Helpdesk Using Domino IQ RAG By Adam Foster and Doug Tait

Discover how AI can transform the way complex systems are supported and maintained. In this session, you’ll learn how to build an intelligent support bot using Domino IQ RAG that empowers both help desk operatives and end users to find accurate answers faster. We’ll explore real-world challenges in managing technical systems and show how AI can reduce downtime, improve resolution times, and enhance user confidence. Through practical examples, you’ll see how to design, train, and deploy an effective support assistant. Whether you’re in IT, operations, or digital transformation, this session will give you actionable insights you can apply immediately. Join us to see how AI can turn complexity into clarity and support into a strategic advantage. Takeaways 1 Domino IQ works nicely with RAG integration 2 Domino IQ is quite easy to integrate with existing databases 3 AI can be added to existing infrastructure without sacrificing security

Transcript

1. Help Desk AI using Domino IQ RAG Adam Foster & Doug Tait | Notes Advocates, Oval Business Solutions LET'SCONNECT 2026 · OSLO 20 OCT · STOCKHOLM 22 OCT
2. Adam Foster Notes Advocate, Oval Business Solutions • • • • • Worked at Oval for 16 years OpenNTF contributor board Keen cricketer and sea angler Session Speaker debut at Engage in April Oval are restarting UK HCL User group linkedin.com/in/adamfoster · adam.foster@oval.uk.com 2
3. Doug Tait Notes Advocate, Oval Business Solutions • • • • Worked at Oval for 17 years Developer and Consultant Guitarist and singer of songs Session Speaking Debut is Today! linkedin.com/in/dougtait · doug.tait@oval.uk.com 3
4. Why did Oval Business Solutions want to do this? Confirm set up of Domino IQ RAG is Sovereign Investigate Domino IQ RAG capabilities Improve support desk response to large customer 4 Free up our staff for other tasks Roll out to other systems and clients
5. What have we been working on since April? Exploring various options on the RAG database configurations on Domino IQ Dedicated RAG database consolidating Request and responses was created. A dedicated RAG database allows: - separate related documents to be consolidated into one text chunk - a subset of original database data to be used - consolidated multiple databases to be queried in one query (e.g. SupportDesk, FAQ and Best Practices in one query.) “Allow a LLM query to use more than one database” submitted to HCL ideas: DDXP-I-1449 5
6. How does AI support query work? User query submitted Answer and references are returned to the form Relevant References retrieved 6 Query is passed to the LLM LLM generates a response using all RAG data
7. 01 - Demo Noisy Coffee Grinder 7
8. 02 - Demo Empty Warning Light 8
9. 03 - Demo Identified User vs Product Manager User 9
10. Key Points and Takeaways Sovereignty, Security, References and Scoring Domino IQ RAG is Sovereign, reuses data and respects security The LLM request references are extracted from logs as notesIDs - can be used to remove unhelpful data (flag/readers) can be used to link to show details (graphics/videos) (UNIDS or custom reference ids may be more useful) Score shows confidence / probability Suggest HCL add notesID property to NotesLLMresponse “add NotesId property to NotesLLMResponse” submitted to HCL ideas: DDXP-I-1448 10
11. Where do we go from here? Next steps and improvement strategies to make use of AI: • Utilise Domino 14.5.1 with Statistics LLM and add new Query • Data preparation: combine databases, remove unhelpful data and improve references • Update RAG automatically from Lotusscript when new support completed • Investigate a larger, more advanced, or task-specific LLM • Investigate speed and results from very large databases 11
12. Thank you to our sponsors 12
13. Thank you! Questions? Adam Foster & Doug Tait linkedin.com/company/oval-business-solutions www.ovalbusinesssolutions.co.uk