Merge’s integration platform uses OpenAI models — GPT‑6 Astra, ChatGPT and Codex — across product reviews, marketing, and business‑development planning. The company aims to help other businesses build connected software and AI products, from syncing customer data to enabling AI agents to act across business systems.
Product set and capabilities
Merge’s offerings include:
- Merge Unified for customer‑data integrations.
- Merge Agent Handler to give AI agents secure access to tools.
- Merge Gateway for model routing and cost controls.
- Merge for Workforce, which brings these capabilities inside companies so teams can manage which models, tools and business systems employees may access.
Together these products support both the software Merge’s customers build and the AI‑augmented work inside Merge itself.
How Merge uses models internally
Shensi Ding, Merge’s co‑founder and CEO, uses Codex to review product code, including work on a new setup interface for configuring tool access in Merge for Workforce. Other team members use GPT‑6 Astra for product testing and marketing. The company also relies on AI for routine company work, from preparing sales updates to planning customer dinners.
Ding emphasizes reliability: Merge reports that tool calls through its connectors succeed 99.9% of the time, compared with roughly 60% success on public MCP servers.
Code reviews and choice of tools
For product development, Merge relies on GPT‑6 Astra because of its speed and improved code‑review performance. Ding says the team appreciated GPT‑6 Astra’s “personality,” and that Codex has also improved significantly. Merge allows employees to choose their preferred AI tools, which has become a lively internal debate about which model to use.
Codex provided code‑review feedback on a proposed implementation of the new provisioning tab in Merge for Workforce; the team discussed UX changes such as making the “Add tool pack” action a clearer, separately opening link. Engineers implemented and validated the final decisions.
AI‑assisted executive dinners
Merge also used AI agents to help organize executive dinners for customers and prospects. For one Dallas event the bot selected the city and target guest types, pulled potential attendees from the CRM and Apollo, checked past invitations and responses, and helped prepare outreach in Ding’s email. It maintained the attendee tracker, proposed a seating chart, and created calendar notes for sales and customer success to add context.
The model handles reasoning and drafting while Merge’s connectors provide access to business systems. Ding reviews and edits the output and curates the guest list using relationship knowledge the bot lacks.
Measured efficiency gains
In October 2026 Merge reported that preparing guest lists and outreach and maintaining the attendee tracker dropped from ‘‘dozens of hours’’ across the company to about 30 minutes per event, comparing the same planning tasks measured over the preceding three months. Hosting, relationship building and onsite logistics remain human responsibilities.
Merge ran 16 executive dinners in Q3, up from eight in Q2, and plans to host more than 30 in Q4. After the Dallas dinner the team reported positive feedback on invitations and new follow‑up demo opportunities.
Ding said: “With GPT‑6 Astra, the dinner bot connects to our Salesforce, Gong, Google Sheets, and Gmail, and it’ll send emails on my behalf to people in our network. It gives me recommendations for people who are a good fit for each dinner, and it surfaces new potential attendees.”
Sales, marketing and reusable workflows
Charlie Lynch on Merge’s sales leadership team built reusable AI workflows for deal approvals, forecast updates, and next steps after renewal calls. Rather than repeatedly gathering context, he uses the workflows to prepare material, then reviews recommendations and decides on actions.
The social media team uses GPT‑6 Astra via ChatGPT to identify posts the company should respond to, and marketing used it to clarify messaging for the Merge Gateway Evaluations launch.
Leadership, experimentation and culture
Ding advises leaders to set the ceiling for AI fluency in their organizations. She recommends finding lower‑risk areas to experiment and giving people time and space to tinker. Merge lets team members pick which tools they want and encourages daily learning and testing, including room to ‘‘go wild and make mistakes’’ when necessary.
Ding sums up the approach: “Anything that I want to do, I can just tell AI to do it, and it can just do it.” Merge continues to use OpenAI technology both to expand what its teams can build and to bring the same experimental spirit into its products and operations.
