Agentic AI M&A: what is happening now?

Is agentic AI entering a major acquisition wave?

Agentic AI is going through a major acquisition wave, with Salesforce, ServiceNow, Workday and other enterprise software companies spending billions to buy AI agents, automation technology and the systems needed to make them work.

ServiceNow completed its $2.85 billion acquisition of Moveworks in December 2025, Workday bought Sana for approximately $1.1 billion, and Salesforce completed its $3.6 billion purchase of Fin in September 2026. Together, those three transactions alone represent roughly $7.55 billion in acquisition spending.

The activity has continued. Automation Anywhere announced an agreement to acquire Boost.ai, Workday agreed to buy AI identity company C1.ai, and Rocket Software purchased Geniez AI to automate mainframe operations.

What's interesting is how different these targets are. Moveworks helps employees find information and complete tasks, Fin handles customer-service conversations, and C1.ai manages access permissions for people and AI agents.

Enterprise software companies are buying several pieces of the same emerging market. Some want the interface employees interact with, others want the agents performing tasks, and others want control over the systems those agents access.

If you want more recent data on this point, see the latest figures in our Agentic AI report ↓.

How much money is going into agentic AI acquisitions?

Agentic AI M&A already involves several billion dollars in confirmed deal values, although the total becomes much larger when we include acquisitions of data, cybersecurity and other technologies needed to run AI agents.

The three major transactions involving Moveworks, Sana and Fin total approximately $7.55 billion. Adding Salesforce's roughly $8 billion acquisition of Informatica brings that selected group to $15.55 billion.

ServiceNow's approximately $7.75 billion deal for cybersecurity company Armis would push the combined figure to about $23.3 billion. However, counting the full value of Informatica and Armis as agentic AI spending would be misleading. Both businesses serve much broader markets.

Many smaller acquisitions also have undisclosed prices, making an accurate market-wide total impossible to calculate from publicly announced figures alone.

The following transactions show the range of businesses attracting buyers, from established agent platforms to the infrastructure supporting them.

Buyer	Target	Announced value	Main business
Salesforce	Informatica	~$8.0B	Data management
ServiceNow	Armis	~$7.75B	Cybersecurity
Salesforce	Fin	~$3.6B	Customer-service agents
ServiceNow	Moveworks	$2.85B	Enterprise AI assistant
Workday	Sana	~$1.1B	Knowledge and AI agents
Workday	Paradox	Undisclosed	Recruiting automation
UiPath	WorkFusion	~$189.5M	Financial-compliance agents
Snowflake	Natoma	Undisclosed	Secure agent connectivity
Automation Anywhere	Boost.ai	Undisclosed	Voice and conversational AI
Workday	C1.ai	Undisclosed	Agent identity and permissions

The announced values represent transaction consideration rather than necessarily the final accounting cost. UiPath's $189.5 million WorkFusion figure comes from its regulatory filing. The table also includes deals at different stages, including completed acquisitions and agreements awaiting closure.

Has agentic AI M&A accelerated lately?

Agentic AI acquisitions have spread across more types of enterprise software companies, although public transaction data does not yet support a reliable year-over-year growth rate.

Salesforce provides a useful example. In 2025, it acquired Convergence.ai, Regrello, Apromore and Informatica, covering AI agents, workflow automation, process analysis and enterprise data.

Its buying continued in 2026 with Fin, followed by the announcement of its agreement to acquire Listen Labs, an AI-powered customer research company.

Other buyers have become increasingly active. UiPath acquired WorkFusion in February 2026, Snowflake announced Natoma in May, and Automation Anywhere agreed to acquire Boost.ai in October. Workday announced C1.ai, while Rocket Software acquired Geniez AI.

Across these two years, at least seven established enterprise technology companies announced or completed acquisitions directly relevant to agentic AI: Salesforce, ServiceNow, Workday, UiPath, Snowflake, Automation Anywhere and Rocket Software.

