Building Vs. Buying In the Age of AI: The Sales Leader’s Guide

As I discovered on vacation last summer, almost anyone can build a software application with Claude. The great democratization of coding has arrived, and you don’t need even five minutes in a C# or Python class to make software that does what you want it to do. This can feel like total freedom from clunky, expensive platforms you have to pay through the nose for.

But before you cancel all your SaaS subscriptions, spin up an account on Claude and start coding, check that you’re not creating more problems than you’re solving. Here’s what you need to know.

1. Think Past The Moment

Coding with AI can be a real rush. So much power at your fingertips. But, as I learned this summer, what you create still has to be maintained. You will still need security reviews, bug fixes, user feature updates, API/MCP updates, and fixes for broken integrations. What started as a productive week can quickly turn into a full-time job chasing problems, unless you think past the moment and plan for the full picture.

2. Avoid the Two Big Traps

I’ve spoken in the past about “point pollution” and “Frankenstacks.” The AI age has made it even easier to fall into the same traps, but faster and cheaper. You spin out an app real quick to solve a specific workflow, and then later spin one out to solve another point problem, and before long you’ve got a labyrinthine mess of code that slows itself down and eats up your people’s time, not to mention your own. The reality check is that even when you code it with AI, the software still has to interface with existing systems, still needs to serve a meaningful purpose, be secure, and still needs your team to actually use it.

3. Evaluate Your Way of Selling

Building custom applications with AI can be a smart choice when you take the time to understand what you’re building on and for. Before you clutter up the software landscape, evaluate your company’s Way of Selling, your strategy, process, methodology, training, reinforcement, team skills, and execution.

Before you clutter up your software landscape, evaluate your company’s Way of Selling.

Sometimes, the fastest way to improve is not found in yet another technology tool, but in how your approach is structured.

4. Evaluate Your CRM

Your salespeople should come to work every day knowing exactly what they need to do and how to do it, and trusting that they will be guided through the process and provided with support when they need it. A B2B tailored CRM gives them this, while also providing you with insight into the analytics that matter. Everything else in your tech stack depends on that and should support that. When your CRM is solid and accessible by AI via connectors, you won’t need so many work-arounds and any custom apps can be tailored precisely to solve the right problems without distracting from the main task.

When you approach your tech stack strategically in this way, your adventures with AI can become more than just a fun holiday activity. Instead, they can become a highly functional part of a highly effective selling system.