Shopping beyond the website: Peter Sommer on Shopify's growth, AI translation, and agentic commerce

Most merchants think about platform migration as a technical project. Peter Sommer sees it as a business model reset.

Peter Sommer is the CEO of Dtails, a Shopify Plus partner working exclusively with Shopify ecommerce since 2016. Before joining Dtails, he spent four years as a Partner in IBM’s consulting division. He joined us as part of our Hello Retail Conversations series, where we talk to people shaping how ecommerce works.

Watch the full conversation with Peter here →

The 80-20 rule, in reverse

The financial argument for Shopify migration is consistent across the customer conversations Peter describes. Merchants arrive spending 80% of their IT budget on maintaining their existing platform and 20% on building new features. The ambition is to flip that ratio entirely.

“They’re looking for speed. They’re looking for self-service. And on top of that, you will get a lower cost.” — Peter Sommer

The business case is not just about switching costs. It is about what becomes possible on the other side. A platform sustained by more than 3,000 software engineers pushing over 100 new features every six months handles the recurring business problems through community-driven development. Integrations, payments, AI capabilities: these evolve as the ecosystem evolves.

Peter has not found another software platform that has grown with the same pace and the same stability over the years he has been part of the Shopify partner community.

“It’s never boring being part of the Shopify partner community.” — Peter Sommer

Curating 100+ features twice a year

Every six months, Shopify marks what arrived and what is coming with a partner celebration. For a consultancy like Dtails, the work is not to implement everything. The work is to filter.

What makes that filtering credible is that the entire customer-facing team at Dtails has been on the merchant side themselves. When new features land, they can assess quickly what is relevant for a given market, a given industry, or a specific type of customer.

“We take great pride that we have all been born on the other side of the table. So when they get new features, they can quite fast find out what’s relevant for this market, what’s relevant for that particular industry or customer.” — Peter Sommer

That filtering is what a partner adds that a platform alone cannot. Shopify’s roadmap is ambitious. Its relevance to any specific merchant is always partial. A team that has worked as merchants knows which slice matters.

AI translation that holds tone of voice

The most concrete case Peter brings from recent client work is internationalization. A manufacturer of high-fidelity audio equipment had built 25 years of product texts, blog posts, FAQ articles, and community content in English. Translation had always been too large a task and the business case had not justified it. So everything stayed in English.

The German pilot changed the calculation. The question was whether AI could hold tone of voice, not just convert words.

German brings a genuine cultural tension: the formal Sie versus the familiar du. The brand did not want to take sides and risk alienating any part of their audience. Dtails structured the translation to use a third-person pronoun throughout. The LLM applied it consistently across the full content library.

“Within a calendar month, we were actually able to translate all of their content.” — Peter Sommer

What made the scale possible, Peter points out, was Shopify’s data model. Clean extraction of product data, articles, blog posts, and page content. Structured input to the translation layer. Quality review. Structured re-import. The architecture of the platform is what makes this kind of automated workflow practical, not theoretical.

When native Shopify search is not enough

Shopify has improved its search and recommendations engine significantly over recent years. For a broad range of merchants, it is sufficient. Peter sees a clear pattern for when it stops being enough.

Two factors come up consistently. One is language: handling the way customers actually type queries, including semantic distance and intent, not just keyword matching. The other is assortment size: large catalogs surface the limitations of a generic engine earlier and more visibly.

The business case sharpens when you account for what manual curation actually costs. Shopify’s native setup requires ongoing configuration and maintenance. Automated recommendations eliminate much of that overhead. The question becomes: what is the cost of the automation relative to the lift in conversions and the staff time recovered?

“The larger the customers are, the bigger their product assortment is, and the further up they are in their advancement of being a mature e-commerce merchant, the more you need tools like Hello Retail to fit the needs of their customers.” — Peter Sommer

The printer story

The section of the conversation Peter uses most precisely is a personal one. He was asked to help his mother-in-law buy a printer.