The activity now reaches well beyond the CRM and IT-service-management markets where some of the earliest large transactions occurred.

We should still be careful about claims of explosive deal-count growth. Many companies now describe ordinary automation, data integration and security acquisitions as agentic AI deals. Without applying the same definition across several years, a precise growth percentage would give a false impression of accuracy.

Why is Salesforce buying so many AI companies?

Salesforce is buying AI companies to turn Agentforce into a platform that can understand customers, access business information and complete tasks across enterprise applications.

Its acquisition history shows how the strategy has developed. Convergence.ai brought technology for agents navigating digital interfaces. Regrello added structured workflow automation. Apromore contributed process analysis, helping companies understand how work moves through their systems.

Informatica added a much larger data-management business, while the $3.6 billion acquisition of Fin brought an established customer-service AI platform.

More recently, Salesforce agreed to acquire Listen Labs, whose agents conduct customer interviews, analyze responses and automate research. According to Salesforce's September 2026 announcement, the company can recruit participants from a network exceeding 50 million people and support interviews in more than 120 languages.

The combination looks deliberate. Salesforce already owns major customer-data and workflow products. By acquiring agent execution, customer-service automation and customer intelligence, it can offer more of the process inside its own ecosystem.

There is a defensive element too. If customers begin using independent AI agents to manage sales, service and customer relationships, Salesforce risks losing some control over the software experience. Buying these capabilities helps it remain the platform through which that work happens.

If you want more recent data on this point, see the latest figures in our Agentic AI report ↓.

Why did ServiceNow pay $2.85 billion for Moveworks?

ServiceNow bought Moveworks because it had already built a widely deployed enterprise AI assistant, giving ServiceNow immediate access to customers, technology and an employee-facing product.

According to ServiceNow's December 2025 acquisition announcement, Moveworks had approximately 5.5 million employee users and around 250 customers that already used both companies' technologies.

ServiceNow also reported that nearly 90% of Moveworks customers had deployed its technology to their entire employee population. That measures deployment coverage rather than daily usage, but it still suggests the product had moved well beyond small departmental experiments.

The existing customer overlap is particularly interesting. Around 250 organizations were already using both companies, providing ServiceNow with a group where deeper integration could be commercially straightforward.

Moveworks also brought more than 100 technology integrations, allowing employees to search information and trigger actions across different business applications.

For ServiceNow, developing another AI assistant internally would have been possible. Acquiring Moveworks gave it an established product, enterprise deployment experience and a substantial user base immediately.

Those assets together help explain the $2.85 billion price.

Are AI customer-service startups becoming the hottest acquisition targets?

Customer-service AI has become one of the most valuable areas of agentic AI M&A, with Salesforce's $3.6 billion acquisition of Fin showing how much buyers will pay for a working product with an established customer base.

Fin had more than 30,000 business customers when Salesforce completed the acquisition in September 2026. Salesforce also reported an average AI resolution rate of 76% across customer inquiries handled by the platform.

That figure is company-reported and does not independently establish performance across every customer or type of inquiry. Still, Fin came with something important: an existing product handling real customer-service workloads.

Automation Anywhere's acquisition agreement with Boost.ai points in the same direction. Boost.ai specializes in conversational and voice AI, extending Automation Anywhere's capabilities into customer-facing interactions.

These acquisitions make commercial sense because customer support offers relatively clear ways to assess performance. Businesses can measure the number of inquiries resolved, the cost per interaction, customer satisfaction and the amount of work still requiring human intervention.

The attraction also extends beyond customer support. Workday acquired Paradox to automate recruiting conversations and scheduling, while UiPath bought WorkFusion for financial-compliance workflows.

Buyers are concentrating on applications where the agent has a defined job and companies already spend money getting that job done.

Why are AI-agent security startups suddenly acquisition targets?

AI-agent security has become an important acquisition category because companies need much tighter control over software that can access records, change information and perform actions automatically.

