A broker asked me to build him an AI CRM. The fix had no AI in it at all

A broker contacted me because he wanted a CRM system that used artificial intelligence. He wanted the package, including predictive lead scoring. The broker thought that his agents were missing some leads and that an intelligent system would catch those leads. He had already chosen a tool that cost around 600 dollars per month. He wanted me to set it up and create the automations for it. Before I agreed to do anything I asked the broker a question. I asked him if his team actually used the CRM system they already had and it turned out that they did have a CRM system and nobody used it. The system was empty because the agents did not log their calls after each showing. This was a task that they all skipped. This is the problem. You cannot use scoring with a system that has no data. The model has nothing to predict from. The broker would have been paying 600 dollars per month for nothing. He would have thought that the leads were being handled. The real problem would still be there. The problem was that nobody was writing anything down. What I created instead was very simple. The calls and texts that the agents were already making were logged automatically into the CRM system. This was done without anyone having to lift a finger. The agents received a message every morning with the names of the people they should call that day and why. That’s all…. There was no scoring model and no AI making decisions. The manual step was that the right names were shown at the right time. Something that stuck with me happened a weeks later. One of the brokers agents a man who had been in the business for twenty years and did not like technology told the broker that the morning list was the software thing, in years that did not feel like more work. He was not impressed by anything. He just liked that it did not ask him to do anything. This is the standard that we should aim for and every new tool misses it. The broker told me later that he had been about to buy the intelligence CRM system and would have blamed his agents when it did not work. This is the trap. You buy something that looks impressive. It does not change anything because the real problem was never the software. Then you think you need something more impressive. The CRM system is a tool and the brokers agents were the ones who were actually using it. The broker learned that he should focus on the CRM system they already had and make sure that the agents were using it correctly before buying an one. I've built 40 something automations for clients across a bunch of industries, and one of the ones I'm proudest of is a job where I argued myself out of most of the scope on the first call… The client appreciated my honesty and tbh I am happy to be the person who tells you that you might not need the thing.