He knew the criteria: wireless, color laser, multifunction, duplex. He typed into Google. The results mixed color and monochrome printers, and he spent time on pages that looked plausible but were wrong. He switched to an LLM. He put in the four criteria. He got five results that matched. He narrowed to two. Then he went back to Google to find the right vendor with the right warranty and local shipping terms.

“I found myself actually getting the advice where I simply put up five criteria. It came back with a range of five printers that matched those criteria.” — Peter Sommer

This is the division of labor he sees taking shape. LLMs handle the structured advisory role: taking criteria, matching them against a product space, narrowing options. Traditional search performs better once the buyer knows the specific product and is comparing vendors and delivery conditions.

The implication for merchants is directional. The brands that appear in the LLM advisory phase as well-documented, credible options are positioned differently from brands that depend entirely on paid search and keyword ranking for discovery.

Agentic commerce is already live

Shopify has launched what it calls agentic commerce: a product catalog spanning all participating merchant stores and a universal cart. A shopper can assemble items from multiple stores and complete a single checkout transaction.

Peter draws the analogy to a department store. You browse across floors, gather what you want from different vendors, and settle the bill once at the exit. That experience, which has always been normal in physical retail, is what Shopify is making structurally possible across online stores.

“They’ve launched a universal cart, which enables a completely new shopping experience.” — Peter Sommer

The feature set is currently available for merchants shipping to North America. Peter’s conversations about it are primarily with customers who have that as a significant market. But the directional signal is clear regardless of geography.

His advice to customers is consistent with how he approaches any new channel: understand what it is good for, identify what your customers are using it for, make the step, and harvest the value before moving to the next one.

What actually changes

Peter’s framing for what merchants should do now is not a five-year projection. It is a near-term sequence.

The first step is data readiness. LLMs can only give good advice about products they can understand. Product data that is incomplete, inconsistent, or structured for legacy keyword-matching will not surface well in an advisory context. That is not a new problem. The stakes have raised.

The second is brand trust. The era of outsmarting Google’s algorithm through keyword density is over, Peter says. What replaces it is real content, authentic voice, credible user feedback, and communities that form around products organically.

“The good will become the victory of the future. Because if you do good and you do well and you excel, the trust communities that are out there, the big changes that any customer will know about that through the AI and the advisory.” — Peter Sommer

The third is loyalty. Customers who have a relationship with a brand do not need to be re-discovered by an algorithm to return. That structural advantage only becomes more valuable as discovery channels fragment across platforms.

What Hello Retail brings to this

Peter’s point about data readiness connects directly to what Hello Retail’s Product Intelligence is built to do.

The behavioral layer, meaning what a shopper viewed, how long they engaged, what they are likely to need next, turns a merchant’s raw product catalog into a working model of customer intent. That model is what makes automated recommendations accurate rather than generic.

Product Intelligence surfaces patterns across the assortment that a merchandiser reviewing manually would miss. For a large catalog, those patterns are the difference between search that finds what the customer meant and search that returns only what the customer typed.

Product Agents extends this into triggered communications. Price drops on viewed products. Replenishment signals based on purchase cadence. Back-in-stock alerts for items a shopper saved. Every message grounded in product intelligence and timed to the customer’s actual situation, not a broadcast calendar.

As agentic commerce matures, the merchants who surface accurately in an LLM advisory context will be those with structured, complete, and behaviorally informed product data. That is the data layer Product Intelligence builds.

Final thought

Peter’s closing is measured. He has been in the industry long enough to have watched voice commerce announced as the next transformation of retail, and has not made a single purchase through a smart speaker in the years since.

His advice is not to project where the market goes in five or ten years. It is to make the next step clearly, then the step after that, with a clear read on the value available right now.

“I’m more into the next step and the next step after that. Look at reality today and then harvest the value.” — Peter Sommer

The channel shift will happen. The pace will be uneven, and the projections will be wrong in ways that are hard to anticipate. The merchants best positioned for it are the ones with clean product data, authentic brand content, and loyal customers who know who they are regardless of which platform surfaces them next.

Watch the full conversation with Peter Sommer on the Hello Retail Conversations page →