Robotic process automation keeps getting pronounced dead, and it keeps refusing to die. This week's flurry of market data and vendor analysis suggests the technology is instead undergoing a quiet absorption into the broader "agentic AI" wave — with real consequences for which vendors win and lose.

Vendor mindshare is fragmenting

Fresh PeerSpot engagement data updated this month shows a striking reshuffle inside the RPA-turned-"Agentic Automation" category. As of August 2026, in the Agentic Automation category, Automation Anywhere's mindshare sits at 14.7%, down from 17.8% a year earlier, Blue Prism has fallen to 9.9% from 20.0%, Microsoft Power Automate dropped to 15.4% from 29.9%, and UiPath declined to 19.1% from 28.1%. That every major legacy vendor lost ground simultaneously is the real story — it points to demand splintering across a growing field of AI-native challengers rather than consolidating around the old RPA "Big Three."

The market keeps getting bigger anyway

Despite that fragmentation, forecasters aren't backing off their growth numbers. A market report refreshed on August 25 pegs the combined global RPA and hyper-automation market at $93.04 billion in 2025, expanding at roughly 23.4% annually through 2034 to reach nearly $617.34 billion. Multiple other trackers, drawing on slightly different methodologies, put pure-play RPA software alone in the $8–35 billion range for 2026 — the wide spread itself illustrating how blurry the line between "RPA" and "AI agent" has become for analysts trying to size the category.

"AI-native" challengers are pressing the attack

The sharpest challenge to incumbents is coming from vendors arguing that bolting generative AI features onto screen-scraping bots isn't enough. Newer entrants frame legacy tools as fundamentally constrained: traditional RPA vendors have added document understanding and generative AI assistants on top of platforms architecturally rooted in screen-scraping, where bots still rely on UI selectors while AI sits layered on the workflow surface. Their pitch is a rebuild from the ground up, with automation logic expressed and audited in plain language rather than confidence scores — a direct shot at RPA's traditional black-box reputation.

Why incumbents aren't panicking (yet)

Even skeptics of the "agents replace RPA" narrative concede the old guard has adapted rather than collapsed. The lazy version of the story says AI agents replaced RPA, but that isn't what happened — RPA survived by being swallowed into broader platforms that now promise to read documents, reason over policies, invoke APIs, update systems, and hand work back to humans when confidence drops. For regulated industries especially, that hybrid framing — bots for deterministic execution, agents for judgment calls — is emerging as the practical consensus, even as the vendor rankings underneath it keep shifting.

Sources

Agentic AIDaily RPA UpdateEnterprise AutomationHyperautomationRPAUiPath