Codex Turned Sales Calls Into Demos: Proaction’s 60% Sales Lift and 75 Hours Saved

Sales Calls Became Interactive Fleet Demos, Not Slide Decks

Proaction, a North American startup that builds fleet management software, has recorded a 60% increase in sales after moving custom demo creation into Codex.

OpenAI published the customer story on September 25, 2026, detailing how Proaction also saved 40 to 60 engineering hours and 33 founder hours each month.

• 40–60 engineering hours saved per month by replacing engineer-built demos with Codex-created demos.
• 33 founder hours saved per month from consolidated sales, support, and product workflows.
• 60% increase in sales from moving initial contacts into solution development with custom demos.

Before the shift, personalized demos required engineering time the team could not spare.

As OpenAI‘s customer story details, the founders often had to rely on conversations and slide decks because highly tailored product walkthroughs were impossible to produce on demand.

Codex flipped that constraint.

Colin Knudsen, co-founder and COO of Proaction, now builds four to six interactive demos monthly, with each taking 30 to 45 minutes.

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.

That capability gave Proaction a concrete reason to move prospects from initial contact into solution development rather than long-term nurture.

The company estimates the share of deals advancing to that stage has risen by 50% to 60%.

Inside Proaction’s Codex Stack: Founder Hours to Fleet Agents

After a sales call, Knudsen points Codex to the Granola recording, prospect email threads, and any spreadsheets the prospect has shared.

Codex uses that context to customize an HTML demo environment that mirrors Proaction’s product and the customer’s own fleet.

When the screen is shared, prospects see their own cars, trucks, or equipment organized around how they work.

They can point to what needs adjusting and help shape the solution without engineering involvement.

If engineers built comparable demos, Knudsen estimates each would take about 10 hours.

Across four to six demos a month, that represents 40 to 60 hours of engineering effort returned to product work.

The same demo becomes a visual reference for implementation.

Engineers receive a concrete picture of what to build, which reduces questions and back-and-forth when a prospect converts.

Knudsen also built a customer solution center where prospects can log in, explore tailored workflows, and review sales materials.

That shifts requirements discovery away from engineering and toward non-technical teammates.

Codex plugins for Granola, Gmail, Slack, Linear, GitHub, and HubSpot pull together customer context in one workspace.

Knudsen uses the environment to prepare follow-ups, create Linear issues, and update HubSpot opportunities.

A scheduled automation reviews recent calls and prepares sales updates, compressing a workload that previously meant jumping between tabs.

Across 15 to 20 daily tasks, Knudsen estimates Codex saves him 25 to 33 hours a month.

On the product side, Proaction is moving from tracking fleet work to executing it.

When customers submit photos with vehicle issue reports, ChatGPT-5.6 Sol helps identify damage.

With GPT-Live-1, Proaction is building what it calls a Managed Execution Layer.

Customers can ask specialized agents to handle tolls, service, or automated workflows.

The agents use GPT-Live-1 and GPT-6 Astra to make voice calls, review documents and images, analyze text, and respond in chat.

One agent, Marty, coordinates vehicle maintenance by talking with a driver, calling repair shops, arranging service, and helping get the estimate approved and paid.

Human teams step in when the work needs review or intervention.

Astra’s computer-use runs are more succinct. With GPT-5.6 Sol, I had a much longer run to execute the same work.

Danny O’Halloran, Head of Product at Proaction, framed that speed increase as a reason the team can build these agent experiences more quickly.

Fleet AI’s Governance Gap Is Growing Alongside Its Wins

Proaction’s story lands inside a broader commercial fleet AI market that is moving at two speeds.

Heavy Duty Trucking reported in early September that fragmented AI adoption is creating serious security, governance, and audit-trail risks.

Trimble authors cited a Gartner prediction that more than 40% of agentic AI projects will be canceled by the end of 2027 because governance and accountability complexity outpaces deployment capacity.

A survey of 68 freight professionals found roughly 55% actively using, piloting, or experimenting with AI in limited areas, but only about 15% using AI agents.

Proaction sits in that smaller group, positioning its Managed Execution Layer rather than a narrow pilot.

Commercial Carrier Journal offers a concrete operational benchmark: PS Logistics started with an AI voice agent for drivers across roughly 17 operating companies and reached a 70% to 75% call-resolution rate.

PS Logistics and Axle Logistics both frame their AI work as a way to return time to employees for customer and driver relationships, not as a replacement for human contact.

Randall Reilly’s Fleet AI Report found that 30% of fleets are not using AI anywhere, and nearly 60% say their biggest obstacle is their own team.

Recruiting produced the best ROI of any AI category tested so far, which complicates the idea that fleet-wide agent adoption is automatic.

The report’s recommended fix is a single secure AI agent with layered defenses, human-in-the-loop controls, and permission constraints.

Proaction’s human review step aligns with that model, but the company’s self-reported sales lift and hours saved have not been independently audited.

Still, the story is not just about results.

It shows that the fastest measurable business value may come from using Codex to accelerate sales motion before full agentic rollout.

Sales Demos Are the Wedge, Fleet Agents Are the Frontier

Proaction’s experience suggests that Codex can collapse the gap between a sales conversation and a working product proof before engineering time is spent.

For teams building AI-assisted content and automation pipelines that need to scale, programmatic SEO and AI automation is how Andres SEO Expert approaches it — talk to us here.