Sana Agents

What is Sana AI?

Sana AI is an enterprise workspace currently marketed on its main product pages as Sana Agents. It gives employees a conversational interface for working with company documents, emails, meeting notes, files, business applications and selected web sources. Rather than functioning only as a general-purpose chatbot, it is intended to work with an organization's existing knowledge and workflows.

A user can ask a question about internal information, investigate a topic across several sources, upload a file for analysis, create a report or presentation, summarize a meeting, or configure an agent to carry out a recurring task. The platform is aimed at organizations that need controlled access to business information and actions, not just individual users looking for a consumer chat assistant.

What Sana Agents does in practice

Enterprise search and research

Sana can search across connected documents, files, emails, meeting notes and business applications, subject to the access permissions configured for the workspace. This makes it useful for questions such as finding relevant company knowledge, preparing for a customer meeting, locating decisions from previous discussions or combining information from multiple internal sources.

It also supports broader research workflows using web search, uploaded material and selected company sources. The result may be a written answer, a structured brief, a table or another work product rather than only a short chat response. This puts Sana closer to an enterprise research and knowledge system than to a standalone web-search tool such as Perplexity.

Custom agents and recurring tasks

Teams can create agents for departments, domains or repeatable processes without writing code. An agent can be configured around particular knowledge sources, instructions and actions. Predefined and custom tasks can handle multi-step work, and eligible plans support scheduled workflows for recurring research, reporting or administrative activities.

The practical value depends on how well the organization has connected its systems and configured permissions. Sana is not simply an autonomous worker that can safely operate everywhere by default: actions depend on supported integrations, administrator settings, user authorization and the design of each task.

Documents, presentations and data analysis

Sana can generate reports, briefs, documents, presentations, tables and other structured outputs from selected sources. It can also analyze uploaded PDFs, spreadsheets, images and connected business data, producing summaries, calculations, charts or tables. These capabilities are useful when the source material already exists inside the company's systems and the goal is to turn it into a usable deliverable.

This overlaps with workspace products such as Notion AI, but Sana's positioning is broader: its central purpose is to connect information and actions across an enterprise rather than only assist inside one document workspace.

Meeting intelligence

Sana supports meeting recording or upload workflows, transcription, summaries, decisions and action items. Meeting integrations include Google Meet, Microsoft Teams and Zoom. The resulting notes can also become part of a wider knowledge and follow-up workflow, where users search past discussions or use extracted actions in connected systems.

Integrations and operating environment

Sana is primarily a web-based platform and also has an iOS app. Documented integrations include Gmail, Google Drive, Google Calendar, Microsoft Outlook, Microsoft Teams, Slack, Notion, Confluence, SharePoint, OneDrive, Dropbox, Box, Jira, Linear, Asana, Monday, HubSpot, Salesforce, BigQuery, Snowflake, Databricks and Workday. The exact availability of integrations varies by plan and workspace configuration.

Enterprise administrators can also connect custom integrations and remote MCP servers. MCP support can extend the tools available to agents, but it introduces another layer of configuration, access management and third-party security review. Sana also documents API access on eligible plans, although the research does not establish a single public API package or pricing structure.

Who Sana AI is for

Sana is most relevant to medium-sized and large organizations with substantial internal information and several business systems. Common users include sales, customer support, operations, finance, HR, legal, research and management teams. Typical use cases include:

• Searching company knowledge during onboarding or day-to-day work.
• Preparing sales and client-meeting briefs from internal records.
• Summarizing meetings and extracting decisions or follow-up actions.
• Creating recurring reports, research briefs and presentations.
• Analyzing spreadsheets, PDFs and data from connected systems.
• Updating CRM or project-management records through approved workflows.
• Building department-specific agents for repeatable research and administrative tasks.

It is less suitable for an individual who only wants an inexpensive general chatbot, or for a small team without enough connected data and repeatable processes to justify enterprise configuration.

Plans, access and pricing

Sana Agents has Free, Team, Civic and Enterprise tiers. The Free tier supports up to five workspace members and includes up to 10 meetings and 20 chat messages per user per month, along with unlimited agents and tasks and selected integrations. Civic users have higher documented limits, including up to 30 meetings and 200 chat messages per month.

Team supports up to 50 members per workspace, unlimited chat messages and recordings, and up to 10,000 documents per integration according to the supplied plan information. Enterprise limits, retention settings, integrations and administration are configurable. Sana does not publicly list a normal starting price for Team or Enterprise; paid subscriptions are based on workspace users and may be billed monthly or annually.

Advanced capabilities such as broader integrations, model selection, APIs, analytics, SSO, SCIM, MCP and enterprise administration are restricted by tier. This means the free version is useful for evaluating the workflow, but it should not be treated as a full representation of the enterprise offering.

Privacy, security and limitations

Sana states that customer data is encrypted in transit and at rest and is not used to train underlying language models. Its documentation describes ISO 27001 certification, SOC 2 compliance, GDPR compliance, role-based controls, SAML SSO, SCIM provisioning, multifactor authentication, penetration testing and enterprise retention controls.

Retention depends on product tier and configuration, and the exact default retention period was not verified. Access to connected information follows workspace permissions, but administrators still need to review how integrations, uploaded material, task actions and remote MCP servers are configured. Third-party tools can create additional security and reliability risks.

There are also practical limitations. Paid pricing is not transparent, free-tier limits are relatively small, and meaningful setup is required before the platform can provide useful organization-wide answers. Output quality depends on the connected sources, user permissions, selected model, task configuration and the accuracy of the underlying business data. Sana's model documentation mentions GPT-4.1 for Sana Search Assistant and selectable OpenAI and Claude models on some tiers, but routing can vary and should not be treated as a fixed model guarantee.

How Sana differs from adjacent AI tools

Unlike a general-purpose chatbot, Sana is designed around enterprise knowledge, permissions, integrations and actions. Unlike a meeting-notes product such as Fireflies.ai, meeting intelligence is only one part of a larger workspace that also covers search, research, content creation and workflow automation. It also differs from focused workplace assistants such as Slack AI because its intended scope extends across many applications and data sources.

The trade-off is complexity. Sana can be a useful central layer for organizations that want one interface over scattered business information, but its value depends on administration, source quality and integration work. For teams that need those controls and have the data infrastructure to support them, it is a credible enterprise AI agent platform. For casual users or organizations seeking simple chat, its setup and pricing model may be excessive.