Case Studies | FinTech

A safe internal assistant in Microsoft Teams

Teams Agent Creation

About

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

Challenge

The challenge had three focus areas.

Answers people can trust

The bot needed to handle both straightforward questions and questions about uploaded documents, returning accurate, contextually relevant answers rather than plausible-sounding guesses, with contextual grounding checks to reduce hallucination.

Safety and privacy

As a financial business, the client required personal information to be detected and redacted, offensive or toxic content to be filtered, and prompt-injection attempts to be blocked, in both what users send and what the model returns.

Access control and data boundaries

The assistant had to respond only when @-mentioned, honour Azure Active Directory group boundaries so knowledge stays partitioned by department, block external collaborators, and connect to AWS privately so data never crosses the public internet.

Solution

Users interact with the bot in Microsoft Teams, where Amazon Q Developer acts as the connector that receives a query and invokes the right workflow. Behind it sits a multi-agent architecture on Amazon Bedrock. Bedrock is the managed service that runs the models and coordinates the agents: a Supervisor Agent interprets each question and decides how to handle it, then delegates to the platform, a General Query Agent for direct answers and a Knowledge Base Agent for anything that needs looking up.

4,000

Queries per day in the sizing basis

200

Internal users in the sizing basis

12 hours

Time-to-live on uploaded files before automatic removal

By the numbers:

  • 4,000 - Queries per day in the sizing basis
  • 200 - Internal users in the sizing basis
  • 12 hours - Time-to-live on uploaded files before automatic removal
Changes

The engagement delivered a functional Microsoft Teams assistant integrated with Amazon Q Developer and Amazon Bedrock, configured with guardrails, private connectivity and an automated data-retention workflow, together with the technical and user documentation to run it.

  • Grounded answers — A Supervisor and the platform agent design on Bedrock routes each question to direct reasoning or knowledge-base retrieval, with contextual grounding checks to keep responses accurate.
  • Private by construction — AWS PrivateLink and VPC endpoints keep traffic to Bedrock off the public internet.
  • Guardrailed — PII detection and redaction, content filtering and prompt-injection protection apply to both inputs and outputs.
  • Scoped access — Mention-based activation, Azure AD group boundaries and blocking of external collaborators keep knowledge partitioned and internal.
  • Self-cleaning data — DynamoDB-tracked time-to-live and Lambda cleanup remove uploaded files automatically, minimising retained data.

The multi-agent framework is built to extend: new the platform can be added as needs evolve, letting the client broaden what the assistant can safely help with over time.

AWS Stack

Amazon Q Developer

For the Microsoft Teams connector and query handling.

Amazon OpenSearch Serverless

As the vector store backing the Knowledge Base.

AWS Lambda

For file ingestion and time-to-live-based data removal.

Amazon DynamoDB

For tracking uploaded files and their retention.

AWS PrivateLink

And Amazon VPC for private connectivity to Bedrock.

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.

We use cookies

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

See our Privacy Policy for details.