Proaction builds software for companies that manage vehicle fleets — from cars and trucks to construction equipment. Because each fleet operates differently, demonstrating how Proaction’s platform fits a prospect’s specific operation is a key part of selling the product. Those personalized demos previously required engineering time the team could not spare, leaving founders to rely on conversations and slide decks.
Codex changed that. Colin Knudsen, co‑founder and COO of Proaction, says, “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 now builds four to six customized, interactive demos per month, each taking 30–45 minutes to create. He estimates this avoids 40–60 hours of engineering work per month.
How the process works
After a sales call, Colin points Codex to the Granola recording, prospect email threads, and any spreadsheets the prospect shared. Codex uses that context to generate a customized HTML demo environment that mirrors Proaction’s product and the customer’s own fleet. When Colin shares his screen, prospects see their own vehicles and equipment arranged around how they work, can indicate needed adjustments, and collaborate on the solution — all without involving engineers.
Colin estimates that the share of deals moving from initial contact into solution development, rather than being put into nurture, has increased by roughly 50–60% thanks to the custom demos. If Proaction’s engineers produced comparable demos, each would take about 10 hours; with four to six demos monthly, that equates to the 40–60 hours of engineering time Colin cites.
The custom demo also serves as a visual reference when a prospect becomes a customer, reducing questions and back‑and‑forth about what to build.
Customer solution center and streamlined workflows
Using Codex, Proaction created a customer solution center where prospects can log in, explore workflows tailored to their business, and review sales materials. This lets customers articulate needs more precisely and gives non‑engineering teammates a way to turn conversations into clearer requirements, so engineers arrive with a more concrete picture of what to build.
Colin’s day covers sales, customer support, and product management. Through Codex plugins for tools including Granola, Gmail, Slack, Linear, GitHub, and HubSpot, he pulls together customer context and acts on it directly in Codex: extracting call transcripts and email history to prepare follow‑ups, creating Linear issues, and updating HubSpot opportunities. He also set up a scheduled automation that reviews recent calls and prepares sales updates for the team.
Previously, his workload forced frequent tab switching and manual copying between tools. Now he can describe what he needs and have Codex gather context and execute the next step. Colin estimates he handles 15–20 distinct tasks per day and that Codex saves him 25–33 hours per month.
OpenAI models across the platform: image analysis and agents
Proaction uses multiple OpenAI models across its product. When customers submit photos with vehicle issue reports, ChatGPT‑5.6 Sol helps identify damage. Using GPT‑Live‑1, Proaction is building agents to handle routine fleet operations; the company calls this its Managed Execution Layer.
Customers can ask specialized agents to handle tasks like tolls or service, or set up workflows that automatically assign the right agent. The agents use OpenAI models, including GPT‑Live‑1 and GPT‑6 Astra, to make voice calls, review documents and images, analyze text, and respond in chat.
One agent, Marty, is designed to coordinate vehicle maintenance: Marty talks with a driver about a problem, calls repair shops, arranges service, and helps get estimates approved and paid. Proaction’s team intervenes when human review or action is required.
GPT‑6 Astra also enables faster development of these customer experiences. Danny O’Halloran, Head of Product, said, “Astra’s computer‑use runs are more succinct. With GPT‑5.6 Sol, I had a much longer run to execute the same work.”
Overall, Codex and GPT‑6 Astra free up Proaction to focus on sales and customer support, while GPT‑Live‑1‑based agents reduce daily burden for fleet managers.
Key figures and effects
- Codex‑based demos: 4–6 demos per month, 30–45 minutes per demo.
- Estimated engineering time saved: 40–60 hours per month (engineer‑built demo ~10 hours each).
- Increase in deals moving to solution development: roughly 50–60%.
- Colin’s daily load: about 15–20 tasks; Codex saves him 25–33 hours per month.
Proaction’s example shows how generative models and automated agents can boost sales efficiency and clarify product requirements without continuously taxing engineering teams.
