Roadrunner

Roadrunner builds AI-native software for enterprise pricing, quoting, and approvals. Its Prompt, Quote, Approve (PQA) workflow lets sales representatives request quotes in natural language and uses an agentic stack to recommend deal configurations, enforce pricing policies, and route approvals. The company announced general availability on May 12, 2026, positioning the platform as a replacement for legacy configure, price, quote (CPQ) software.[4][5][2]

On the same day, Roadrunner announced $27 million in aggregate funding across a seed round led by Mamoon Hamid at Kleiner Perkins and a Series A led by Trae Stephens at Founders Fund. Stephens was joining its board. The announced use of funds included expanding the team, deepening enterprise partnerships, and building toward broader quote-to-cash coverage, including billing; that broader coverage was a development goal.[4][5]

The launch followed Roadrunner’s founding in 2025 and incubation at Kleiner Perkins. Its earlier development explains the decision to rebuild the underlying data model rather than add another layer to existing CPQ software.[4][1]

Roadrunner was co-founded by Joubin Mirzadegan, Ajay Natarajan, and Eugene Shao. It was Kleiner Perkins’s first incubation since Glean. Mirzadegan remained a Kleiner Perkins partner while serving as Roadrunner’s CEO. The company’s team page, as of September 18, 2026, listed Natarajan as chief product officer and Shao as chief technology officer; the two had previously co-founded the AI college-application platform Athena.[4][5][3]

Mirzadegan traced the idea to his sales career, recalling custom additions to Salesforce CPQ that still failed to support the deals he needed to close. He also described conversations with enterprise CIOs and CROs at Kleiner Perkins about shortcomings in sales software. His stated decision was explicit: “We didn't set out to improve CPQ. We set out to replace it.”[4]

Hamid told Fortune that Mirzadegan brought a prototype he had vibe-coded over a weekend to their one-on-one. According to Hamid, prospective buyers had said they would buy the product if it were built, and the team closed its first seven-figure deal within three months of building the prototype.[5]

By December 2025, Roadrunner’s development thesis combined a rebuilt data model with enterprise design-partner co-development. A Latent Space introduction described the pricing problem as a shift from seats toward consumption, bundles, renewals, and growing SKU catalogs. Using LLMs to recommend deal structures was then a prospective goal; the May 2026 launch announcement described deal recommendations as part of the platform.[1][4]

As of September 18, 2026, Roadrunner’s product offering centered on a single conversational agent for quoting, approvals, and deal-structure questions. Deal desk, finance, legal, and executives could approve in parallel on a shared surface. The company described scoring each quote against the customer’s own closed-won history using chosen metrics—such as margin, discount, and term shape—with customer-defined confidence intervals.[2]

Administrator changes also used a conversational workflow. An admin could request a new discount tier, approval threshold, or product rule; the admin agent would draft the configuration change and generate user acceptance testing (UAT) scenarios against real quote history. The administrator had to sign off before the change shipped.[2]

The product description paired a modular pricing data model with composable quotes for expansions, cancellations, upgrades, downgrades, and renewals. Transactions rolled into a time-bound, account-level entitlement view showing what the customer owned and why. Roadrunner also supplied engineers to model customer catalogs, rebuild configuration rules, and migrate from existing CPQ systems, making implementation part of its enterprise offering.[2]