AI engineering project

Askly: RAG Customer Support Project

Askly is a customer support bot built around retrieval-augmented generation (RAG). It retrieves information from an organization's knowledge base to help resolve queries and assist support agents.

By Ansh Sharma · AI Engineer & Full-Stack Developer

  • Retrieval-augmented generation
  • Knowledge retrieval
  • Customer support

What problem does Askly address?

Support teams need answers grounded in their organization's own documentation. Askly connects those knowledge resources with incoming queries so that customer support can use information specific to the product or organization.

How does the RAG workflow work?

The core workflow connects a support question to relevant knowledge-base material, then uses that retrieved context to support an answer. Retrieval is the architectural link between organizational information and the support conversation.

Where do human agents fit?

Askly's scope includes assisting support agents as well as answering queries. The public product site describes document ingestion, answer citations, and handoff to agents, connecting knowledge retrieval with the wider support workflow.

How would I evaluate a support RAG system?

A useful evaluation would include questions with documented answers, questions missing from the knowledge base, and ambiguous requests. I would check whether retrieval finds the relevant material, whether answers are supported by that material, and when a question should be handed to a person. These are evaluation criteria, not reported benchmark results.

Explore the project

Explore Askly ↗
Ansh Sharma monogram

Let's connect

Let's build something.

Open to AI engineering and full-stack roles and collaborations. See my work on GitHub and let's talk on LinkedIn.

or email me ↗