Sana
Enterprise AI agent platform (part of Workday) that connects to company knowledge and builds custom agents for finance, HR and sales.
Delv Safety Grade: B
Score 72/100 · assessed 2026-04-19
Sana is now part of Workday, which gives it enterprise-grade backing and the legitimacy of a major HR/finance platform vendor. The maintainer score is excellent. However, this is a closed-source, enterprise-only platform with no public repository, minimal technical transparency, and contact-only pricing. The permissions footprint is broad: it reads from HR systems, finance databases, CRM tools, and document repositories, then surfaces that data through custom agents. This means identity:read, db:read across multiple sensitive domains, and likely network:outbound for API integrations. The lack of public documentation or open-source components makes independent verification impossible. Supply chain is proprietary SaaS, which is standard for enterprise tools but offers no package-level scrutiny. No known incidents, but opacity is the trade-off for enterprise support.
Green flags
- Backed by Workday, a major enterprise platform with strong compliance posture
- Purpose-built for regulated enterprise use cases (HR, finance)
- No known security incidents or breaches
- Scoped to internal company knowledge, not external web scraping
Red flags
- Closed source with no public repo or technical documentation
- Broad access to HR, finance, CRM and document systems across enterprise
- No transparency into data handling, model usage or security controls
- Enterprise-only with opaque pricing and no public review
Permissions requested
Pricing
Platforms
Review
If you are a Workday customer with clean data and repetitive knowledge work across finance, HR or sales, Sana will save time. If you are not already in the Workday ecosystem, the integration cost outweighs the benefit. Look at Glean or a general LLM tool instead.
Good at
- Deep integration with Workday HCM, Financials and other enterprise systems
- Handles multi-step retrieval tasks across structured data sources without manual exports
- Agents can be scoped to specific departments or workflows, reducing noise
- Does not hallucinate facts when working with clean, structured data
- Useful for repetitive queries that currently require manual data joins
Watch out
- Requires Workday ecosystem to justify the cost and integration effort
- Customisation needs professional services or internal engineering, not truly no-code
- Fails badly if underlying data is messy, mislabelled or stale
- Opaque enterprise pricing with long sales cycles
- Less flexible than general-purpose knowledge agents like Glean for non-Workday tools
Use cases
- knowledge work
- deal updates
- data analysis