The website is the company you wish existed.

Every company website describes a slightly better company than the one that exists. That is not dishonesty. It is the job of a website.

The site was written in a quiet month by the people who most believe in the business. It leads with the service the founder wants to sell more of, the industry she hopes to win, the outcome from the best customer's best quarter. It is a statement of direction, and a good one should be a little ahead of reality, because it is pulling the company toward something.

The website is the company you wish existed. The calls are the company your customers bought.

Which makes it a strange thing to train a salesperson on, human or otherwise. Hand a new hire nothing but the homepage and they will pitch the deck: the flagship service, the aspirational industry, the feature the founder loves. Then they get on a real call and discover that customers keep buying the dependable thing, for a reason the site never mentions, after an objection nobody wrote down.

The real company lives in the conversations. Which questions buyers ask in the second meeting. Where reps get stuck. Which part of the pitch makes people lean in and which part makes them check the time. None of it is on the website, because none of it was known when the website was written.

A new rep figures this out in about a month, by sitting in on calls and getting a few wrong. Anything else that sells on the company's behalf has to learn it the same way, from the same place, or it will keep pitching the deck with total confidence long after every person on the team has stopped.

So the useful design starts from the website and does not stop there. On the first day there is nothing else to go on, and the site is a fair first draft of what the business says about itself. After that, where the business switches it on, the platform learns how the team actually sells from the calls it records. The coaching scores on those calls are what it learns from. What it learns arrives as a proposal - a revised playbook for how the conversation should go - and a person approves it before it changes anything. Where a business has not switched that learning on, the agents keep working from what the business told them.

Start from what you say. Learn from what they hear. Ask before confusing the two.

The approval is not a formality. The website is the company's intent, and intent should outrank the habits of any one quarter. Sometimes the calls reveal a better pitch than the one on the homepage, and the homepage should change. Sometimes they reveal a bad habit that has quietly spread across the team, and the habit should change. From inside the transcripts the two look identical. Only someone who knows where the business is trying to go can tell them apart.

The calls tell you what is working. Only you can say whether it is where you meant to go.

It is worth being exact about what this knows. It learns only from the calls it scored, and from nothing else. It does not know why a deal was lost unless someone said so out loud, and it does not pretend to. A score on a call is a reading of that conversation against the rubric the team chose, not a verdict on the market.

Keep writing the website. It is the company you are building toward, and it deserves to be written with ambition. Just do not ask it to explain what customers bought.

The website says where the company is going. The calls say where it is. Anything that sells on your behalf should read both, and should ask before it mistakes one for the other.