Proaction’s Sales Growth and Efficiency Gains
Proaction, a startup specializing in fleet management software, reports a significant 60% increase in sales and a reduction in internal workload after integrating Codex alongside OpenAI models such as GPT‑Live‑1 and GPT‑6 Astra.
Streamlined Personalized Demos
Due to the unique operations of each fleet, Proaction’s sales process previously depended heavily on engineering resources to create tailored product demonstrations. Colin Knudsen, co-founder and COO, notes that Codex has transformed this process. As a non-technical user, he is now able to independently build four to six customized, interactive demos each month, with each demo taking approximately 30 to 45 minutes to assemble. This shift has allowed Proaction to save an estimated 40 to 60 engineering hours monthly, as producing similar demos through engineering would have required about 10 hours each. Furthermore, the introduction of these personalized demos has led to a 50% to 60% increase in the percentage of deals progressing from initial contact to solution development, rather than remaining in the nurturing phase.
Efficiency Across Sales and Support
Colin utilizes Codex plugins for various tools, including Granola, Gmail, Slack, Linear, GitHub, and HubSpot, to consolidate call transcripts, email histories, and spreadsheets into a single environment. He estimates that Codex saves him between 25 and 33 hours each month across his responsibilities in sales, customer support, and product management. The platform also facilitates scheduled automations that prepare sales updates for the team. When combined with the engineering time saved, Proaction claims a total of over 75 hours saved per month.
Advanced Agent Capabilities with OpenAI Models
Proaction employs ChatGPT‑5.6 Sol to assist in identifying vehicle damage from submitted photos. Additionally, the company is developing agents using GPT‑Live‑1 and GPT‑6 Astra that can perform tasks such as making voice calls, reviewing documents and images, and engaging in chat interactions. One notable agent, named Marty, is responsible for coordinating vehicle maintenance by communicating with drivers, contacting repair shops, arranging services, and assisting with estimates and payments. The Proaction team intervenes only when human oversight is necessary. Danny O’Halloran, Head of Product, mentions that GPT‑6 Astra produces more succinct computer-use runs compared to GPT‑5.6 Sol, resulting in shorter execution times for similar tasks. Overall, Proaction leverages Codex, GPT‑Live‑1, GPT‑6 Astra, and ChatGPT‑5.6 Sol to enhance demo creation, minimize engineering back-and-forth, and support new agent workflows for customers.
Original source: OpenAI News