📰 Key Takeaways

Merge has brought several OpenAI tools into its day-to-day operations, covering three products: GPT-6 Astra, ChatGPT, and Codex. On the product side, the team uses these models to review product code, with AI helping inspect the code and keep quality in check. On the marketing side, the team also relies on AI to support related work. Beyond product and marketing, other teams inside Merge also use AI for more administrative and sales-oriented tasks, like planning executive dinners and preparing sales progress updates. Overall, this is a case of a company stretching AI from engineering into marketing, sales, and everyday admin workflows. The original summary doesn’t give any concrete benefit numbers, rollout results, or technical implementation details, so check the original article for more.

💬 JudyAI Lab Take

Merge has put GPT-6 Astra, ChatGPT, and Codex into its daily operations, from reviewing product code and supporting marketing all the way to planning executive dinners and preparing sales progress updates. What’s worth noticing isn’t any single tool, it’s that AI is being used across the whole workflow, not just in engineering.

The original summary doesn’t give any numbers or implementation details, so we can’t judge how well it’s working. But the case does reveal a design mindset: when companies bring in AI, they often start with an engineering scenario, like using AI to help keep code quality in check, and then it spreads to marketing, sales, and admin, where the work is repetitive and the format is fairly fixed. For AI builders, this means your product’s use cases won’t stay confined to what you originally designed for. When the same set of models gets used for very different tasks across departments, the tool needs to be general enough, and it also needs to be something non-technical teammates can pick up right away. Another thing worth noting: Merge uses several products at once, each matched to a different need, rather than relying on one single entry point to do everything.

If you’re designing an AI tool, try listing the 3 most tedious, repetitive tasks on your team, like scheduling or compiling progress updates, and pick one to hand off to AI first.

📅 Source Info

🔗 Further Reading