The vendors aren't lying. That's what makes them dangerous.
The Solution Evangelist believes in the product. Genuinely. They have watched the model produce impressive output in a controlled demo, seen a customer tell a good success story, and absorbed the idea that the tool is transformational. They are not delivering a sales pitch so much as giving testimony.
That sincerity works. People leave the meeting thinking, "That person really believes in this," rather than, "That person is trying to sell us something." Both can be true. Belief does not prove that a product works at your scale, with your data, inside your constraints.
The Evangelist's job is to make the product feel understood before it has been tested. They answer objections with confidence because they have rehearsed the answers. They describe the happy path, where the model has clean context, the workflow is well prepared, and the result can be shown in a five-minute screen share. The difficult parts are real, but they tend to appear later, after the purchase, when the vendor's language has become your team's expectations.
Then comes the Vendor Bard.
The Bard arrives with the Inevitability Deck: the conference keynote, the partner briefing, the executive presentation. The Deck does not need to prove that the product is good. It argues that the category is transformational, the market is moving, and organizations that do not act now will fall behind.
The product becomes almost incidental. Urgency does the selling.
This is why the Deck can remain vague. Specific claims create obligations. A claim that a tool will reduce support costs by 30 percent can be checked. A claim that the technology will transform the customer experience can survive indefinitely. It sounds important without specifying what anyone should measure.
The Account Manager stays after the sale. Their stated job is to make sure the team is "getting value." In practice, that means making sure the team is using the product, then increasing usage, adding seats, expanding scope, or moving to the tier that includes the feature someone mentioned three months ago.
The Account Manager may be helpful. They may care about the project. They also have a quota. Those facts do not cancel each other out.
A check-in that begins as a troubleshooting conversation can become a discussion about expansion before anyone notices the turn. The proposal feels organic because it follows a real success: one team used the tool, one workflow improved, one manager wants to try it elsewhere. But success is not proof that the next purchase is worthwhile. It is a reason to evaluate the next purchase.
That distinction matters because vendors are usually not operating as cartoon villains. The Evangelist believes. The Bard creates urgency. The Account Manager expands the relationship. Each person may be acting in good faith while the combined system steadily moves you from evaluation to adoption to expansion.
You need a process that interrupts the motion.
Treat every expansion as a new decision. Ask what problem the additional seats or higher tier solves. Ask what evidence supports the change. Ask what the cost is if the team does nothing. Ask who will own the result after the contract changes. "We are already using it successfully" is not an answer to those questions.
Keep the original evaluation criteria visible. If the team bought the tool to improve code review, do not let a successful code review pilot quietly become a justification for deploying it across every engineering workflow. If the product now has a different role, evaluate that role on its own terms.
Also separate the product from the person presenting it. A convincing Evangelist is evidence that the Evangelist is convincing. A polished Deck is evidence that the vendor knows how to make a polished Deck. Neither tells you whether the tool is reliable, affordable, or appropriate for your environment.
The vendors are characters in a system built to create momentum. You do not have to resent them. You do have to notice what each interaction is trying to make feel inevitable.