Gift shoppers who cannot phrase a keyword search get a curated recommendation out of a conversation instead.
Jason, Therabody's AI Shopping Assistant
Guided, sourced product recommendations on the storefront, live in under 90 days and owned by the brand end to end.
A storefront assistant that asks what a shopper needs, recommends one wellness device, cites the research, and checks out in the chat.
What the agent does
Asks about goal, budget, and lifestyle, recommends a product, explains the research behind it with sources, then adds to bag and checks out in the chat.
01
The problem
Why this exists
Therabody sells at the meeting point of consumer goods and medical devices, and it operates as an FDA-regulated manufacturer. Shoppers arrive with very different goals: recovery for a training routine, relief from everyday aches, stress or restless sleep, or anti-aging. In that setting a wrong recommendation costs more than a returned product, because it can steer someone away from the solution they came looking for, and shoppers want confidence that the right evidence supports the product they are shown.
Read more
After evaluating the market, Therabody concluded that turnkey tools built for general audiences did not give it the level of control its category requires over product claims. The company also sells through more than 10,000 retail doors, so it carries customer-experience obligations both as a brand and as a retail partner.
02
Outcomes
What changes when this is in place
- Launch in under 90 days, configured around Therabody's own products, customers, and compliance rules.
- Shoppers get a guided recommendation with the research behind it cited in the chat.
- Add to bag and checkout complete without leaving the conversation.
Read more
- Custom dashboards track success metrics without the company building a new analytics team.
- Therabody is among the first flagship brands to launch the Agentic Shopping Assistant on AWS.
03
How it works
Architecture
Jason runs on Therabody's own infrastructure on the Agentic Shopping Assistant on AWS, with Anthropic Claude Haiku 4.5 as the model behind its answers. The chat surface is the branded Chat to Shop button on Therabody.com. Commerce actions are handled by Shopify, so a shopper can add an item to the bag and check out inside the conversation.
Read more
Custom context and guardrails bound what the assistant may say about product claims and are revised as Therabody iterates. Answers are grounded in Therabody's own research, with the sources cited in the chat next to the recommendation. Custom dashboards carry the success metrics, so tracking performance did not require a new analytics team.
Tech stack
Agentic Shopping Assistant on AWSAnthropic Claude Haiku 4.5ShopifyCustom context and guardrailsCustom dashboards
Next step
Request an intro
AWS is behind this solution. Ask about deploying a compliance-aware shopping assistant on your own infrastructure, and the MACH Alliance will put you in touch with the person who can answer.
Not ready to talk to anyone? MACH Insider is the Alliance mailing list. New solutions, program openings, and the research behind them, once a month.
Everyone credited on this solution
- AWSShopping assistant platform the assistant runs on
Every organization named here has a page listing the rest of its work on the Hub.
Related solutions
Merchandisers act on performance questions without leaving their workspace, and nothing the agent proposes applies without approval.
Coveo Merchandising Agent
Take a shopper from search to confirmed order in one conversation, personalized from the first message.
Valtech Agentic Commerce Agent
Verification
Contributor submission
The detail on this page comes from the contributor's own account of their own work. The MACH Alliance reviews a submission before publishing it and does not independently verify the figures in it.
What this page does not establish
- No traffic, conversion, revenue, or support-deflection figures are published for the assistant.
- No measurement of recommendation accuracy, or of how often shoppers act on a recommendation, is reported.
- The work behind the claim guardrails is not quantified, and no review or running costs are stated.
- The account rests on a single vendor case study with Therabody's participation; no independent evaluation is cited.
Where this came from
Drawn from AWS's customer case study, October 2026, and Therabody's launch announcement of October 21, 2026.
Status: Live in productionAdded Oct 29, 2026A listing is an editorial decision, not a MACH Alliance certification.
Built something worth showing?
Put your agentic solution in front of the enterprise leaders who use the hub to decide what to build next. Submissions are open to MACH Alliance Members, Collaborators, or Ambassadors.