UiPath

UiPath spent a decade teaching software robots to click through the boring parts of enterprise work. Now it is making a much bigger bet: that AI agents can think, robots can do, and people can just lead. The company frames it as "Act 2," a shift from rule-based RPA to agentic automation, and it is putting the whole platform behind it.

The origin story is unusually good. UiPath started in 2005 as a 10-person shop in a small apartment in Bucharest, Romania, founded by Daniel Dines and Marius Tirca. It went public on the NYSE in April 2021 under the ticker PATH at a valuation around $29 billion. Today it is a global public company with roughly 4,000 employees and about $1.6 billion in annual revenue, and in fiscal 2026 it turned its first full-year GAAP operating profit.

UiPath sells a single governed platform that ties together the old and new worlds of automation: classic RPA robots that mimic clicks and keystrokes, API workflows, document understanding, process mining, and now AI agents that plan and act on their own. The pitch is that most companies have AI pilots stuck in a drawer, and UiPath is the layer that gets them into production safely.

The glue is orchestration. A control plane called Maestro coordinates agents, robots, and humans across long-running processes, with governance and audit trails baked in so a regulated bank or insurer can actually deploy it. That governance-first posture is the whole sales motion: automate aggressively, but keep a human in the loop and a paper trail behind every decision.

The problem UiPath keeps pointing at is that AI agents rarely make it past the demo. They are non-deterministic, hard to govern, and scary to unleash on core banking or claims systems. Most enterprises have siloed AI investments that never turn into real outcomes.

UiPath's answer is orchestration plus controlled agency: fine-grained role-based access, real-time monitoring, and compliance controls that de-risk letting an agent touch a mortgage file or a sanctions alert. It sells that as the difference between an AI experiment and an automated process a Fortune 500 will trust with money.

This is enterprise software in the fullest sense. UiPath's customers are large companies and governments in banking, insurance, healthcare, manufacturing, and the public sector, the kinds of buyers who care more about governance and auditability than about moving fast and breaking things. The company reports more than 10,000 customers across 100-plus countries.

Inside those accounts the buyers are a mix: IT and automation leads who run the platform, business owners who want processes fixed, and citizen developers building low-code automations. UiPath leans hard into that spread with low-code, pro-code, and prebuilt templates so the whole org can build, not just engineers.

The UiPath Platform is really a suite that grew around the original robots. The newer, splashier pieces are the agentic ones: Agent Builder for creating and deploying AI agents, Maestro for orchestrating them alongside robots and people, and Autopilot, a conversational assistant aimed at everyone from interns to CEOs.

Underneath sit the workhorses that made UiPath famous: document understanding (now IXP), process and task mining, and Test Cloud for agentic software testing. The platform runs as SaaS on Automation Cloud or on-premises for customers who need to self-host, which matters a lot to the regulated buyers it courts.

UiPath makes its money on recurring software subscriptions, and it is one of the few enterprise automation vendors that actually posts a starting price. A Basic plan runs from $25 a month for individuals and small teams, but the real business is the Standard and Enterprise tiers, both quoted as "Contact Sales," where cost scales with users, robots, AI usage, and support level.

That is the trade-off buyers describe: a genuinely mature platform, but licensing that gets expensive fast and rarely fits a small company. A separate Test Cloud track (Standard and Enterprise) prices agentic testing the same way. Annual recurring revenue reached about $1.85 billion, with net retention around 107%.

UiPath grew up as the leader of the RPA pack, a 7-time Leader in Gartner's Magic Quadrant for Robotic Process Automation, and its main rivals are the other automation heavyweights: Automation Anywhere, Blue Prism, and increasingly Microsoft Power Automate. Microsoft is the awkward one, since its automation is transparently priced from around $15 a user and often bundled into Microsoft 365 that customers already pay for.

UiPath's counter is depth and openness: a mature ecosystem, hybrid and on-prem deployment, and an orchestration layer it pitches as vendor-agnostic, coordinating third-party agents rather than locking you in. Whether that moat holds as agentic AI commoditizes is the open question the stock is arguing about.

UiPath's case studies read like an operations greatest-hits list. Johnson Controls automated 6,500 daily invoices and cites $18 million and 900,000 hours saved. Uber is on pace to save $22 million. EY went from 5 robots to 500 in 18 months, and dentsu says automation saves it more than 125,000 hours.

