Case study

Bonnoo. Shared demand, coordinated.

Bonnoo brings shoppers together around the same product so their combined demand can earn a better offer. We built the AWS platform that carries a request through matching, negotiation and a final decision, with people in charge of the commercial terms.

Practice AWS Cloud · DevOps Engagement Platform build · collective buying · generative AI
Designed capacity
60K+
Monthly active users in the sizing model; not a reported user count.
Delivery
12 weeks
From discovery to go-live, followed by hypercare.
Proof of concept
6 weeks
Pilot groups with test suppliers exercised the full journey.
Languages
2
Arabic and English, including Egyptian dialect and Franco-Arabic input.
The brief

Demand only helps when it reaches the same group.

Shoppers can describe the same product in different languages, spellings and levels of detail. Without matching, demand fragments across separate requests and loses its negotiating value.

Groups also need reliable deadlines, trustworthy participant counts and clear offer terms. Joining a group is free and non-binding; a shopper decides whether to buy only after Bonnoo publishes the final offer.

What we did

Give each stage an explicit next step.

Match requests before opening another group

Amazon Bedrock turns free-text requests into structured product descriptions, and embeddings help find existing groups. The same flow supports Arabic and English input.

Run deadlines as part of the workflow

EventBridge schedules and SQS queues drive thresholds, negotiation and decision windows. Each stage can retry independently, with failed work retained for investigation.

Keep negotiation accountable

AI prepares supplier demand briefs and compares quotes. Bonnoo's team approves the terms, and shoppers choose whether to accept through the payment provider's checkout.

Protect the demand signal

One-join-per-account controls and bot protection keep participant counts credible. Aurora holds the durable record, while live counters, reconciliation and duplicate protection support the storefront and order flow.

The outcome

A buying group the team can operate end to end.

The delivered platform includes request matching, group stages, staff-approved negotiation, decision windows, payment and courier integration, and an assistant grounded in live platform data. Pilot groups ran through the full flow before the production build-out.

The platform is sized for more than 60,000 monthly active users. Targets include 99.9% storefront and API availability and at least 90% correct duplicate-request matching; these are operating objectives, not reported performance measurements.

The brief

The person you brief is the person who solves it.

One conversation · 30 minutes · no monologue