Driving scalable growth with OpenAI o3, GPT-4.1, and CUA
OpenAI
Unify's sales platform now runs on OpenAI's o3, GPT-4.1, and its computer-using agent to automate prospecting and outreach. That means reps spend less time researching leads and more time actually talking to them.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Sales teams have always drowned in busywork before they ever get to a real conversation: scraping data, writing personalized emails, tracking who opened what. Unify, a go-to-market platform, is betting that OpenAI's latest models can drain most of that swamp automatically.
The company has plugged o3, GPT-4.1, and OpenAI's Computer-Using Agent (CUA) into its workflow to handle prospecting and research end to end. Instead of a rep manually digging through LinkedIn profiles, company filings, and news mentions, the system does that legwork on its own, then drafts outreach that actually reflects what it found rather than a generic template with a first name swapped in.
What makes this setup notable isn't just the automation, it's the always-on part. CUA lets the system interact with web interfaces the way a person would, clicking through pages and pulling context, which means the research doesn't stop the moment a human logs off. Pair that with GPT-4.1's speed for drafting and o3's heavier reasoning for prioritizing which accounts are worth chasing, and you get a pipeline machine that runs continuously in the background.
The pitch to sales teams is straightforward: let the software do the grunt work of finding and qualifying prospects, and let humans spend their limited hours on the calls and negotiations that actually close deals. That's the theory behind most sales-tech automation, of course. The difference here is that Unify is leaning on general-purpose frontier models rather than a narrow, purpose-built tool, betting that reasoning models get good enough at judgment calls, like which lead is actually worth a rep's time, that they start doing real qualification work instead of just formatting emails.
My take — AI-written commentary, not fact-checked reporting
This is the least flashy, most plausible use of frontier models I've seen lately: not writing a novel, just quietly doing the tedious research a junior SDR would otherwise burn hours on. GTM automation has promised this for a decade with mixed results, but stapling it to o3-level reasoning instead of brittle rule-based scripts is the first version that might actually hold up. My skepticism is reserved for how these companies talk about 'high-impact' human time, that's usually code for smaller sales teams, not happier ones.
Read more about this at: OpenAI