Three transactions help explain what's happening.

ServiceNow acquired Veza to strengthen identity governance across enterprise systems. Snowflake announced its acquisition of Natoma to improve secure connections between AI agents and business applications. Workday then agreed to acquire C1.ai, formerly ConductorOne, to manage permissions for both employees and autonomous agents.

These companies approach the problem from different directions. ServiceNow manages IT workflows, Snowflake handles enterprise data, and Workday manages employees and business operations. All three have decided that controlling agent access deserves dedicated technology.

Workday's October 2026 announcement highlighted research indicating that 95% of large US companies already run autonomous agents, while Deloitte's research found that only 21% of organizations reported mature agent governance. The figures come from different research sources and should not be treated as directly comparable samples, but they illustrate the gap companies are trying to address.

An employee might need permission to read a payroll record. An AI agent could need permission to retrieve the same record, update another system and trigger an approval process. Companies must decide exactly which actions are authorized and when that access should expire.

That creates an attractive market for startups offering agent identity, access management and security monitoring.

Are software companies buying AI infrastructure instead of chatbots?

Enterprise software buyers are increasingly purchasing the infrastructure behind AI agents, particularly workflow orchestration, data access, identity management and security.

Looking at the acquisitions together, several layers are taking shape.

Moveworks gives employees a way to interact with AI through workplace search and assistance. Convergence.ai contributes technology for performing tasks through digital interfaces. Regrello automates structured business processes. Apromore helps analyze those processes.

Further down, Informatica provides data management, Natoma offers controlled connections to applications, and C1.ai manages identity and access rights.

These capabilities work together. A customer-service agent might need to identify a customer, retrieve an order, issue a refund and update a CRM record. Each action depends on different systems and permissions.

Buyers appear to expect enterprise agents to become much more involved in daily operations. Owning the systems that support and control those agents could become just as important as owning the agents themselves.

Agent capability	Acquisition example	What it provides
Employee assistance	ServiceNow / Moveworks	Enterprise search and task initiation
Workplace knowledge	Workday / Sana	Search, learning and knowledge agents
Digital task execution	Salesforce / Convergence.ai	Actions across software interfaces
Workflow automation	Salesforce / Regrello	Multi-step business processes
Process analysis	Salesforce / Apromore	Understanding existing workflows
Data management	Salesforce / Informatica	Access to governed business data
Agent connectivity	Snowflake / Natoma	Controlled connections to tools
Agent permissions	Workday / C1.ai	Identity and access governance
Compliance automation	UiPath / WorkFusion	Specialized financial investigations

If you want more recent data on this point, see the latest figures in our Agentic AI report ↓.

Are specialized AI-agent startups more attractive than general-purpose agents?

Specialized AI-agent companies are becoming particularly attractive acquisition targets because they already understand the workflows, customers and rules of a specific industry.

UiPath's acquisition of WorkFusion is a good example. WorkFusion builds AI agents for financial-crime compliance, including anti-money-laundering investigations and customer screening. Banks already have teams, software budgets and regulatory responsibilities dedicated to these activities.

Workday's acquisition of Paradox followed a similar principle in recruitment. According to Workday's announcement, Paradox had supported more than 189 million AI-assisted candidate conversations. Its technology could automate recruiting interactions and scheduling within an established hiring process.

Rocket Software's acquisition of Geniez AI applies the same idea to mainframe operations. Geniez AI builds agents that help monitor complex systems, diagnose problems and execute technical workflows.

The industries are very different, but each acquisition gives the buyer expertise that would take time to reproduce.

General-purpose agents can certainly perform some of these tasks. However, specialized companies may have better integrations, operational knowledge and established relationships with customers who already trust them.

Those advantages look especially useful in regulated or technically complex industries.

Are AI-agent startups being bought for their technology or their customers?

Agentic AI acquisitions are increasingly about buying technology and customer access together, although smaller deals are often driven more heavily by engineering expertise.

