The demo goes well. They always go well. Someone from the vendor drives, they pick the example, and the room watches a tool do the one thing it was built to do, perfectly, for eleven minutes. Somebody says the word transformation. Two people take photos of the screen.
Then Tuesday comes.
On Tuesday nobody from the vendor is driving. The example is whatever is actually in front of you, which is messier than the one in the deck. And the person who was supposed to use this has eleven other things open and a call in four minutes.
Six months later somebody says the pilot failed. Usually what failed was the rollout, and it failed for reasons that were visible on day one if anyone had been looking for them.
The demo is the least predictive hour you will spend
This is the part that gets missed, and it is worth sitting with, because it is not a criticism of demos. It is a description of what they are.
A demo is optimized. An expert is driving. The example was chosen because it works. There is no competing priority in the room, because the room was booked for this. Everyone present has decided, for forty-five minutes, that this is the most interesting thing happening.
Every one of those conditions is absent on Tuesday. So organizations make a buying decision on the single hour least representative of the thousand hours that follow, and then treat low usage as a surprise.
"The demo shows you the tool at its best. Adoption is decided by how it behaves at your team's worst: rushed, distracted, and halfway through something else."
The gap nobody budgets for
I have spent sixteen years in leadership development, most of it watching capable people get handed things that were going to change how they work. The ones that stuck had almost nothing in common technically. What they shared was that they asked for less than the person already had.
Failure one: friction is a tax you pay every single time
Everybody knows friction matters. What gets underestimated is that it does not cost you one decision. It costs you a decision every time.
A tool that takes six clicks to reach the useful part has not added six clicks. It has added a small negotiation that your user has to win against themselves, today, and again tomorrow, and again on the day they are behind. Most days they will win it. On the day they do not, the habit breaks, and habits do not usually come back on their own.
This is why "it only takes two minutes" is not the reassurance people think it is. Two minutes is not the problem. Two minutes that you have to decide to spend, repeatedly, while something louder is happening, is the problem.
Failure two: you explained the why to the company
Go back and read the rollout email. Almost always it says something like: this will improve our win rate, or shorten ramp time, or give leadership better visibility.
Every one of those is a reason for the organization. None of them is a reason for the person being asked to open the thing on a Tuesday.
People are not cynical about this. They are just busy, and a benefit that lands on someone else's scorecard does not survive contact with a full calendar. The rollouts that work answer a smaller and more specific question: what does this do for me, this week, that I actually want.
If the honest answer is "nothing yet, but in a quarter the dashboard will look better", that is worth knowing before you buy, not after.
Failure three: you bought it for the team you want
This is the one I see most, and it is the most human of the three, because it comes from optimism rather than carelessness.
Tools get chosen by people picturing the team at its best. The demo shows an advanced user doing advanced things, and it is genuinely impressive, and the person watching thinks: this is where we are going.
Then a new rep opens it on day one and the first thing it asks of them is something they cannot do yet. Their first experience of the tool is failing at it. And for most people the first experience is simply the experience. They will not go back to find out whether it gets easier.
A starting point set for who you want your team to be is not ambition. It is a design decision that filters out everybody who is not already good.
What meeting people where they are actually looks like
It is less inspiring than it sounds, which is rather the point. It means the tool does the adjusting.
In Parlare that shows up in two places, and both are deliberately unglamorous.
Your world, not a generic one. You set your industry and the market you work in, once, in your profile. From then on the AI on the other side of the conversation talks like a buyer in that market: your comparables, your regulator, the words your clients actually use. A broker in Toronto and a broker in Calgary are not practicing against the same generic prospect. Nobody configures anything per account. It is one field, set by the person, and it is optional, so a blank one just leaves the conversation geography-neutral.
Your level, measured rather than assumed. The difficulty of the simulated counterpart is not a setting an administrator picks. It is read from your own scored history on whichever framework that scenario leans on. A rep with no history gets someone open but cautious. It gets harder as your average climbs, and on the sales side the hardest setting needs a sustained average across at least five scored sessions, so one strong afternoon does not unlock it. You have to earn it.
That second one is the whole argument in miniature. The system could start everybody at the top and call it ambition. It starts them where they are and moves when they move.
What it does not do, which matters more than what it does
Two honest limits, because a piece about empty promises should not end with one.
It needs history before it can adapt to anything. On a brand new account both apps give you their easiest setting, and they say so. If someone demos this to you on a fresh login and calls it personalization, they are showing you the one state where it is not personalizing at all.
And the market grounding is the model's own knowledge of a place, not a data feed. It will talk like someone who works in your market. It is not quoting you verified comparables, and anyone who tells you their AI is doing that should be asked where the data comes from.
Neither of those is a caveat I enjoy writing. Both are the kind of thing you find out in month three anyway, and finding out in month three is how tools end up unused.
The commitment, stated plainly
We are not replacing anybody. That is not what this is and it is not what it is for.
The bet is narrower and, I think, more useful: people do not resist getting better. They resist being handed something that does not fit their day. So build the thing that fits the day, start it where the person actually is, and let it get harder only when they do.
That is not a transformation story. It is the reason anyone is still using it a year later.