Job titles tell you who picked up the phone. They rarely tell you who can say yes. Across five long calls in our latest batch, conversations ran 13 to 20 minutes each, 82 minutes of talk time in total, and only one ended with a booked meeting.

One bank’s VP of GRC technology strategy sounded like the obvious buyer. Early in the call, it became clear they were an individual contributor working under a vendor policy they did not write. They declined on the spot. The call was friendly, and the outcome was zero.

A second bank’s head of AI engagement took the opposite path. She said she was open to a pitch the following week and asked for an email so she could coordinate with the people who would need to be involved. Similar seniority, different answer. The difference was whether she knew how vendor decisions got made inside her company.

We now spend the first minute on decision authority, not the pitch. Ask what their role is in evaluating a tool like this, who approves a new AI vendor, and what that approval takes, whether that’s a security review, procurement, or a pilot budget. A clear answer tells you whether the person in front of you can move the deal. A vague one tells you to find the person who can.

The next objection is usually build over buy. Three of the five calls ran into an internal build. One head of AI at a 1,000-person technical org had already built tools for AI proxying, key management, and usage monitoring, and cited a build-over-buy mandate to reject a meeting. Another VP of Engineering had a bespoke internal gateway for token management and governance and declined because of a release cycle, inviting a January follow-up. A third prospect said others on the team were building an in-house gateway and wanted to understand fit with their architecture before committing to any meeting, while noting that a major data platform offered something similar.

The one meeting we booked was the shortest call in the set, at 772 seconds. The two longest, at 20 and 18 minutes, ended without a meeting. Length didn’t predict the outcome. Who was on the call did.

Our rule for the next batch: no pitch until we know who decides, how the vendor gets approved, and whether an internal build already exists. Each of those can be asked in under a minute.