How Therabody built an AI shopping assistant in 90 days with AWS

A new reference deployment shows how regulated consumer brands can launch agentic commerce without surrendering the customer relationship to a general-purpose AI platform.

Retailers face a dilemma with AI shopping: turnkey SaaS assistants are fast to deploy but generic, while custom builds take months and strain compliance. For an FDA-regulated wellness brand, a wrong product recommendation is more than a return — it undermines trust in science-backed claims. Therabody solved this by building Jason, an AI assistant on the Agentic Shopping Assistant on AWS, live in under 90 days. Softprom is the official partner of Amazon Web Services.

What was announced

Therabody announced the launch of Jason, an AI-powered shopping assistant on Therabody.com, built with the Agentic Shopping Assistant on AWS. Jason asks customers about their needs — pain relief, athletic recovery, anti-aging — along with budget and lifestyle, then recommends a Therabody product and explains the clinical evidence behind the choice, with citations accessible directly in chat.

Therabody is among the first flagship brands to deploy the service. The assistant was configured around Therabody's product catalog, customer base, and compliance requirements, and shipped in under 90 days. It integrates with Shopify to enable add-to-bag and checkout inside the chat.

Agentic commerce shouldn't mean settling for a one-size-fits-all tool. With the Agentic Shopping Assistant on AWS, Therabody had the control to build exactly what it needed in under 90 days

Why this matters

For CIOs, CDOs and digital commerce leaders in regulated categories, the deployment is a reference pattern: retailer-owned infrastructure, vendor-neutral LLM orchestration, and custom guardrails that keep product claims accurate. Therabody sells direct and through more than 10,000 retail doors, so the assistant must handle both brand voice and compliance across a sophisticated SKU lineup without a dedicated analytics team.

Technical details

• Platform: Agentic Shopping Assistant on AWS
• Deployment time: under 90 days from kickoff to production
• Hosting model: runs on Therabody's own AWS infrastructure, not a third-party AI platform
• Commerce integration: native Shopify add-to-bag and checkout from the chat
• Guardrails: custom context, source citations, FDA-compliant product claim handling
• Analytics: custom dashboards for success metrics without a new data team
• Learning loop: continuous adaptation based on real customer queries

FAQ

What is the Agentic Shopping Assistant on AWS?

It is an AWS service for building branded AI shopping assistants on the retailer's own infrastructure, with configurable guardrails, product context, and commerce platform integration.

How fast can an enterprise brand deploy it?

Therabody shipped its assistant Jason in under 90 days, including Shopify integration and FDA-compliant product claim handling.

Does the retailer keep ownership of the customer relationship?

Yes. The assistant runs on the retailer's infrastructure with full UX, UI and data ownership, rather than through a third-party AI platform.

Softprom and Amazon Web Services

Softprom is the official partner of Amazon Web Services. Our team supports architecture design, migration and AI service enablement for enterprise customers.

Prepared by the Softprom editorial team based on vendor materials (Softprom DistriFlow). Source: original article.