Moveworks is a good example of a business that offered both. Alongside its technology, it brought millions of deployed employee users, hundreds of integrations and relationships with large enterprise customers.

Fin offered Salesforce a similar combination in customer service, with more than 30,000 customers and its own specialized AI models. That installed base would have been difficult to reproduce quickly through internal product development.

Other acquisitions look more technology-driven. Convergence.ai contributed expertise in agents that navigate and operate digital interfaces. Rocket Software explicitly highlighted the technical leadership and engineering team joining through Geniez AI.

The distinction helps explain why acquisition values vary so widely.

A young startup with a small customer base can still attract a buyer if its engineering team solves a difficult technical problem. A more established agent company can command a much higher price when it also brings revenue, enterprise contracts and a product already deployed at scale.

Buyers are purchasing different combinations of technical expertise, products and distribution, and the balance changes from one deal to another.

Are acquired AI-agent startups delivering real products yet?

Several agentic AI acquisitions have already led to commercial products and integrations, although we still have much less evidence about whether the deals are financially paying off.

Workday completed its acquisition of Paradox in October 2025 and subsequently offered the Workday Paradox Candidate Experience Agent for purchase. The product brings automated candidate conversations and hiring assistance into Workday's recruiting portfolio.

Salesforce has also integrated technology acquired through Regrello into Agentforce Operations, extending its ability to automate back-office processes.

ServiceNow's Moveworks acquisition offers another useful indicator of commercial maturity. The acquired company already served millions of deployed employee users, and ServiceNow identified approximately 250 mutual customers at the time of completion.

These examples establish that buyers are acquiring working technology rather than relying entirely on research prototypes.

But there is an important financial gap. Salesforce, Workday and ServiceNow do not generally disclose enough acquired-product revenue, customer retention, operating costs or profits to calculate acquisition-level returns.

Product integration is encouraging, especially when it happens quickly. Whether the purchase prices were justified will depend on how much additional revenue and customer retention the acquired technology ultimately produces.

Are companies overpaying for AI-agent startups?

The largest agentic AI acquisition prices are aggressive, but we cannot conclude that buyers are systematically overpaying without better information about the targets' revenue and profitability.

Consider three transactions. Workday paid approximately $1.1 billion for Sana, ServiceNow paid $2.85 billion for Moveworks, and Salesforce agreed to pay around $3.6 billion for Fin.

These valuations are large, but the companies being acquired had different customer bases, products and levels of commercial maturity.

Moveworks brought a substantial enterprise deployment footprint. Fin had more than 30,000 business customers. Sana offered enterprise search, learning and AI-agent capabilities that could be integrated into Workday's existing platform.

The missing variable is revenue. Without comparable annual recurring revenue figures, gross margins and growth rates, calculating credible acquisition multiples is difficult.

There is another complication: AI-agent operating costs vary widely. A customer-service agent handling thousands of conversations can have a very different cost structure from a research agent performing long, complex tasks.

Buyers also consider the cost of building competing technology internally and the risk of allowing competitors to acquire it.

Established AI-agent products are clearly attracting billion-dollar valuations. Whether those prices are excessive remains an open question, especially when acquisitions are justified partly by future revenue that has not yet materialized.

If you want more recent data on this point, see the latest figures in our Agentic AI report ↓.

Can independent AI-agent startups still compete with Salesforce and ServiceNow?

Independent AI-agent startups can still compete, but they face a tougher market as Salesforce, ServiceNow and Workday bundle more automation capabilities into products their customers already use.

A company using Salesforce Service Cloud may find it easier to deploy a customer-service agent inside its existing CRM environment. Workday customers can increasingly access recruiting automation without buying a separate system. ServiceNow can connect AI assistance directly to its established IT workflows.

This gives large vendors an advantage in distribution and integration.

However, independent startups have several ways to compete. They can offer better performance on a specialized task, work across software platforms or serve customers whose needs are poorly covered by existing enterprise products.

