Before an enterprise can trust AI to do the work, someone has to know exactly how the work is done. 8Flow watches the clicks, tabs and copy-pastes that make up real jobs - then turns that map into deployable AI agents.
8Flow, Palo Alto. The company's mark, photographed for the record. A 17-person team behind a platform that has logged more than 600 million moments of enterprise work.
There is a quiet, expensive problem inside most large companies: nobody can say precisely how work gets done. A support agent resolving a refund might touch six systems, a spreadsheet, and two browser tabs in ninety seconds - and none of it is written down. 8Flow was built to see that hidden choreography, and to decide what a machine should take over next.
8Flow is an enterprise software company based in Palo Alto, California, founded in 2022. Its product, the Workflow Context Platform, observes how individual employees actually complete tasks across the systems they already use - ServiceNow, Salesforce Service Cloud, Zendesk and the sprawl of spreadsheets, databases and internal tools around them.
Rather than asking teams to document their processes, 8Flow learns them by watching. The platform performs task mining at the micro level, capturing the granular steps that make up a workflow. It then runs continuous process discovery to reveal bottlenecks, quantifies where time and money leak out, and produces what it calls Agentic Blueprints - implementation-ready specifications for deploying AI agents against a specific piece of work.
The distinction 8Flow draws is deliberate. Traditional process mining tells you what happened - which steps ran, in what order. 8Flow aims to show the mechanics of how work is executed by real people, the part that usually stays invisible. That difference is the company's central bet: that the bottleneck in enterprise AI is not the model, it is the map.
"Enterprises know they need AI but don't know where to apply it," says co-founder and CEO Boaz Hecht. The Workflow Context Platform, in his framing, reveals how work really happens - and therefore where AI can drive impact instead of guesswork.
8Flow often arrives quietly - as a Chrome extension or desktop app that learns alongside a team, no rip-and-replace required. The pipeline runs from passive observation to concrete automation.
A browser extension and desktop apps capture the real micro-steps agents take across every system.
Task mining maps those steps into workflows, exposing bottlenecks and repeated patterns.
The Opportunities view ranks where time and dollars are recoverable, with hard numbers.
Agentic Blueprints turn the map into implementation-ready AI agents, exposed via 8Flow's MCP.
* Figures cited by 8Flow across a demonstrated set of workflows; savings are estimates, not audited results.
The core system capturing micro-workflows for continuous process discovery, people intelligence and automation-opportunity identification.
Implementation-ready blueprints, generated from observed human work, that specify how to deploy an AI agent for a given workflow.
A Model Context Protocol interface exposing 8Flow's workflow context to AI models and agent frameworks.
The original engine that learns from agent actions and generates automations across ticketing, CRM, spreadsheets and databases.
A Chrome extension plus Mac and Windows apps that observe workflows in context - without replacing existing systems.
Analytics that quantify recoverable hours and dollars and rank high-impact automation opportunities for managers and executives.
8Flow sells to enterprise operational teams: customer support, IT service management and back-office functions - the departments where high-volume, repetitive, multi-system work concentrates. Early adopters include support organizations such as FloorFound, whose team credited 8Flow with cutting the constant switching between browser tabs.
The business model is B2B subscription SaaS. Crucially, 8Flow integrates with the systems a company already runs rather than trying to replace them - a posture that lowers the barrier to a first deployment and reflects its founders' enterprise-platform instincts.
The competitive field is crowded with process-intelligence and task-mining vendors - UiPath, Celonis, Nice and Microsoft's Power Automate Process Mining among them - plus a wave of customer-support AI tools. 8Flow's differentiation is its focus on the how of work at the individual level, and on translating that directly into agent deployment rather than static reports.
It also sits, by design, at a convergence point: automation, productivity and AI enablement. That positioning is what drew strategic investors ServiceNow and Okta, whose own platforms 8Flow both observes and complements.
8Flow's expertise is narrow and deep: understanding the texture of frontline enterprise work. The founders spent years inside ServiceNow, one of the largest workflow platforms in the world, and that background shows in the product's instincts - it treats the messy reality of a support desk as the source of truth, not an inconvenience to be standardized away.
Most analytics tools reason from system logs, which capture the events an application chose to record. 8Flow instead watches the human layer above those systems: the tab a person opens, the value they copy, the field they check twice. That vantage point is what lets it surface work that never touches a formal audit trail - the shadow processes that quietly consume an organization's hours.
The business model follows the same logic. 8Flow is sold as a subscription to operational departments and is designed to land small and expand: begin with a browser extension observing one team, prove recoverable time and dollars in the Opportunities view, then widen into blueprints and deployed agents. Because it augments incumbent systems rather than replacing them, adoption does not require a disruptive migration.
Compliance is part of the pitch too. SOC 2 Type II and HIPAA certification signal that 8Flow intends to operate inside regulated support and back-office environments, where observing employee work carries real governance weight. For a company whose whole method is watching, earning that trust is not optional - it is the product.
8Flow's three co-founders worked together long before this company - first at SkyGiraffe, then through its 2017 acquisition by ServiceNow, where CEO Boaz Hecht went on to serve as VP of Platform.
Co-Founder & CEO. Previously co-founder and CEO of SkyGiraffe (acquired by ServiceNow, 2017) and later VP Platform at ServiceNow.
Co-Founder. An early SkyGiraffe employee who continued through the ServiceNow acquisition.
Co-Founder. Joined SkyGiraffe early and followed the team into ServiceNow before starting 8Flow.
ServiceNow and SkyGiraffe veterans Boaz Hecht, Josh Russ and Yev Goldin start the company in Palo Alto.
8Flow emerges with a self-learning automation engine for customer support, backed by Caffeinated Capital and notable angels.
Chrome extension and desktop apps extend task mining across ServiceNow, Salesforce Service Cloud and Zendesk.
Strategic backing from ServiceNow Ventures and Okta Ventures; launch of the people-and-process-intelligence platform, showcased at CCW Nashville.
It maps how work actually gets done across enterprise systems, performs task mining and process discovery, quantifies automation opportunities, and generates blueprints for deploying AI agents.
It was founded in 2022 by Boaz Hecht (CEO), Josh Russ and Yev Goldin, who previously worked together at SkyGiraffe and ServiceNow.
About $16.6M total: a $6.6M seed round in 2023 and a $10M Series A in 2025 led by Caffeinated Capital with ServiceNow Ventures and Okta Ventures.
It integrates with platforms including ServiceNow, Salesforce Service Cloud and Zendesk, and works alongside AI providers like Anthropic and OpenAI.
Traditional process mining shows what happened; 8Flow captures the micro-level mechanics of how work is executed by individuals, then turns that into automation and AI-agent opportunities.
Sources: 8flow.ai, TechCrunch, VentureBeat, Process Excellence Network, Crunchbase, LinkedIn. Figures are drawn from public reporting; financial estimates are approximate.