Bain & Company has estimated a US$100 billion market in the US for SaaS companies using agentic AI. The firm said the market is tied to automating coordination work in enterprise systems.
The estimate comes from the second report in Bain’s five-part series on the software industry in the age of AI. The report examines where agentic AI could create new software markets and how SaaS companies can capture them.
Coordination work in enterprise systems
Bain said the market lies in the manual work employees perform between enterprise applications. These workflows often span ERP, CRM and support systems. They may also involve vendor management tools and email.
That work includes pulling data from one system and checking it against another source. It can also involve interpreting unstructured messages and deciding whether to approve, respond, escalate, or wait.
Bain said rules-based automation and robotic process automation are limited in workflows involving ambiguity and information spread in multiple systems. Agentic AI can interpret information from different sources, coordinate actions in systems, and operate in policy guardrails.
The report argues that agentic AI is not primarily a replacement for SaaS platforms, but that the market comes from converting labour-intensive coordination work into software spending.
It estimates vendors are already capturing US$4 billion to US$6 billion of the US market. More than 90% remains untapped, according to the firm.
Outside the US, Bain estimated that Canada, Europe, Australia, and New Zealand could add a similar-sized market. That would bring the total in those regions and the US to about US$200 billion.
Market size by function
The market is not evenly distributed in enterprise functions. Bain estimates that sales represents the largest single share at about US$20 billion. This is mainly due to the number of sales employees, not unusually high automation potential.
Cost of goods sold and operations account for about US$26 billion. The large size of the operational workforce means even modest automation rates can translate into a large addressable market. R&D and engineering, customer support, and finance each represent about US$6 billion to US$12 billion in addressable market size. These functions have sizeable workforces and higher automation potential in specific workflows.
Customer support and R&D or engineering have the highest automation potential, with roughly 40% to 60% of workflow tasks automatable. Bain said both areas have structured data, standardised processes, and clearer output signals. Finance and human resources fall in the 35% to 45% range. The report said accounts payable and payroll have higher automation potential, while financial planning and employee relations involve more judgement.
Sales and IT sit at 30% to 40%. Bain pointed to relationship nuance, deal-by-deal variation, and the unpredictable nature of security incidents as limits on automation in those areas. Legal has lower overall automation potential, at 20% to 30%. Bain said contract review and compliance are repeatable, but the consequences of errors create a need for tighter oversight.
Bain’s automation factors
The report identifies six factors that determine how much of a workflow can realistically be handled by an AI agent. They include output verifiability, consequence of failure, digitised knowledge availability, and process variability. Bain said workflows with clear verification signals are easier to automate than work involving subjective judgement. Examples include compiling code, reconciled invoices, and resolved support tickets.
Workflows involving regulatory or financial risk require closer human supervision, even where agents are technically capable, according to the report. These include tax filings, legal compliance, and security incident response.
Bain also identified digitised knowledge availability as a constraint. Agents need access to structured data and documented context. They also need machine-readable inputs, including decision logic that often sits informally with experienced employees.
Integration complexity affects automation when workflows pass through several systems and APIs. Authentication layers and exception-handling processes add further complexity, and these workflows are harder to automate end-to-end than workflows contained in a single platform. The highest-value areas are concentrated where no single system of record controls the full outcome. These workflows often span ERP, CRM and support systems, the company says.
David Crawford, chairman of Bain’s global technology and telecommunications practice, said SaaS companies have spent the past two decades building positions around systems of record with the next source of advantage being “cross-workflow decision context,” which is defined as the ability to interpret and act in workflows that move through multiple systems.
Company examples and adjacent workflows
The report cited Cursor, Sierra, Harvey, Glean, Salesforce, ServiceNow, and Workday in its discussion of agentic AI adoption. Cursor has surpassed US$16.7 million in average monthly revenue, according to Bain, after doubling in a single quarter. Sierra has crossed US$150 million per annum, Harvey passed US$190 million pa, and Glean US$200 million pa.
The report also pointed to GitHub for example of a company using data from an existing core workflow to move into adjacent work. GitHub’s core business is developer collaboration and source control, but its repository and workflow data helped support expansion into AI-assisted developer productivity and security automation.
Bain said SaaS companies can expand through two types of workflow automation. The first is automating core workflows, where they already have domain knowledge and customer trust. Bain said existing system integrations can support automation of core workflows. The second is automating adjacent workflows that the company does not currently serve directly. These areas can be harder to identify because they require detailed mapping of customer workflows and the underlying data that supports decisions.
Pricing models can change when agents deliver completed outcomes. Bain said outcome- and use-based pricing can become more relevant when agents resolve issues or process invoices. The report contrasts this with traditional pricing based on seats and logins.
Bain’s recommendations for SaaS companies
Bain recommended that SaaS companies begin by identifying which customer workflows are now automatable with agentic AI. The firm said companies should assess automation at the subprocess level not treating entire functions as equally automatable.
The report also said companies should assess the quality of their data. Bain said relevant factors include whether the data is comprehensive, tied to outcomes, and usable for automation.
Bain said companies could close ability gaps through internal development, acquisitions, or partnerships. The report cited AppLovin’s in-house development of its Axon platform, ServiceNow’s acquisition of Moveworks, and Salesforce’s partnership with Workday as examples of different approaches.
The firm also pointed to the need for AI engineering talent, cloud-native architecture for multi-agent orchestration, and funding for model training and inference. It said companies should align pricing and sales incentives with AI-driven outcomes not legacy seat-based models.
Bain said SaaS companies will also need data and product foundations designed for agentic workflows, including machine-readable hand-offs and systems that capture decisions and outcomes from each workflow run.
Crawford said the timeframe for SaaS companies is “measured in quarters, not years,” as AI-native companies gather more deployment data with each customer workflow they automate.
(Photo by engin akyurt)
See also: Google tests Remy AI agent for Gemini as focus turns to user control
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.
AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