Cross-platform compatibility could become particularly important. Many businesses use Microsoft, Salesforce, ServiceNow, SAP and other applications simultaneously. An agent that connects these systems reliably may have an advantage over one designed mainly for a single vendor's ecosystem.

The challenge for independent companies is demonstrating that their product delivers enough additional value to justify another software contract.

Generic assistant products will probably face the most pressure, while specialized automation, agent security and cross-platform infrastructure look like more defensible opportunities.

Which AI-agent startups could be acquired next?

The most plausible future agentic AI acquisition targets are companies working in security, specialized automation, customer service and orchestration, based on the types of businesses large software vendors have already purchased.

There is a reason to watch security companies such as Zenity, Noma Security, Aembit and Astrix Security. As seen above, the acquisitions of Veza, Natoma and C1.ai show that enterprise buyers are paying attention to the controls surrounding autonomous software.

Customer-service companies remain relevant too. Decagon, Sierra, Parloa and PolyAI operate in markets where the commercial value of automation can be measured through support volumes, resolution rates and operating costs.

Legal, healthcare and accounting agents offer another possibility. Companies such as Harvey, Legora, Hippocratic AI and Basis work in specialized environments where industry knowledge and integration with existing procedures can be valuable.

We can group the most relevant categories without pretending that these companies are currently negotiating sales.

Category	Examples to watch	Potential buyer interest
AI-agent security	Zenity, Noma Security	Monitoring autonomous agents
Agent identity	Aembit, Astrix Security	Controlling access and credentials
AI orchestration	CrewAI, LangChain	Managing complex agent workflows
Customer-service AI	Decagon, Sierra, Parloa	Automating customer support
Voice agents	PolyAI, Retell AI	Automating phone-based interactions
Accounting agents	Basis, Rillet	Financial workflow automation
Legal AI	Harvey, Legora	Legal research and documentation
Healthcare agents	Hippocratic AI, Infinitus	Specialized healthcare operations
Coding agents	Factory	Software development automation

These are examples of companies in strategically relevant categories, not confirmed sale candidates. Their ownership and independence would need to be checked again before any specific takeover assessment.

The companies with the greatest strategic appeal should be those possessing capabilities that buyers cannot easily reproduce: valuable customer relationships, established integrations, difficult technical expertise or access to specialized workflows.

If you want more recent data on this point, see the latest figures in our Agentic AI report ↓.

What is really happening in agentic AI M&A now?

Agentic AI M&A is becoming a major consolidation movement in enterprise software, with established vendors buying the technology, customers and infrastructure needed to make AI agents part of everyday business operations.

The size of the transactions already makes the shift hard to ignore. Moveworks, Sana and Fin alone account for approximately $7.55 billion in announced acquisition values, while the broader group of relevant buyers now includes Salesforce, ServiceNow, Workday, Snowflake, UiPath and Automation Anywhere.

The direction of the acquisitions is even more interesting than the prices. Early attention centered heavily on AI assistants and task automation. Recent deals increasingly involve customer-service platforms, industry-specific agents, secure connectivity and identity management.

Salesforce has been assembling a broad set of capabilities around Agentforce. ServiceNow is strengthening the connection between employee assistance, operational workflows and security. Workday is building automation around HR, knowledge and access management.

We now have enough repeated acquisitions across independent buyers to conclude that large enterprise software companies want direct ownership of agentic capabilities.

What remains uncertain is the financial outcome. Many of the acquired products are commercially available, but buyers have disclosed relatively little information about their returns on investment.

For independent startups, the market is becoming more demanding. A generic AI assistant will have difficulty standing out as established vendors add similar features. Companies with specialized workflows, deep integrations, proven customers or technology for controlling agents have a stronger position.

Agentic AI M&A has moved well beyond an experimental phase. The next question is whether these billion-dollar acquisitions will generate enough additional software revenue and customer value to justify what buyers are paying.