ServiceNow just announced Otto at its Knowledge 2026 conference in Las Vegas. This announcement deserves a careful read from anyone evaluating enterprise AI right now, not because every claim will survive scrutiny, but because the underlying problem it is trying to solve is real, and the way it is positioned will shape buying decisions across the industry for the next twelve to eighteen months.
ServiceNow’s platform processes more than 100 billion workflows per year. The Moveworks acquisition was strategically coherent. And the core argument ServiceNow is making – that enterprise AI has a "completion problem" where tools can find information but cannot actually finish work – is something every CIO we have spoken with in the past six years since founding Atolio has experienced firsthand.
But the distance between what Otto promises and what it currently delivers is worth understanding clearly before making decisions or investments.
What is ServiceNow Otto?
Otto is ServiceNow's new unified AI experience. It brings together three angles the company has been building or acquiring: Now Assist (their existing generative AI layer), Moveworks (the conversational AI and enterprise search company acquired in late 2025 for $2.85 billion), and AI Experience (a newer agentic layer introduced last year).
The pitch is direct: employees ask Otto something in natural language, and Otto handles the rest. It routes the request to the right system, crosses departmental boundaries, and completes work end-to-end without the employee needing to know which system is involved. Beneath the experience sits the AI Control Tower – ServiceNow's governance layer – which logs every interaction, enforces policies, and provides explainability for every decision Otto takes.
What is ServiceNow Otto?
Otto is ServiceNow's new unified AI experience. It brings together three angles the company has been building or acquiring: Now Assist (their existing generative AI layer), Moveworks (the conversational AI and enterprise search company acquired in late 2025 for $2.85 billion), and AI Experience (a newer agentic layer introduced last year).
The pitch is direct: employees ask Otto something in natural language, and Otto handles the rest. It routes the request to the right system, crosses departmental boundaries, and completes work end-to-end without the employee needing to know which system is involved. Beneath the experience sits the AI Control Tower – ServiceNow's governance layer – which logs every interaction, enforces policies, and provides explainability for every decision Otto takes.
That architecture is ambitious. And for the narrow set of use cases where it works now, early results are promising. ServiceNow reports that EmployeeWorks, the first product to surface Otto, closed six deals exceeding one million dollars each in its first month.
The first live AI specialist is an IT Level 1 Service Desk agent. That tells you what is currently working and where the deepest investment has been made.
The completion problem is real, and worth taking seriously
Before getting into the harder questions, it is worth steelmanning ServiceNow's argument, because they are right about the problem even if the solution is more limited than presented.
Most currently-available enterprise AI tools are what ServiceNow calls "sidecar AI": copilots bolted onto individual applications, each with their own knowledge, their own interface, and no ability to cross systems. An employee who needs to request software access, find the policy document that governs it, check whether a similar request was approved last quarter, and submit a ticket has to navigate four different tools to do it.
ServiceNow's argument is that the completion layer, the layer that takes intent and turns it into finished work across approval chains, permissions, and cross-system workflows, is what has been missing. That argument is correct. Whether Otto is the right implementation of that layer for your organization is a different question.
Five questions every CIO should ask before adopting Otto
1. Where does your data actually go?
Otto is a SaaS platform. When employees ask questions and Otto retrieves answers, that retrieval happens in ServiceNow's cloud infrastructure. For many organizations this is acceptable. For organizations in financial services, healthcare, defense, or any industry with data sovereignty requirements or IP considerations, this is often a structural barrier.
The question is not whether ServiceNow has strong security controls – they do. The question is whether your architecture requires that enterprise knowledge, board communications, client data, competitive strategy, and employee records never leave your own environment. If the answer is yes, a SaaS deployment model does not satisfy that requirement regardless of the contractual language around it.
This is not a hypothetical concern. The lead investor in Atolio's Series A, Translink Capital, specifically sought out Atolio because a major financial institution asked them to find an enterprise search platform that could deploy entirely within the client's own private cloud. It was the only requirement on the table. Most platforms could not meet it.
For a deeper look at what sovereign AI deployment actually requires in regulated environments, including federal agencies, see: Sovereign AI Enterprise Search for Federal Agencies.
2. How deep is the permission-awareness across non-ServiceNow systems?
ServiceNow's own platform has strong role-based access controls and has been building permission infrastructure for years. That is a material strength.
But Otto's enterprise search capability, powered by Moveworks, needs to honor permissions across external systems too: SharePoint document-level permissions, Salesforce record-level visibility rules, Jira project permissions, and the access control lists of every system an employee's query might touch. The depth of that permission federation across third-party sources is not detailed in the Knowledge 2026 announcements.
