Will AI fix our sales problems?

Not on its own. AI applied to a broken process just makes the mess faster. Fix the process around sales, service and marketing first, then automate the parts that are working.

Where do scaling bottlenecks usually sit?

Rarely in the product. Far more often in the handoffs between teams, the steps no one owns, and the parts of the funnel no one can see clearly.

What should we sort out before adding AI?

Visibility and ownership. You need to see what is actually happening in the funnel and know who owns each stage. Once that is true, AI has something solid to build on.

AI won't fix a broken sales engine

The best AI companies I have worked with have genuinely impressive product, and they lean hard on automation. That is exactly as it should be. But the parts of the business that actually turn interest into revenue, the sales engine, the customer service team, the marketing function, are still run by people and process. When a company is growing fast, that is almost always where it breaks, not in the tech.

This is the COO side of what I do, and it is less talked about than launches and raises because it is not glamorous. It is also where some of the biggest, fastest impact on output and revenue tends to hide.

Growth exposes process. It does not create it.

When a company is emerging or scaling fast, the cracks that did not matter at five people start to break the business at thirty. Leads come in faster than anyone can work them. Handoffs between marketing and sales get missed. Customer service is drowning and nobody has decided what good looks like. Everyone is busy, the numbers are growing, and underneath it value is leaking out of every join in the system.

The founder usually feels it before they can name it. More activity, more headcount, more spend, and somehow the output is not scaling at the same rate. That gap is the process, or the lack of one.

The AI trap

AI-native teams have a particular version of this. Because the product runs on automation, there is an instinct to assume the rest of the company does too. It does not. A model does not chase a warm lead that went quiet. It does not rebuild trust with a customer who is about to churn. It does not get sales and marketing pointed at the same number. Those are human systems, and they need designing just like the product did.

Used well, AI makes a good process faster. It also makes a bad one fail faster, because now you are doing the wrong thing at scale. The teams that win are the ones who fix the engine first, then point the automation at it.

Automation on top of a broken process does not fix the process. It just helps it break more efficiently.

The revenue is usually already inside the business

When a company wants more revenue, the reflex is to go and find more. More channels, more spend, more leads. Often the faster win is the revenue already sitting inside the business, lost to friction. The lead that never got a second touch. The trial that never got onboarded. The customer who would have stayed if anyone had called. Fixing conversion and retention in the engine you already have is usually cheaper, quicker and more durable than buying more at the top.

What this looked like in practice

On one engagement I came in as fractional COO and marketing lead for a fast-scaling education business. The company was growing and the CEO was rightly focused on growth, but the sales and marketing engine underneath needed to scale with it. I built and scaled those processes, managed and executed a full CRM migration, and stood up a bespoke AI management and monitoring solution so the CEO and CMO could finally see lead generation and sales conversion in real time, rather than guessing.

That visibility was the unlock. Once you can see the funnel, you can fix it, and you can keep fixing it without the founder in the middle of every decision. The result was a business that ran on systems instead of on the CEO's time, which is the whole point. It frees the person at the top to keep growing the company.

How I work inside a company

The pattern is fairly consistent. Map what actually happens now, not what the org chart says happens. Find the joins where value is leaking, usually the handoffs nobody owns. Cut the friction, put clear ownership on each part, and set a rhythm so the work ties to a number the business cares about. Then instrument it, so you can see what is working and what is not. And do it in a way that leaves the team stronger, running the system themselves once I am gone.

Signs the bottleneck is your process, not your product

Headcount and spend are rising faster than output.
Leads come in but nobody can tell you what happens to them.
Sales and marketing are working to different numbers.
Customer service is reactive, with no agreed standard.
The founder is still the glue holding the operation together.
You are adding automation on top of steps no one has actually fixed.

Hyper-growth is the best time to do this work and the easiest time to put it off, because everything feels like it is going well. But the companies that turn a fast-growing top line into real, durable output are the ones that fix the engine while they scale, not after it has started to cost them. That is the work, and it is usually where the biggest impact is sitting.