How Sana Labs and Workday AI Automate Enterprise Learning and Knowledge

In the rapidly evolving landscape of enterprise technology, the friction between finding information and applying it has long been a bottleneck for productivity. Organizations often struggle with "tool sprawl," where critical knowledge is scattered across Slack, Google Drive, GitHub, and legacy Learning Management Systems (LMS). Sana Labs emerged as a notable solution to this problem by treating all company data as a single, searchable, and actionable intelligence pool.

As of August 2026, Sana has transitioned from a high-growth startup to a core component of the Workday ecosystem. This integration allows users to not only search for information but to execute tasks—like onboarding a new hire or processing a financial approval—directly through a conversational AI interface. By merging the functions of an AI agent with the structured environment of a learning platform, Sana stands out as a top choice for enterprises aiming to achieve "AI fluency."

What Is the Difference Between Sana Learn and Sana Agents?

To understand the platform's value, one must distinguish between its two primary pillars: Sana Learn and Sana Agents. While they share the same underlying AI infrastructure, they serve different operational needs within the enterprise.

Sana Learn: The Evolution of the LMS

Traditional LMS platforms are often viewed as static repositories for compliance videos. Sana Learn reimagines this by using generative AI to assist in every stage of the learning lifecycle. Administrators can upload a single PDF or record a meeting, and the AI will instantly generate a structured course, complete with quizzes and interactive elements. For the learner, it provides a personalized experience where an AI tutor can answer specific questions about the course material in real-time, adapting the difficulty based on the user's progress.

Sana Agents: Conversational Enterprise Search

Sana Agents serve as the conversational interface for the entire company. Unlike a basic chatbot, these agents are connected to over 100 enterprise applications. A user can ask, "What was the final decision on the Project Phoenix budget?" and the agent will scan Slack threads, email chains, and meeting transcripts to provide a cited answer. According to Sana's official documentation, this capability allows employees to find answers 95% faster than manual searching.

The true power lies in the interaction between the two. An employee might use a Sana Agent to find a technical document on GitHub. If they realize the team needs more formal training on that topic, they can instantly convert that document into a Sana Learn module with a single command. This creates a seamless loop between knowledge discovery and skill development.

How the Workday Acquisition Changes Everything for Users

The $1.1 billion acquisition of Sana by Workday in November 2025 was a pivotal moment for the enterprise AI sector. For existing Workday customers, this isn't just another integration; it is a fundamental shift in how they interact with their HR and Finance data. The platform is now frequently referred to as "Sana from Workday," signaling its role as the intelligence layer sitting atop the Workday core.

One of the most significant benefits of this acquisition is the concept of Agentic HR. In the past, an HR manager might need to navigate multiple menus in Workday to initiate a performance review or approve a leave request. With Sana, these tasks are handled via natural language. The AI agent understands the context of the request, checks the relevant company policies stored in the knowledge base, and executes the action within Workday's secure environment.

Furthermore, the acquisition has addressed the "data silo" problem. Because Sana is now a part of Workday, it has deeper, native access to employee records, organizational charts, and financial workflows. This allows for highly specific queries, such as "Identify all employees in the engineering department who haven't completed their security certification and send them a reminder." This level of automation was previously difficult to achieve without custom coding or manual intervention.

Why Sana Uses Knowledge Graphs Instead of Basic Vector Search

Most modern AI tools rely on a technique called Vector Search (or RAG - Retrieval-Augmented Generation). While effective for finding "similar" pieces of text, vector search often fails when faced with complex, structured queries. Sana Labs has taken a different approach by building a Knowledge Graph.

As detailed in the Bitter Lessons philosophy shared by the company, Sana believes that leveraging massive computation is more effective than trying to hard-code human knowledge. Their Knowledge Graph maps the relationships between people, projects, documents, and concepts. For example:

• Vector Search: Might find documents containing the words "European" and "Client."
• Knowledge Graph: Understands that "Client A" is located in "Paris," which is in "Europe," and that "Client A" has a "Contract Value" of "$6M."

This distinction allows Sana to answer highly specific business questions like, "List all European clients with more than $5M in annual recurring revenue that we met with last month." Standard AI wrappers typically cannot handle this level of relational reasoning.

How the R-4 Agent Solves the Hallucination Problem

A common critique of enterprise AI is the tendency to "hallucinate" or provide incorrect information. Sana addresses this through its R-4 Agent architecture. This system employs a multi-step planning process and a "self-reflection" loop. When a query is received, the R-4 agent doesn't just generate a response; it creates a plan to find the data, executes the search, checks the results against the original source for accuracy, and then summarizes the findings. If the agent detects a contradiction in its own reasoning, it restarts the process before the user ever sees the output.

Comparing Sana Labs to Glean and Microsoft Copilot

Choosing the right AI platform depends heavily on an organization's existing tech stack and specific goals. While Microsoft Copilot is ubiquitous, it may not offer the same depth of learning integration as Sana. Similarly, Glean is a top-rated choice for pure enterprise search but lacks the LMS capabilities that define Sana Learn.

