Stop marketing emails from promoting products a shopper cannot buy in their size.
The Loyalty Loop Campaign Agent
Run more personalized loyalty campaigns without the build time that personalization usually costs.
An agent that builds loyalty campaigns from approved templates, then picks the best offer for each customer at send time.
What the agent does
Analyzes customer data, sets up campaigns from approved templates, and picks each customer's best offer at send time.
- Built by
- Talon.One's team at Bloomreach's Composable AI Hackathon 2026
- Contributed by
- Talon.One (Built the agent at the hackathon)
- Industry
- Retail
- Audience
- B2C
- Category
- Marketing & Personalization
01
The problem
Why this exists
Loyalty marketers want exciting campaigns at scale, but personalization and speed pull against each other. Building and analyzing reports to understand members takes time before a campaign is even started, the build spans coupon, loyalty, messaging, and redemption tools, and every extra variation adds more work.
Read more
Classic next-best-action engines need large volumes of interaction data to learn, so their early decisions are close to random.
02
Outcomes
What this demonstrates
- The agent answers analysis questions, such as which product categories a segment engages with most, from live customer data.
- Incentive rules and send scenarios are created from templates marketers trust, with approval at every step.
- At send time, each customer gets the offer the agent judges best for them, in under a second.
Read more
- Offers are personal from day one because the agent reasons over customer facts rather than waiting for training data.
03
How it works
Architecture
The agent has two parts. The core agent connects to Bloomreach and Talon.One over MCP; it runs data queries in Bloomreach's EQL and builds campaigns and scenarios in both platforms from approved templates, which the marketer reviews and approves.
Read more
At send, a Bloomreach scenario iterates through the target segment and calls a Cloud Run function with each customer's context. The function retrieves that customer's available offers from Talon.One, uses Gemini reasoning to pick the best one, and returns it. A webhook creates the coupon and the email goes out.
Redemption runs through Talon.One's standard Shopify connector, which validates the coupon and applies the offer. Feeding results back into the customer profile is the team's planned next step.
Watch the demo (recorded by the Talon.One team for the hackathon).
Open standards used
Model Context Protocol (MCP)Webhooks
Tech stack
Talon.OneBloomreachGoogle GeminiGoogle Cloud RunShopify
Everyone credited on this solution
Next step
Request an intro
Talon.One is behind this solution. Ask about personalizing loyalty offers at send time, 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.
Related solutions
Let shopping agents offer the right incentive without blanket codes that give away margin.
Otto Deal Concierge
Keep hesitating shoppers by responding in the same session, not in a batch email after they leave.
Real-Time Customer Retention 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
- Built in a hackathon on a demo store, so there is no measured lift in redemption or revenue.
- Results do not yet feed back into the agent's choices; the team names that loop as the next step.
- There is no public code or live demo; the recorded demo video is the public record.
Where this came from
Written from the team's demo video and summary in the demo album for Bloomreach's Composable AI Hackathon 2026.
Status: Reference prototypeAdded 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 Agent Solutions Hub to decide what to build next. Submissions are open to MACH Alliance Members, Collaborators, or Ambassadors.