Inspiration
Consumer complaints can be surprisingly difficult to resolve. A missing delivery or incorrect charge might seem straightforward, but customers often struggle to gather evidence, explain the problem clearly, contact the right business and keep track of what happens next.
I wanted to explore how a conversational AI assistant could make that process simpler, more accessible and less intimidating.
SIR — Smart Issue Resolver is a prototype AI consumer advocate designed around a straightforward principle: the assistant does the organisational work, while the customer remains in control.
What it does
SIR guides a customer through the complaint process using a structured, conversational experience designed to demonstrate an Alexa+ use case.
The prototype allows users to:
- Describe a consumer problem through guided interaction.
- Upload supporting evidence, including PDF order confirmations.
- Extract and review relevant details from that evidence.
- Prepare a clear, evidence-backed complaint.
- Review and explicitly approve an email before exporting it.
- Record correspondence, outcomes and case history.
The demonstration uses a completely fictional case involving an £89 camping stove that was ordered but never delivered.
SIR does not automatically send emails or contact businesses. The customer retains responsibility for reviewing and sending their correspondence.
How I built it
SIR was developed as a local, Windows-based demonstration prototype, combining a guided complaint workflow, document evidence extraction, structured case information and an email-export process.
The experience includes a scripted simulation of Alexa+ interaction to demonstrate how a conversational assistant could support consumers through a multi-stage task.
The project is packaged with a Windows judge distribution, fictional sample evidence, instructions and a public source-code repository so the workflow can be reviewed and reproduced.
Challenges I faced
One of the biggest challenges was making the entire experience coherent rather than simply demonstrating isolated AI features. Evidence extraction, complaint drafting, email approval and case tracking needed to feel like parts of one continuous journey.
Another challenge was ensuring that the demonstration was safe and transparent. The prototype distinguishes simulated Alexa+ behaviour from live integration, uses fictional evidence, and requires explicit customer approval before email export.
Preparing a reliable judge distribution also required careful testing, documentation and a final security review of the packaged demonstration files.
Accomplishments I'm proud of
I'm particularly pleased that SIR demonstrates a complete consumer complaint journey, from the initial problem through to a prepared email and tracked outcome.
Rather than stopping at generating a piece of text, the prototype shows how a conversational assistant could help people organise evidence, make informed decisions and maintain control over a real-world administrative process.
What I learned
Building SIR reinforced the importance of designing AI-assisted workflows around user trust, transparency and practical outcomes.
It also highlighted how valuable conversational interfaces could become when combined with structured evidence and clearly defined approval steps.
What's next for SIR
Future development could include genuine Alexa+ integration, more document formats, additional consumer complaint scenarios and richer case-management capabilities.
The longer-term vision is a consumer assistant that makes resolving everyday problems easier without taking control away from the person it serves.
Log in or sign up for Devpost to join the conversation.