Bain: Agentic AI could create $100B US SaaS market

Bain & Company estimates agentic AI could create a roughly $100 billion addressable market in the US for SaaS vendors by automating manual coordination tasks that sit between enterprise systems such as ERP, CRM and support platforms. The estimate appears in the second report of Bain’s five-part series on software in the age of AI.

The report says the opportunity comes from work employees perform outside single systems of record, including pulling and reconciling data across platforms, interpreting unstructured messages like email or chat, and deciding whether to approve, escalate, respond or wait. Bain argues that rules-based automation and traditional robotic process automation struggle with ambiguity and dispersed information, while agentic AI can interpret multiple sources, coordinate actions across systems and operate within policy guardrails.

Bain estimates vendors have captured between $4 billion and $6 billion of the current US market and that more than 90% remains untapped. The consultancy projects that Canada, Europe, Australia and New Zealand could add a similar-sized opportunity, bringing a combined total for those regions and the US to about $200 billion.

The firm breaks the US addressable market down by function. Cost of goods sold and operations account for about $26 billion, and sales represents roughly $20 billion. R&D and engineering, customer support, and finance each represent about $6 billion to $12 billion of potential market size.

The report scores technical automation potential by function. Customer support and R&D or engineering show the highest potential at about 40% to 60% of workflow tasks, driven by structured data and clear output signals. Finance and human resources fall in the 35% to 45% range, with accounts payable and payroll more automatable than financial planning or employee relations. Sales and IT are estimated at 30% to 40% because of relationship nuances and unpredictable incidents, while legal is lower at 20% to 30% due to the consequences of error and the need for tighter oversight.

Bain identifies six factors that determine how much of a workflow an AI agent can handle: whether outputs can be verified, the consequence of failure, availability of digitized knowledge, process variability, integration complexity and exception handling. The report notes that workflows spanning several systems and authentication layers are harder to automate end-to-end than those contained in a single platform.

The report names Cursor, Sierra, Harvey, Glean, Salesforce, ServiceNow and Workday as examples of agentic AI adoption and highlights revenue milestones for some vendors. Cursor is reported to have reached about $16.7 million in average monthly revenue, Sierra more than $150 million per year, Harvey over $190 million per year and Glean about $200 million per year. GitHub is cited as an example of using repository and workflow data from a core product to expand into adjacent automation areas such as developer productivity and security.

Bain outlines two expansion paths for SaaS firms: automate core workflows where they already have domain knowledge and integrations, or map and automate adjacent workflows that require detailed mapping of customer processes and data. The report also says agentic agents that deliver completed outcomes make outcome- and use-based pricing more relevant than traditional seat- or login-based licensing.

The consultancy recommends SaaS companies assess automation opportunities at the subprocess level, evaluate the quality and machine-readability of their data, and close capability gaps through internal development, acquisitions or partnerships. AppLovin’s Axon platform, ServiceNow’s acquisition of Moveworks and Salesforce’s partnership with Workday are cited as different approaches. The report also calls for AI engineering talent, cloud-native architectures for multi-agent orchestration, funding for model training and inference, and product and data foundations that capture decisions and machine-readable hand-offs from each workflow run.

David Crawford, chairman of Bain’s global technology and telecommunications practice, wrote that the next source of advantage for SaaS vendors is ‘cross-workflow decision context’ and that the timeframe to capture these opportunities is ‘measured in quarters, not years.’