Your company bought the AI tools. The training happened. A month later, half the team is doing the work the way they always did. Nobody stood up and said no. They just routed around it.
This is the quiet version of resistance, and it is easy to miss. Usage dashboards look fine. Adoption looks like a line going up. But the real work, the judgment calls and the answers people actually rely on, is still happening in the old channels.
What routing around looks like
It rarely looks like a flat no. It looks like a person opening the tool, glancing at what it produced, and then doing the task by hand anyway. It looks like a Slack message that starts with "I know we have the AI thing, but can you just tell me." It looks like a summary that gets generated and never read.
People are not being difficult. They are protecting their own results. When someone is responsible for an outcome, they will not hand that outcome to a tool they do not yet trust. So they keep a human path open, just in case. That parallel path is the resistance.
Why it happens
The reasons are ordinary. The tool gives an answer, but the person cannot see where it came from, so they have to check it themselves. Checking takes as long as doing the work, so the tool has added a step instead of removing one. Or the answer sounds confident and turns out to be wrong once, and that one time is enough. Trust is slow to earn and fast to lose.
There is a broader pattern worth naming here. In BCG's AI at Work 2026 survey of 1,488 US workers, adding more AI tools correlated with lower reported productivity. That is a correlation, not proof that the tools caused the drop. But it fits what people describe on the ground: more tools can mean more surfaces to check, more places an answer might live, and more doubt about which one to believe. Piling on software is not the same as making the work easier.
The trust gap
The core problem is not that people dislike AI. It is that most rollouts ask for trust before they have earned it. A tool that answers without showing its work is asking the reader to take it on faith. Busy people do not take answers on faith when their name is on the result.
You can see the gap clearly in one test. Ask someone to act on an AI answer without checking it. If they hesitate, the tool has not closed the gap. It has just moved the checking to a different moment. And every check is time, which is the thing the tool was supposed to give back.
What actually gets used
The tools people keep are the ones that make the next step faster and safer at the same time. That usually means three things.
- It shows its source. The person can see where the answer came from and trust it without re-doing the work.
- It stays inside what is known. When the answer is not there, it says so, rather than filling the gap with something plausible.
- It fits the work people already do. No new place to file things, no separate habit to maintain, no extra login to remember.
Adoption is not a training problem or a change-management problem. It is a trust problem. People route around tools they cannot verify, and they lean on tools that let them verify fast. If you want the parallel human path to close on its own, make the tool worth trusting.
Start from the work, not the tool
One reason rollouts stall is that they lead with the software instead of the job. The question that earns adoption is not "how do we get people to use this," it is "what does this remove." If the honest answer is "a step," people will keep it. If the answer is "nothing, but you should trust it," they will quietly go around it, and you will not find out until the numbers stop matching the reality.
This is the thinking behind StandIn. Instead of a new tool to adopt, it works from the work you already did. You write one short brief at the end of your day, and while you are off, it answers your teammates from what you actually wrote, with a source under every answer. When the answer is not in the record, it says so instead of guessing. There is nothing to route around, because there is nothing to take on faith.
When you're off, your StandIn is on.
It answers your teammates' questions from work you've already done, in your words, with a source under every answer.