This matters because permission-aware retrieval is not a UI-level problem. It is an index-level problem. An AI assistant that retrieves a document a user should not have access to, even if that document never renders in the final answer, has already created compliance risk. The architecture that prevents this has to be built at the layer where content is indexed, not the layer where it is displayed.
Today or in due time, ServiceNow offers 30-plus enterprise integrations. What remains unclear is whether those integrations enforce permissions at the index layer or the display layer for each connected system.
3. Which functions in your organization actually need this?
ServiceNow and Moveworks both have deep roots in IT service management. That is where the tooling is most mature, the AI specialists are most proven, and the integrations are deepest. The Level 1 IT Service Desk agent resolving cases 99% faster than human agents is the showcase statistic for a reason.
The "any function" framing in the announcement includes CRM, legal, finance, HR, workplace services, procurement, and more. These are real roadmap intentions. But the specialists covering those functions are newly announced, not widely deployed. A company evaluating Otto for its legal team, R&D organization, or marketing function should weigh current capability against roadmap intent.
The Deloitte 2026 State of AI in the Enterprise report found that only 25% of organizations have moved 40% or more of their AI pilots into production. The gap between piloting and deploying at scale is almost always an integration problem, not a model problem. Buying into a platform's future function coverage before integrations are proven and the product is battle-tested only further jeopardizes the success of pilots and deployments.
The question for any executive is: which functions in my organization have the problem Otto solves now, versus the problem it intends to solve in twelve months? Buying a workflow execution platform based on its future function coverage is a meaningful risk, especially when the implementation costs are high and the switching costs are higher.
4. What is your LLM strategy, and does Otto respect it?
ServiceNow has announced partnerships with Anthropic, OpenAI, Google, and Microsoft Azure. That is a meaningful set of relationships. But in a SaaS architecture, the customer does not control which model is used, when it changes, or what the terms of that model's data handling look like. ServiceNow makes those decisions.
For organizations with model governance requirements, those already committed to a specific LLM provider for compliance reasons, or those operating in jurisdictions with regulations around AI model usage, this is a real constraint. The ability to choose your model, run it locally, and ensure no inference happens outside your infrastructure is a different capability than having vendor partnerships with multiple LLM providers.
5. Are you solving an enterprise-wide knowledge discovery problem or a ServiceNow workflow execution problem?
This is the most important question, and the one most buyers skip.
Otto is built to complete work. Its value proposition requires that the work to be completed lives on ServiceNow's platform or flows through systems ServiceNow deeply integrates with. If an employee asks Otto to provision software access, that works because ServiceNow owns the ITSM workflow underneath. If an employee asks Otto to find the most recent competitive analysis, the customer contract from 18 months ago, and the engineer who built the integration that is now breaking, the answer to that question is not an IT workflow. It is a knowledge discovery problem that benefits from spanning more than just one platform.
Most organizations have both problems. But they have the knowledge discovery problem first. According to McKinsey, employees spend 1.8 hours every day searching and gathering information. The MIT 2025 study of enterprise AI pilots found that 95% of failures traced back to data quality and integration problems, not the AI models themselves. The foundation for effective agentic AI is not a better action layer. It is a better knowledge layer.
Otto does not solve the knowledge discovery problem for functions that do not run on ServiceNow. It solves it for the workflows ServiceNow already owns. For a fuller treatment of why the sequence matters, see The Enterprise AI Maturity Hierarchy.
What Otto does not change about the enterprise search problem
The enterprise knowledge problem that Atolio was built to solve is not primarily a workflow problem. It is a relevance and access problem. It’s the sales rep who cannot find the customer contract. The engineer who cannot find who built the system they are now responsible for. The legal team that cannot surface the precedent that exists inside their own organization.
These problems exist across every function, every tool, and every organizational level. They are not solved by a better IT ticketing system with a conversational interface. They are solved by a platform that indexes the full knowledge surface of an organization, enforces permissions at the index level, and returns answers that are hyper relevant to the specific person asking.
Otto and Atolio are not the same kind of tool. Otto is an action completion platform that includes enterprise search as a component, and whose real value proposition requires deep integration with ServiceNow's workflow layer. Atolio is a knowledge discovery platform, deployed entirely within the customer's own infrastructure, that serves every function in the organization equally and does not require any workflow migration or platform consolidation to work.
Moveworks handles your IT helpdesk queue. Atolio handles your entire company’s institutional knowledge. If you’re evaluating Moveworks or Otto as your enterprise search solution, you’re solving the wrong problem.
