The most revealing demo in enterprise artificial intelligence may be a password reset. It is boring, frequent and surprisingly difficult. The assistant must know who is asking, understand the request, check permissions, call the right identity system, complete the change and record what happened. A fluent paragraph does not count as success. The employee needs to log in.
Aisera built its company in that gap between a good answer and a finished task. Founded in 2017 by Muddu Sudhakar, a serial entrepreneur whose earlier companies were acquired by EMC, VMware and Splunk, the Santa Clara software maker began with AI service management. It placed a conversational layer in front of help desks, knowledge bases and ticket queues, then connected that layer to the software where work gets done.
The wedge was practical. Corporate support is full of small requests that ricochet between portals: provision an application, unlock an account, find a policy, open a case, approve leave, trace an order. Aisera lets a worker start in familiar channels such as Microsoft Teams, Slack, email, web chat or voice. Behind the conversation, the platform can search company knowledge, interpret intent and trigger a workflow in systems including ServiceNow, Workday, Salesforce, Zendesk, Jira and identity-management tools.
01 / The jobSoftware for the moments between systems
That position explains where Aisera fits in the market. It is partly conversational AI, partly enterprise search, partly IT service management and partly workflow automation. It does not need to replace a company's systems of record. It tries to become the front door to them. The user asks for an outcome; the platform finds the knowledge, application and sequence required to deliver it.
The company expanded well beyond IT. Its current catalog covers employee support, HR onboarding and leave, finance analysis and expenses, procurement and contracts, customer service, retail order tracking, healthcare scheduling and public-sector knowledge access. Human agents can use Agent Assist for summaries, similar cases and suggested next actions. Operations teams can use AIOps features to cluster incidents, predict failures and generate remediation playbooks. Builders get Agent Composer, prompt and event studios, low-code Hyperflows and an agent library.
“AI should make work more meaningful, not less human.”Aisera's statement of mission
The phrase is softer than the software. In practice, Aisera is selling fewer handoffs, shorter queues and less time spent hunting for the correct portal. The target buyer is generally a large organization where those minutes compound: a railway, hospital, retailer, bank, telecom company, university or government agency. Named customers have included Adobe, Workday, T-Mobile, Gilead Sciences, Amgen, BNSF Railway, Zoom, Grant Thornton, NJ Transit, OmniTRAX, Big 5 Sporting Goods and BDO Canada.
Its case studies read like a census of uncelebrated office friction. NJ Transit named its internal assistant Travis and reported a 60 percent increase in agent productivity. OmniTRAX said it auto-resolved 70 percent of tickets. Big 5 Sporting Goods reported saving 24,000 user hours a year. LifeScan said 65 percent of incoming support requests were automatically resolved, producing $2.2 million in support-cost savings. These are vendor-published customer results, useful as examples rather than universal forecasts.
02 / The differenceNot another all-knowing oracle
Aisera's differentiation is less about owning the smartest general-purpose model than about operating around models. Enterprises can use Aisera's domain-specific language models, bring their own or route requests to outside foundation models. An LLM gateway can select among them. Knowledge graphs and permission-aware search ground answers in organizational context. More than 100 integrations connect the conversation to actual records and tools.
The company also wraps autonomy in a governance idea called TRAPS: Trusted, Responsible, Auditable, Private and Secure. The name sounds like a warning label because it is one. A system allowed to approve, provision and modify cannot be treated like a casual text generator. Audit trails, access control, model choice, observability and policy enforcement are product requirements, not paperwork added at the end.
Answers the question
Retrieves an article, offers a link, collects a form or hands the conversation to a person.
Completes the request
Plans the steps, calls governed tools, coordinates systems, confirms the result and escalates exceptions.
The competitive field is crowded. ServiceNow can place AI directly inside the dominant ITSM suite and acquired Moveworks to deepen its conversational reach. Microsoft has Copilot Studio and the distribution advantage of Teams. Salesforce offers Agentforce near customer records. Kore.ai, Amelia, Espressive and Rezolve.ai sell overlapping conversational and service-automation products, while UiPath and Automation Anywhere approach the same work from process automation.
Aisera's answer is openness and specialization. Its 2025 Unify product was designed to coordinate Aisera and third-party agents rather than insist that one vendor own every task. It supports emerging protocols including Model Context Protocol, Agent-to-Agent and AGNTCY. A closed-loop engine can discover available agents, build a plan, route work and watch the outcome. The ambition is not one bot with an implausible résumé. It is a managed team of narrow agents.
03 / The economicsA large bet on invisible labor
Investors funded that ambition in stages. Norwest Venture Partners led a $20 million Series B in February 2020. Icon Ventures led a $40 million Series C in April 2021. Goldman Sachs Asset Management and Thoma Bravo led a $90 million Series D in August 2022. Those three disclosed rounds total $150 million, excluding the undisclosed Series A.
Aisera sells enterprise software through negotiated contracts rather than a public price card. The business resembles a familiar software-as-a-service model, with deployment, integration and customer-success work around the subscription. The economic case rests on avoided tickets, higher agent productivity, faster resolution and fewer overlapping tools. After the acquisition, Automation Anywhere has also promoted outcome-based pricing - charging for work performed instead of underused seats.
That shift matches the product's philosophy. A seat is easy to count, but an automated outcome is what the buyer wanted in the first place. It also raises the standard of proof. A useful agent must be reliable enough to act, cheap enough to beat the manual process and observable enough that a customer can see what it did.
04 / The next ownerConversation meets process automation
In November 2025, Automation Anywhere acquired Aisera for an undisclosed amount. The logic was unusually legible. Aisera brought natural-language understanding, enterprise search and self-service agents. Automation Anywhere brought a broad process-automation system, an orchestrator and experience drawn from hundreds of millions of automations. One side understood what the person wanted. The other had spent two decades making software execute repeatable work.
Aisera's more than 100 AI engineers joined the parent company. Automation Anywhere said it would continue product development and customer support, while the Aisera website and product identity remained active. In 2026, the company referred to Aisera as part of its Agentic Process Automation system. IDC named Aisera a Leader in its inaugural MarketScape for conversational AI platforms used in back-office work, specifically noting flexible model options, connectors and no-code workflows.
The strategic question is no longer whether a bot can talk. It is whether a company will trust it to touch the machinery.
For customers, the practical use remains recognizable. An employee can ask for software access without learning the service catalog. A support agent can receive a concise case brief rather than reread the history. An operations team can spot related incidents before the ticket pile becomes an outage. An HR department can route onboarding steps through several systems from one conversation. A customer can track an order without waiting for a representative.
None of this makes the complexity disappear. It relocates it. Connectors must stay current, permissions must be precise, knowledge must be clean and exception handling must work. Agentic AI adds another concern: a confident system can now do the wrong thing, not merely say it. Aisera's future depends on turning its governance language into dependable daily behavior across messy, old and highly specific enterprise stacks.
Still, the company chose a revealing place to begin. The help desk is where grand claims meet ordinary reality. Either the employee can log in, the application appears, the leave is routed and the ticket closes - or it does not. Aisera's story is the enterprise AI market in miniature: conversation was the opening act; accountable execution is the business.