The named roster runs deep: Omega Healthcare, Canon, Xerox, Generali, TQL, Fiserv, State Street, and United Wholesale Mortgage among them. On its own numbers UiPath claims 93% of reviewers on Gartner Peer Insights would recommend it, which lines up with an unusually high 4.6 on G2 across thousands of reviews.

The recurring praise is consistent: reviewers call UiPath one of the most capable and mature RPA platforms, with strong scalability, a deep integration ecosystem, an active marketplace, and good documentation and community support. It rates around 4.6 on G2 across thousands of reviews, near the top of its category.

The recurring gripe is just as consistent, and it is money. Licensing is widely described as expensive, hard for small and mid-sized organizations to justify, and there is a real learning curve to get productive. The pattern is clear: people who need enterprise-grade automation are happy they bought it, and people who are small wish it were cheaper.

UiPath's venture years are behind it. It raised from backers including Accel and CapitalG on the way up, then IPO'd on the NYSE in April 2021 at a valuation near $29 billion, one of the largest software listings of that year.

The stock has since come back to earth, with a market cap in the mid-single-digit billions and shares well below their debut. The more interesting number now is the business: about $1.6 billion in fiscal 2026 revenue, a first-ever full-year GAAP operating profit, roughly $1.7 billion in cash, and a new $500 million buyback. UiPath is using that balance sheet to acquire its way into verticals, picking up Peak for retail AI and WorkFusion for financial-crime compliance.

UiPath is trying to convert a decade of RPA dominance into leadership of the messier, higher-stakes agentic era. The math is encouraging on the ground floor: it is now profitable, growing revenue in the low-to-mid teens, and sitting on enough cash to keep buying its way into verticals like financial services.

The risk is right there in the competition. If agentic AI gets cheap and good enough from Microsoft and the model providers, UiPath's edge narrows to governance and orchestration, and it has to prove those are worth an enterprise premium. Guidance points to roughly $1.75 billion in fiscal 2027 revenue, so the near-term story is steady execution while the bigger bet plays out.

UiPath runs on four stated values, Humble, Bold, Immersed, and Fast, and a "failing forward" mindset that tells employees to take smart risks and learn from the misses. The engineering job posts back it up with a builder culture: own it from concept to customer impact, ship early, treat every complaint as a design input, and measure honestly instead of chasing vanity metrics.

Work is flexible by team rather than by decree; the mix of hybrid, office, and remote varies with the role and business need. The company hires globally, with heavy engineering concentration in Bucharest and Bengaluru and a sales and go-to-market presence spread across dozens of countries, a reflection of its Romanian roots and its enterprise reach.

Disclosed US pay skews high, as you would expect for a public enterprise-software company. Engineering roles run roughly $140K to $288K in base, with product, design, and marketing roles landing in a similar six-figure range. Sales bands look lower on base because the real upside sits in commission on top of it.

All disclosed figures are US annual base salaries; UiPath is a public company, so equity is part of the package for many roles and commission drives sales comp, even where the postings do not spell out the exact split.

UiPath's privacy policy names two data controllers, a detail that matches its dual-home story. UiPath Inc. is incorporated in the US at 452 5th Avenue, 22nd Floor, New York, NY 10018, and UiPath SRL is incorporated in Romania in Bucharest. The company says it aims to uphold GDPR-level data protection even for its non-EU entities and honors California and other regional privacy rights.

On the trust side, UiPath positions governance as a product feature, not an afterthought: an AI Trust Layer, role-based access, real-time audit logs, customer-managed encryption keys and credential vaults on the Enterprise tier, and public-sector work that touches FedRAMP authorization. In early 2026 it also announced an AIUC-1 certification aimed at AI agent security and reliability.

The last year has been about proving the second act is real. UiPath hit full-year GAAP profitability for the first time in its history in fiscal 2026, its platform landed on TIME's Best Inventions of 2025 list for agentic automation and orchestration, and it kept acquiring to fill in verticals.

The headline deal was WorkFusion, an AI-agent specialist for financial-crime compliance, announced in February 2026 to push deeper into banking. That followed the Peak acquisition for retail and manufacturing AI, and a steady drumbeat of analyst wins across document mining, intelligent document processing, and AI-augmented testing.