cd /news/ai-agents/how-retailers-can-turn-interactions-… · home › topics › ai-agents › article
[ARTICLE · art-173358] src=sierra.ai ↗ pub= topic=ai-agents verified=true sentiment=↑ positive

How retailers can turn interactions into lasting customer relationships

Sierra is developing the Personal Agent Protocol, or Poppy, with more than 100 businesses — including Muse, Meta's AI agent, Genesys, Shopify, Stripe and Walmart — as an open standard governing how consumers' personal agents interact with retailers' own agents. Sierra positions Poppy around choice and control: consumers decide what their personal agent can access, while retailers decide what it can do on their site, such as allowing a return but not a change to payment details. Sierra also pitches its Context Engine, which links a retailer's catalog, commerce systems and customer records with what shoppers tell the agent, as the way retailers turn shopping conversations into compounding customer context.

key takeaway

Sierra's Poppy protocol splits permissions between consumers and retailers — a personal agent might be allowed to start a return but not change payment details — so retailers should scope agent capabilities per task rather than granting blanket access.

by read3 min views3 publishedNov 20, 2026
How retailers can turn interactions into lasting customer relationships
Image: Sierra (auto-discovered)

Retailers were the earliest of early adopters of conversational AI, and they started with customer service: tracking orders, processing returns, handling exchanges. Over the last year, that ambition has grown quickly, from reactive support to proactive discovery that helps shoppers find what they love and fills bigger baskets.

Shoppers usually know roughly what they want. The hard part is narrowing hundreds of options down to the two or three that will truly work. Websites and apps are great in so many ways, but the choice can be overwhelming and the grids of products fiddly to navigate.

Shopping agents change all that. The conversation (“show me red dresses for my anniversary”) becomes the interface. It’s one you don’t have to learn, and it adapts to what you say (“I like that one, but will my size arrive in time?”). And once you’ve found the perfect red dress, the same agent can track the order, handle an exchange, or find the shoes to go with it. Every conversation deepens the relationship between a retailer and its customers.

Starting the conversation #

Most shoppers still won’t open a chat window on their own. But the ones who do have a better experience and are more likely to buy. So agents shouldn’t wait to be asked. They should start the conversation.

Picture someone looking at a rain jacket. They see a suggested question, “Will it fit over a fleece?”, and tap it. The agent answers using the product details, customer reviews, and what it knows about the shopper’s past purchases. It recommends the next size up, shows that it’s in stock, and adds it to the bag.

One agent before and after the purchase #

For shoppers, browsing, buying, and customer service are one experience. They expect a retailer to remember what they’ve shared, whether they’re making a purchase or returning something. And they shouldn’t have to navigate a company’s org chart to get the basics done. With one agent handling everything, a retailer can help at exactly the right moment. When something won’t fit, the agent can suggest the right size before it goes in the cart. When someone doesn’t love the color, it can suggest a better one before processing a refund, and keep the sale.

Every conversation also teaches the agent something. Sierra’s Context Engine connects a retailer’s catalog, commerce systems, and customer records with what shoppers tell the agent. That context is unique to each retailer, and it compounds over time.

Working with personal agents #

Personal agents like Muse and Instinct are still relatively new, but their use is growing fast, and shopping is one of the reasons people give most often. That raises an important question for retailers: how do you respect your customers’ choices while staying in control of the shopping experience?

We believe the first step is for retailers to build their own agent: a digital front door for their brand that personal agents can interact with, alongside their website and APIs.

We’re also developing the Personal Agent Protocol (Poppy for short) with 100+ leading businesses, including Muse, Meta’s AI agent, Genesys, Shopify, Stripe, and Walmart. It’s an open standard for how personal agents and businesses work together, built on choice and control. Consumers choose what their personal agent can access, and retailers decide what it can do on their site. A retailer might let a personal agent start a return, for example, but not change payment details. And when a task needs more information, like a fit question, the personal agent can work directly with the brand’s agent to get the right answer.

Built on what you already have #

Retailers want agents that look and sound like their brand, so Ghostwriter, our agent for building and improving agents, uses brand guidelines and product knowledge to build them. And soon teams will also be able to track shopping agent performance directly in Agent Studio.

Our goal at Sierra is to help the great retailers of the world build better customer experiences that also deliver better business outcomes: larger baskets, fewer returns, and more growth.

── more in #ai-agents 4 stories · sorted by recency
── more on @sierra 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/how-retailers-can-tu…] indexed:0 read:3min 2026-11-20 · —