Sana: learning, knowledge, and custom agents

Sana is what happens when a learning platform and an AI assistant meet. It's more than corporate LMS + AI — it's a coherent rethink of how teams learn and use institutional knowledge.

The three surfaces

• Learn. Course creation, delivery, pathways. Like a modern LMS but built around how people actually take in information.
• Assistant. Chat with your org's content, grounded in Sana's knowledge base + connected sources.
• Agents / custom AI. Purpose-built AIs trained on specific knowledge for specific workflows.

What Sana is good at

• Creating learning content quickly. AI helps write courses; humans curate.
• Mixed formats. Video, text, quizzes, live sessions, all in one platform.
• Learning as a two-way street. Not just consumption; the assistant helps people apply what they've learned.
• Analytics. Who's learning what, where gaps are.

It blurs the line between "learning platform" and "knowledge platform." That's a feature, not a confusion.

Where Sana shines vs. alternatives

vs. generic LMS (Docebo, Cornerstone): Sana's AI integration is years ahead. vs. Glean: Glean is a better general search tool; Sana is better for structured learning. vs. Notion + Notion AI: Notion is a better document tool; Sana is better for delivering training at scale.

Many orgs run Glean + Sana — Glean for "where is this?" and Sana for "teach me this."

Typical deployment

• HR/L&D team owns the platform.
• Initial content: existing compliance training, onboarding, product training migrated in.
• AI assistant enabled with access to the content library + select external sources (company docs, sales collateral).
• Rollout expands with custom learning paths for specific roles or skills.

Custom AI as a differentiator

Sana's custom AI feature lets you build specialized assistants:

• Onboarding AI that walks new hires through their first 30 days.
• Sales enablement AI grounded in product knowledge, compete battle cards, case studies.
• Compliance AI that answers policy questions with citations.

These aren't chat bots — they're curated experiences with guardrails and tracking.

Content creation with AI

The content creation workflow is one of the better AI implementations in L&D:

1. Upload source materials or describe the topic.
2. Sana generates a course outline.
3. You edit, reorder, expand.
4. Sana generates slides, quizzes, summaries.
5. You review and publish.

Cuts course creation time from weeks to days. The humans still own pedagogy; AI removes drudgery.

What to watch

• Adoption alongside existing LMS. Migration from a legacy system takes effort. Plan for a dual-run period.
• Permission model. Sana's is solid but differs from your existing tools'. Get it reviewed.
• Content quality bar. AI-generated course drafts are a starting point, not a shipping bar. Editorial discipline matters.

Data handling

Sana is enterprise-grade with appropriate certifications. Data stays in your tenant; training on your content is off by default.

For regulated industries: verify per-customer specifics with Sana's security team. Standard disclaimers about "verify current options" apply.

Integration with other tools

Sana connects to:

• Identity providers (Okta, Azure AD) for SSO.
• HRIS (Workday, BambooHR) for role-based content.
• Communication tools (Slack, Teams) for in-flow learning reminders.

Integrations are solid; few rough edges.

The honest assessment

Sana is valuable for orgs that:

• Treat learning as strategic, not compliance-checkbox.
• Want AI-assisted content creation at scale.
• Can dedicate team ownership (not "HR admin handles it in spare time").

It's expensive relative to basic LMS alternatives. The value is real for the above profile. Below that bar, cheaper tools suffice.