its own internal quality model. Copy from its previous provider typically scored between 14 and 18. Copy from its Workforce flow now scores between 38 and 41.
How do you govern AI-generated content?
You set the rules before anything is generated, review what comes out, and keep a full record you can check and reverse afterward.
That matters more than many teams realize. Forrester found that 61% of AI decision-makers believe genAI always gives the same output for the same prompt, and 76% believe it’s good at checking facts. Neither is safe to rely on when your brand is on every page.
The platform is designed to keep people in charge at every step. Brand and tone rules shape the work before it starts, and reviewers approve it before it goes anywhere. Every change leaves a trail you can follow and undo. Workforce already works this way today, and the new platform carries the same controls across everything. When an operation is about to update thousands of pieces of content, it will tell you exactly how many and wait for you to say yes.
In John’s words, “Agentic doesn’t mean uncontrolled.” Nothing publishes itself unseen.
Back in Studio One, what turns a room full of instruments into a film score has never been a better violin. It’s everyone playing from the same score, in time. Commerce content works the same way. The teams, tools and agents were never the problem. They just need to play together.
John puts the shift another way. “As commerce becomes machine-mediated, content becomes infrastructure.”
If you’d like to see the Amplience Content Operations Platform, book a demo and we’ll show you how it could work across your catalog, markets and channels.