Case Studies | Retail

An AI concierge for bespoke spirits

Agent Build PoC

About

A Retail business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The proof of concept had to prove itself against four focus areas, each tied to a measurable success criterion.

Speed of capture

Clients needed to move from first contact to an usable output in under five minutes. Anything slower breaks the interactive feel that makes a self-serve concierge worth using and pushes customers back towards a manual, human-paced process.

Recipe validity

The agent had to return three genuine bespoke recipes drawn from the client's own flavour library and matched to the client's stated preferences, not generic outputs a general model might invent. Recipe quality is the whole product, so this was the make-or-break test.

Contextual pairings and brand fidelity

Each recipe needed at least three cocktail or serve suggestions contextualised to the client's menu or usage scenario, alongside label mock-ups that incorporated the client's logo and respected the client's brand guidance.

Contained, recorded behaviour

The agent had to stay strictly on task, conversing only about bespoke recipes, and every session had to be recorded so the client could analyse conversations and outputs after the fact.

Solution

The engagement is scoped as a three-phase journey. Phase 1, the proof of concept described here, was delivered inside Cloud Combinator's AWS environment. Phases 2 and 3 are the agreed roadmap that takes the agent into the client's own AWS account and on to a production service. Costs and scope for the later phases are indicative and subject to sign-off.

<5 min

Target client journey from brief to output

3

Bespoke recipes generated per brief

5 wk

Proof-of-concept build and handover

By the numbers:

  • <5 min - Target client journey from brief to output
  • 3 - Bespoke recipes generated per brief
  • 5 wk - Proof-of-concept build and handover
Changes

The proof of concept was accepted against its success criteria, demonstrating that an Amazon Bedrock agent can carry the client's early creative onboarding work and confirming the AWS environment is suitable for continued development. The figures below are the design targets the PoC was validated against.

  • A sub-five-minute journeyCaptured client inputs and returned recipes, pairings and label mock-ups inside the target window, keeping the experience interactive.
  • Recipes grounded in the client's libraryDrew on the client's flavour data rather than generic model output, so every recipe stayed on-brand and usable.
  • Contextual pairings and branded labelsGave each recipe cocktail and serve suggestions matched to the client's menu, plus mock-ups built around the client's own logo and brand guidance.
  • A contained, recorded agentStayed on task and wrote every session to DynamoDB and S3, giving the client the conversation and output data it needs for analysis.
  • HandoverDelivered CloudFormation infrastructure-as-code, a private repository with all code, prompts and example payloads, deployment docs and a knowledge-transfer workshop so the client's team can extend the agent independently.

With the proof of concept validated, the agreed roadmap takes the agent into the client's own AWS account for a MVP and then a production-grade, authenticated client workspace, turning a promising demonstration into a scalable digital ordering channel for bespoke spirits.

AWS Stack

Amazon Bedrock

For the agent and foundation-model inference that generates recipes, pairings and label imagery.

AWS Lambda

For serverless orchestration of the agent's tools and session flow.

Amazon API Gateway

For secure client access to the service.

Amazon DynamoDB

For low-latency, per-session state and conversation records.

Amazon S3

For storage of session artefacts and generated assets.

AWS CloudFormation

For infrastructure-as-code provisioning and a repeatable handover.

YOU MIGHT LIKE

Related success stories

View all case studies

Case Studies | Insights

Utilising Language Recognition, Speed, and Enhanced Security to Make Social Media a Force for Good

  • Here, we take a detailed look at how the Cloud Combinator team collaborated with another cutting-edge AI service provider that provides intelligent systems to “make social media more social” for brands and users alike.
  • Arwen AI is a UK-based startup specialising in AI solutions to manage and enhance brands’ social media interactions. Founded in 2020 by Matt McGrory, Dr. David Cole, and Joel Bailey, Arwen. AI focuses on using AI to automatically detect and remove spam, toxic comments, and other unwanted content from social media platforms.
  • The team at Arwen have three core products. ‘Moderate’ is focused on identifying and removing toxic content from social media channels. ‘Engage’ helps brands identify and engage with meaningful conversations on social media, and ‘Customize’ allows brands to apply bespoke algorithms to their channels - creating an even more effective moderation and engagement.
Read more
CONTACT US

Ready to turn AI into impact?

We'll help you spot the highest-value opportunities, reduce risk around your first AI initiative, and define a clear path to results from day one.

Why talk to us:

Outcome-driven recommendations

AWS-recognised delivery expertise

Risk-aware AI adoption

Clear next step, not a sales pitch

Start with a focused 20-minute conversation about your goals — no pressure, no commitment.

This website uses cookies to enhance user experience and to analyze performance and traffic on our website.

See our Privacy Policy for details.