The two products can coexist. An organization running ServiceNow for ITSM can run Atolio for cross-functional knowledge discovery. Otto’s search is strongest within the ServiceNow ecosystem. Atolio’s search is designed for everything else, which is most of what the organization knows. Here is how to quantify what that fragmented search is costing you.
A decision framework for enterprise buyers
Given what was just announced, here’s how executives evaluating enterprise AI should frame their decision:
- If your primary problem is IT helpdesk volume and you are already a ServiceNow shop: Otto is a logical next step. The tooling is proven, the integration is native, and the ROI case for IT ticket deflection is well established.
- If you need cross-functional knowledge discovery across sales, legal, R&D, finance, and marketing: Evaluate purpose-built search platforms. The function coverage in Otto does not extend to these teams in a meaningfully deployed way.
- If you are in a regulated industry with data sovereignty requirements: Otto’s SaaS model is a structural barrier regardless of its capabilities. Any solution that requires data to leave your environment should require explicit approval from your security and compliance leadership, not just a contract review.
- If your LLM strategy requires model control or local inference: A platform that runs on vendor-managed infrastructure does not satisfy this requirement, regardless of partnership announcements.
- If you want both workflow execution and knowledge discovery: The two are not mutually exclusive. The right architecture may include both a workflow execution layer and a separate, privacy-first knowledge layer. Buying one platform and hoping it solves both problems is a common mistake.
Frequently asked questions
1. What is ServiceNow Otto?
ServiceNow Otto is a unified AI experience announced at Knowledge 2026 on May 5, 2026. It combines Now Assist, the Moveworks conversational AI platform, and ServiceNow’s AI Experience layer into a single interface. Employees submit requests in natural language and Otto routes them across systems and completes work end-to-end, governed by ServiceNow’s AI Control Tower.
2. How does ServiceNow Otto handle data privacy?
Otto is a SaaS platform, meaning enterprise data is processed within ServiceNow’s cloud infrastructure. For organizations with data sovereignty requirements or those in regulated industries requiring that data never leave their own environment, this architecture requires careful evaluation as it does not support private cloud or air-gapped deployment. For organizations where this is a hard requirement, see Atolio’s guide to sovereign AI enterprise search.
3. Is ServiceNow Otto available now?
Otto is currently available through two products: ServiceNow EmployeeWorks and the AI Control Tower interface. Broader rollout across ServiceNow’s product portfolio is expected over the course of 2026. The AI specialists announced for CRM, legal, finance, and other functions are not yet widely deployed.
4. Does ServiceNow Otto work for all enterprise functions?
Not yet. Otto’s deepest functionality is in IT service management and HR, reflecting the heritage of both ServiceNow and the Moveworks acquisition. The “any function” positioning reflects a roadmap, not the current product. Functions like legal, R&D, marketing, and finance planning have specialists announced but not yet broadly shipped.
5. How does ServiceNow Otto compare to enterprise search tools like Atolio?
They solve adjacent but distinct problems. Otto is an action completion platform: it executes workflows, routes requests, and completes tasks, primarily within the ServiceNow ecosystem. Atolio is a knowledge discovery platform: it surfaces information, experts, and answers across every enterprise system, deployed entirely within the customer’s own private cloud. Most organizations need both capabilities, and the two can coexist.
6. What is the difference between ServiceNow Otto and AI Control Tower?
Otto is the employee-facing experience: the interface people use to submit requests and get work done. AI Control Tower is the governance layer: it logs every AI interaction, enforces enterprise policies, detects security threats like prompt injections, and provides the audit trail. Otto runs through AI Control Tower. The Control Tower also governs AI agents from other platforms, not just Otto.
ServiceNow’s recent announcement is significant. It reflects a serious, well-resourced bet on a real problem. The completion problem is real. The governance problem is real. And for organizations already deeply invested in ServiceNow, Otto represents a meaningful step toward making that investment work harder.
The “any function, any system” framing reflects ambition. The current product is narrower, and that distinction matters for buyers evaluating fit today. And for organizations evaluating enterprise AI broadly, the questions of data sovereignty, permission depth, functional coverage, and model control matter as much as the question of workflow execution.
The goal here is not to discourage evaluation of Otto. It is to encourage evaluation of the right questions alongside it.
About Atolio
Atolio is the secure enterprise search platform trusted by leading organizations across financial services, healthcare, defense, and technology. Deployed entirely within the customer’s own cloud environment, Atolio indexes organizational knowledge across every system with full permission-awareness, surfacing the right answer to the right person without data ever leaving the customer’s infrastructure. Learn more at atolio.com.
