AI providers

Choose which AI models power each agent and task.

The Settings → AI Providers page controls which models run Motif's agents. Configuration is per task and per agent. Pick a stronger model where quality matters most and a faster, cheaper one for routine work.

---

What you configure

Model selection works at three levels, from broad to specific:

• Processing Scopes set the default model for a class of work: Chat & Q&A, Document Digest, Enrichment & Classification, Deep Analysis, and Simulation.

• Per-operation overrides let you switch an individual operation off its scope default. This includes a Clarity/KG Query group for retrieval and answer generation.

• Per-agent selection applies to each agent in the table: Profiling, Investment Strategist, Market Commentary, Reporting, Market Research, and QA.

Operations inherit from their Processing Scope until you override a row, so you set sensible defaults once and refine only where it matters.

Clarity is tuned through the Chat & Q&A scope and the Clarity/KG Query operations rather than as a standalone agent.

For each level, choose a stronger model where quality matters most and a faster, cheaper one for routine work. For example, run a strong model for Market Research and Market Commentary and a lighter one for QA.

---

Cost visibility

Model choice drives spend. The Observation deck tracks token usage and cost per agent and per run, so you can see the effect of a model change and adjust. Start with Motif's defaults, watch the AI-cost breakdown, and move models up or down per agent from there.