Dust vs Notable

Dust and Notable do not overlap in capability. Dust is a horizontal agent builder: you define an agent with instructions, attach knowledge via connectors (Slack, Notion, Google Drive, GitHub, Salesforce, Zendesk and 70+ more), pick a model per agent from 20+ options, and run it on a schedule, an event trigger or in chat. It adds Frames for dashboards, Pods for shared workspaces, Spaces for per-team data segmentation, a dual-layer permission model, SSO/SCIM and audit logs. Its July 2026 release of self-improving skills — reusable instruction sets that adapt from usage — targets the ops builder who wants agents to compound expertise over time.

Notable is vertical and prescriptive. Its agents execute named healthcare workflows: real-time eligibility checks, copay estimation and collection, prior authorization, denial prevention and appeal-letter generation, chart review for risk adjustment, referral management (14 to 3 days turnaround), care-gap outreach, and a Voice AI Agent for patient calls and pre-procedure instructions. Flow Builder is a low-code surface for these automations, Flow AI generates net-new ones, and Sidekick assists staff in natural language. It plugs into Epic, Cerner, athenahealth and Twilio.

The honest summary: Dust gives builders a general toolkit and asks them to assemble the workflow; Notable ships pre-built healthcare workflows and asks you to configure them. Neither substitutes for the other.

Dust is freemium with credit-based usage. Credits vary by model and tool use and do not roll over, which Dust itself flags as a poor fit for buyers wanting predictable flat-rate pricing. That is the real cost consideration: an agent mixing frontier models with heavy connector use will drift month to month, and you should budget on consumption, not seats. Enterprise controls (SSO, SCIM, audit logs, Spaces) imply a paid tier, though the published facts here do not give a number.

Notable does not publish a price at all — the listing type is contact-sales, and its own not-for list explicitly excludes procurement teams that need a price list before scheduling a demo. Expect a sales-led evaluation and a contracted deployment rather than a self-serve signup, with cost tied to the scope of workflows and sites of care rather than a per-seat rate you can look up.

For a buyer, the pricing models reinforce the comparability verdict: one is a bottom-up, metered SaaS you can start free; the other is a top-down enterprise contract sold into hospitals. There is no scenario where price alone decides between them, because the buying motions do not resemble each other.