Our Take
Bain's $100B estimate rests on converting manual coordination work into software spend, but 90% remains theoretical with only $4-6B captured today.
Why it matters
SaaS companies need concrete frameworks for identifying which workflows justify agent development investment versus which remain too complex or risky to automate profitably.
Do this week
Product teams: Map your customer workflows at subprocess level this month so you can identify automation opportunities before competitors claim adjacent workflow territory.
Bain estimates $100B US market for agentic automation
Bain & Company projects a $100 billion US market for SaaS companies deploying agentic AI to automate coordination work between enterprise systems (per company report). The consulting firm's analysis targets manual workflows spanning ERP, CRM, and support systems where employees pull data from multiple sources, interpret unstructured messages, and make approval decisions.
Current market capture sits at $4-6 billion, leaving over 90% untapped (company estimate). Sales represents the largest single function at $20 billion addressable market, driven by workforce size rather than exceptional automation potential. Operations and cost of goods sold account for $26 billion combined.
Customer support and R&D show highest automation rates at 40-60% of workflow tasks, while legal sits lowest at 20-30% due to error consequences requiring human oversight. The firm extends the global opportunity to $200 billion when including Canada, Europe, Australia, and New Zealand.
Coordination workflows create new software categories
Bain argues agentic AI addresses limitations of rules-based automation in workflows involving ambiguity and distributed information. The opportunity lies not in replacing existing SaaS platforms but converting labor-intensive coordination into software spending.
Six automation factors determine workflow viability: output verifiability, failure consequences, digitized knowledge availability, process variability, integration complexity, and regulatory requirements. Workflows with clear verification signals like compiled code or reconciled invoices automate more easily than subjective judgment tasks.
The report cites revenue examples including Cursor at $16.7 million monthly average, Sierra crossing $150 million annually, Harvey at $190 million, and Glean at $200 million (all company-reported figures). GitHub exemplifies using core workflow data to expand into adjacent automation areas.
Assessment starts at subprocess level
Bain recommends SaaS companies assess automation potential at subprocess granularity rather than treating entire functions as equally automatable. Companies should evaluate whether their data is comprehensive, outcome-tied, and automation-ready.
Three capability-building approaches emerge: internal development (AppLovin's Axon platform), acquisitions (ServiceNow buying Moveworks), or partnerships (Salesforce with Workday). Requirements include AI engineering talent, cloud-native architecture for multi-agent orchestration, and funding for model training and inference.
Pricing models shift from seat-based to outcome-driven when agents deliver completed work products. David Crawford, chairman of Bain's global technology practice, frames the competitive advantage as "cross-workflow decision context" spanning multiple systems rather than single-platform control.
The firm warns SaaS companies face a timeline "measured in quarters, not years" as AI-native competitors accumulate deployment data with each automated customer workflow.