Get a Free Quote
Article

HCL Domino IQ: Secure, Private AI for Domino Applications

HCL Domino IQ is the AI engine built into HCL Domino 14.5 and later. It runs a large language model next to your Domino server, or connects to an AI endpoint you choose, so Domino applications can summarize, translate, classify and draft content while your data stays inside infrastructure you control. Developers call it from LotusScript, and administrators configure it centrally in the Domino IQ database.

RockTeam helps enterprises plan, deploy and govern Domino IQ as part of our HCL Domino consulting and support services. This guide explains how Domino IQ works, what it takes to run it and how to move from pilot to production.

How Domino IQ Works

ElementWhat it does
Domino IQ serverA Domino 14.5 or later server on Windows or Linux configured to run Domino IQ
Local modeRuns an AI inference engine on the Domino server itself, using a single LLM in GGUF format and NVIDIA GPU hardware
Remote modeSends requests to an OpenAI V1 compatible endpoint, such as a private AI server you host
Domino IQ database (dominoiq.nsf)Central administration of configurations, models and system prompts
System promptsAdministrator defined instructions for each use, such as summarizing or classifying
LotusScript classesNotesLLMRequest and NotesLLMResponse let applications send AI requests and use the answers
Guard models (14.5.1)Screen requests and responses, and block content flagged as unsafe

Domino IQ Use Cases in Domino Applications

  • Built in mail assistance. Once enabled, Domino IQ powers summarize and reply drafting in the Notes client and HCL Verse.
  • Summarize long mail threads, case histories and documents.
  • Translate content for global teams without sending it to a public service.
  • Classify and route incoming requests, incidents or approvals to the right queue.
  • Extract key fields from unstructured text into Domino documents.
  • Draft replies, descriptions and reports for users to review.

Existing applications can usually add these capabilities without being rebuilt; see Domino application development and modernization.

Local or Remote Mode?

FactorLocal modeRemote mode
Where the model runsOn the Domino IQ serverOn a separate AI server or service you choose
HardwareNVIDIA GPU; HCL recommends at least 8 GB of GPU memory, with higher capability GPUs for productionNo GPU on the Domino server
Data pathStays on the Domino serverTravels to the endpoint, so choose a private or trusted host
Best forAir gapped, regulated and sovereignty sensitive environmentsShared AI infrastructure or proofs of concept without new GPU servers

Requirements and Planning

  • Domino version. Domino IQ needs Domino 14.5 or later; guard models need 14.5.1. See our HCL Domino upgrade guide.
  • Infrastructure. For local mode, plan GPU servers and drivers that match HCL’s requirements; for remote mode, plan the endpoint, TLS and network path.
  • Model choice. Each local Domino IQ server runs one model, so size the model to the use cases and hardware.
  • Licensing. Confirm your Domino entitlements cover the servers and users involved; see Domino licensing and renewals.

Governance, Security and AI Strategy

Domino IQ fits best when it is part of an enterprise AI strategy rather than an isolated experiment:

  • Decide which data AI can use, respecting Domino access control lists and document level security.
  • Standardize system prompts so behavior is consistent and reviewable.
  • Enable guard models and logging so requests and responses can be audited.
  • Involve security leadership, such as a vCISO, in approving use cases.
  • Align with other platforms. Where Microsoft 365 or other AI tools are in use, define which system is the source of truth for each workflow.

From Pilot to Production

  1. Select one or two use cases with clear value and willing business owners.
  2. Prepare the environment: upgrade to a current release and configure local or remote mode.
  3. Build and test the LotusScript integration and system prompts with real data.
  4. Measure results such as time saved, accuracy and user adoption.
  5. Scale to more applications with standard prompts, guard models and monitoring.

Ready to explore Domino IQ?

RockTeam’s Domino consultants and developers design Domino IQ pilots, prepare servers and build AI into your Domino applications. RockTeam is an official HCL Business Partner. Start with a no cost assessment.

Talk to a Domino Consultant

Related Articles

Domino IQ FAQs

Domino IQ is the AI capability in HCL Domino 14.5 and later. It runs an AI inference engine with a large language model on the Domino server, or connects to an AI endpoint you choose, and lets Domino applications use AI through LotusScript.

Not unless you configure it to. In local mode the model runs on your Domino server; in remote mode requests go only to the endpoint you specify, which can be a private AI server.

Local mode needs NVIDIA GPU hardware on a Windows or Linux Domino server; HCL recommends at least 8 GB of GPU memory. Remote mode needs no GPU on the Domino server.

Domino IQ was introduced in Domino 14.5. Domino 2026 (14.5.1) adds guard models that screen AI requests and responses.

Through the NotesLLMRequest and NotesLLMResponse LotusScript classes, combined with system prompts that administrators define in the Domino IQ database.