Feature	Sana Labs	Glean	Microsoft 365 Copilot
Primary Focus	AI-Native Learning & Search	Enterprise Search & Discovery	Productivity & Office Automation
LMS Integration	Native (Sana Learn)	Third-party connectors only	Limited to Viva Learning
Data Architecture	Knowledge Graph + R-4 Agents	Vector Search + Knowledge Graph	Large Language Model (GPT-4)
Workday Synergy	Native (Workday Company)	Standard API Integration	Standard API Integration
Best For	Workday users & L&D teams	Pure search across many apps	General office productivity

For organizations that are already heavily invested in the Workday ecosystem, Sana is a highly recommended option. However, if a company's primary pain point is simply finding files across a massive array of 500+ different applications without a need for integrated training, Glean remains a strong contender.

Transforming Meetings into Searchable Company Knowledge

One of the most underutilized sources of knowledge in any company is the verbal exchange that happens during meetings. Sana treats Zoom, Microsoft Teams, and Google Meet as "net new" knowledge sources. Instead of just providing a transcript, Sana indexes the content of these meetings so they appear in global search results alongside PDFs and Slack messages.

Consider a scenario where a Product Manager misses a critical technical sync. Instead of watching a 60-minute recording, they can ask the Sana Agent, "What was the specific decision made regarding the API rate limits?" The agent will not only provide the answer but also link to the exact timestamp in the video where the decision was made. Furthermore, Sana can automatically generate a Jira ticket or a summary document based on the meeting's action items, significantly reducing administrative overhead.

"Sana is not an LMS with Gen AI added. It’s AI at its core, designed to change how knowledge flows through an organization." — Josh Bersin, Global Industry Analyst

Using Scenario Cards for AI-Driven Role Play

Beyond simple content generation, Sana Labs offers advanced tools for skill development, most notably Scenario Cards. These allow employees to practice high-stakes interactions in a safe, AI-driven environment. For example, a sales representative can practice a discovery call with an AI prospect that has been programmed with specific objections and personality traits.

The Studio Cards feature takes this a step further by allowing administrators to turn a single prompt into a "choose-your-own-adventure" style training module. This is particularly effective for compliance or safety training, where learners must make decisions in real-time and see the consequences of their actions. According to user reports, this approach has reduced course production time from three weeks to just three hours for many teams.

What Does Sana Labs Actually Cost in 2026?

In recent years, Sana has shifted its pricing strategy to align with its status as an enterprise-grade platform. While earlier iterations offered mid-market tiers starting around €7,200 per year, the 2026 pricing model is focused on large-scale deployments. Current estimates suggest that enterprise contracts typically range from $50,000 to over $100,000 per year, depending on the number of seats and the depth of integrations required.

There is currently no transparent "self-serve" tier for the full suite, as the implementation process usually involves a dedicated success manager to ensure the Knowledge Graph is correctly mapped to the company's data. However, the ROI data remains a strong selling point for the platform. For instance, the Polygon Group reported saving 28 working days per year through Sana’s automations, while other companies like Berner reached 90% adoption within just 40 days of launch.

Common Implementation Pitfalls and How to Avoid Them

Despite its sophisticated capabilities, implementing Sana is not without challenges. Organizations should be aware of several common friction points to ensure a smooth rollout:

The "Opinionated UX" Challenge

Sana features a very specific, minimalist design language. While many users find it beautiful, teams accustomed to traditional folder-based LMS structures may initially find it restrictive. It is important to provide "change management" training to help users adapt to the search-first workflow.

Integration Friction with Legacy Tools

While Sana offers over 100 connectors, connecting to highly customized legacy systems (like on-premise SharePoint versions) can require significant technical oversight. Ensure your IT team is involved early in the process to handle API permissions and data mapping.

Data Privacy and Silos

A common concern is whether the AI will expose sensitive information (like executive salaries) to the general staff. Sana uses strict permission-based indexing, meaning an employee will only see search results for documents they already have permission to view in the source application (e.g., Google Drive or Slack).

Frequently Asked Questions

Is Sana Labs a good company to work with?

Sana Labs is widely regarded as a leader in the AI-native learning space. With the backing of Workday and significant funding from firms like NEA, the company has the stability of a major enterprise player while maintaining the innovation speed of a startup. Their high adoption rates (often exceeding 90% in the first month) suggest a strong product-market fit.

Who is the CEO of Sana Labs?

The company was founded and is led by Joel Hellermark. He has been a vocal advocate for the "Bitter Lessons" philosophy in AI, which emphasizes the power of general-purpose computation over human-designed heuristics. Under his leadership, Sana has grown from a Swedish startup to a global enterprise AI powerhouse.

Does Sana Labs offer a free version?

As of 2026, there is a limited "Free" tier for Sana Agents that allows individuals or small teams to test the conversational search capabilities. However, the full enterprise suite, including Sana Learn and the deep Workday integrations, requires a paid contract and a formal implementation process.

How does Sana integrate with Slack?

Sana integrates with Slack by indexing public channels and authorized private conversations. Users can interact with Sana directly within Slack by tagging the Sana bot. The AI can summarize long threads, answer questions based on past discussions, and even suggest relevant training modules based on the topics being discussed in a channel.

Is my company data used to train public AI models?

No. Sana employs a strict enterprise data silo model. Your company's data is used to populate your specific Knowledge Graph and fine-tune your internal agents, but it is never fed back into public models like ChatGPT or used to train the AI for other customers. The platform is SOC2 compliant and adheres to global data privacy standards.