How Proaction Saved 60% Sales Growth With OpenAI Codex

Overview

Proaction, a North American startup specializing in fleet management software, reports a 60% increase in sales growth since integrating OpenAI’s Codex and related models into its workflow. The company, which builds software for managing diverse vehicle fleets ranging from cars to construction machinery, found that traditional demo processes were creating a significant bottleneck in its sales cycle.

By leveraging AI tools, Proaction has reduced the engineering hours required for personalized customer demonstrations while simultaneously improving conversion rates. The initiative has saved the company an estimated 40 to 60 engineering hours per month and allowed co-founder and COO Colin Knudsen to reclaim approximately 33 hours of his own time monthly. This shift has enabled the startup to scale its operations without proportionally increasing its technical headcount.

Accelerating Sales Cycles with Custom Demos

For a SaaS company like Proaction, demonstrating how its platform fits a prospect’s unique operational needs is critical. However, creating these personalized demos previously required significant engineering resources that the small team did not have. Founders were forced to rely on static slide decks and verbal explanations, which often failed to convey the full potential of the product.

The introduction of Codex changed this dynamic entirely. Knudsen now creates four to six customized, interactive demos each month, with each demo taking between 30 to 45 minutes to build. These demos are tailored to the specific vehicles and workflows of the prospect, allowing potential clients to see their own data integrated into the Proaction interface.

“As a non-technical person, I used to have to loop engineers in if I wanted a demo. Now I do it myself in Codex.” — Colin Knudsen, Co-founder and COO, Proaction

The process begins after a sales call, where Knudsen directs Codex to review Granola recordings, prospect email threads, and shared spreadsheets. Codex then generates an HTML demo environment that mirrors both the Proaction product and the customer’s specific fleet. During live meetings, prospects can point to elements on their screen and request adjustments in real time. This collaborative approach removes the need for immediate engineering intervention and makes the sales process more engaging and concrete for the buyer.

Operational Efficiency and Engineering Conservation

The impact of Codex extends beyond the sales floor, significantly reducing the burden on Proaction’s engineering team. When prospects become customers, the customized demo serves as a visual reference for the development team. This reduces ambiguity and minimizes the back-and-forth typically required to clarify project specifications.

If Proaction’s engineers had produced comparable demos manually, Knudsen estimates it would have taken approximately 10 hours per demo. With four to six demos generated monthly, this represents a substantial saving of 40 to 60 engineering hours. This efficiency allows the technical team to focus on core product development rather than bespoke sales support.

Additionally, Proaction has built a customer solution center using Codex. This platform allows prospects to log in, explore tailored workflows, and review sales materials independently. It empowers non-engineering team members to translate customer conversations into clearer requirements, ensuring that when engineers are finally involved, the scope of work is well-defined.

Consolidating Daily Workflows

Knudsen’s daily responsibilities span sales, customer support, and product management. Previously, this required constant context switching between multiple applications. By utilizing Codex plugins for tools such as Granola, Gmail, Slack, Linear, GitHub, and HubSpot, he can now execute tasks within a single interface.

He pulls call transcripts and email histories to prepare follow-ups, creates task issues in Linear, and updates opportunities in HubSpot without leaving Codex. Furthermore, he has implemented scheduled automations that review recent calls and prepare sales updates for the broader team.

“Everything that I do is centered around working in Codex. I don’t leave it much.” — Colin Knudsen, Co-founder and COO, Proaction

This consolidation has saved Knudsen an estimated 25 to 33 hours per month. Handling 15 to 20 distinct tasks daily with greater efficiency, he notes that he cannot imagine operating as a startup founder without Codex at the center of his workflow.

Enhancing Product Capabilities with Advanced AI Models

Beyond sales enablement, Proaction is integrating OpenAI models directly into its platform to offer more advanced features to its customers. The company utilizes ChatGPT-5.6 Sol to analyze photos of vehicle issues, helping to identify damage and streamline insurance or maintenance claims.

More notably, Proaction is developing a "Managed Execution Layer" powered by GPT-Live-1 and GPT-6 Astra. This layer enables voice agents to handle day-to-day fleet operations, moving beyond mere management and tracking to active execution of tasks.

One such agent, named Marty, is designed to coordinate vehicle maintenance. Marty interacts with drivers to understand problems, calls repair shops to arrange service, and assists in getting estimates approved and paid. Human team members intervene only when work requires complex review or specialized handling.

“We’re building the ability for Proaction to execute work for our customers, beyond helping them manage and track it. The advances in OpenAI voice are a big reason we can do it.” — Colin Knudsen, Co-founder and COO, Proaction

Danny O’Halloran, Head of Product at Proaction, highlights that GPT-6 Astra allows the team to build these experiences faster. He notes that Astra’s computer-use runs are more succinct compared to previous models, reducing the time required to execute similar tasks. By combining Codex for internal efficiency with advanced models for external product features, Proaction is driving significant growth while maintaining a lean operational structure.