S4E1 – The DYI Pricing Trap

Summary

Everyone’s using AI. Michael and Avy break down why the DIY pricing trap is real, and why an AI answer still isn’t the same as a strategy your team trusts.

Key takeaways

AI can produce a pricing analysis in minutes, but it cannot get a sales team, finance, or leadership to actually trust and act on it.
The scariest part of DIY pricing is invisible: a homegrown model rarely shows you the margin, win rate, or revenue you left on the table.
Internal pride in a self-built AI pricing tool can make teams less willing to question it, not more, which quietly compounds the risk.
Change management, not analysis, is the part of pricing work that most companies underestimate and AI cannot do for them.
A durable pricing capability still requires three connected pieces: a clear strategy, a properly built tool, and a plan for adoption.

Everyone has an answer now. Not everyone has a strategy.

The question is not new, but AI has made it louder. If a model can generate a price recommendation, a discount threshold, or an elasticity estimate in seconds, what is left for a pricing expert to do?

Michael and Avy do not dismiss the question. AI genuinely compresses the analysis. Where it falls short is everything that happens after the analysis: getting a regional sales lead to change a discounting habit, getting finance comfortable with a new margin target, getting a customer-facing team to hold a price increase under pressure. An answer that nobody in the business trusts is not a strategy. It is a document.

The cost you never see is the one that should worry you most

A DIY pricing model does not usually fail loudly. It fails quietly, by leaving value on the table that nobody ever notices was there. There is no line item for a deal that could have closed at a better margin, or a price increase that could have held if the team had pushed a little further. That invisibility is what makes it dangerous. A team can run a homegrown model for a year, see revenue holding steady, and conclude the tool is working, without ever knowing what a properly built pricing capability would have captured instead. The absence of an obvious problem is not the same thing as good pricing.

Homegrown pride is quietly setting companies up to fail

There is a specific trap Michael and Avy call out directly: when a team builds its own AI-assisted pricing tool, it tends to develop real pride in it. That pride is understandable. It also makes the tool harder to question.

A model that came from a vendor or an outside expert gets scrutinized. A model the team built itself, and is proud of, tends to get trusted by default. That is precisely backwards. The origin of a pricing model has no bearing on whether its assumptions hold up, and the more ownership a team feels over a homegrown tool, the less likely anyone is to stress test it before it makes a costly recommendation.

Change management is the part of pricing no one wants to do

Michael and Avy return to a theme that shows up across every season of the podcast: most pricing work is not analysis, it is execution. Convincing a skeptical regional lead. Holding the line with a demanding customer. Getting a cross-functional team to actually use a new tool instead of quietly reverting to spreadsheets.

AI has done almost nothing for that part of the job. No model walks into the room and makes the case in person, and no prompt builds the organizational muscle to hold a pricing decision once the room pushes back. That is the work that determines whether a pricing strategy shows up in the numbers or just sits in a deck.

What this means for your pricing capability

The episode’s conclusion tracks with how RML structures its own work: a pricing capability that lasts is built from three connected pieces, not one clever model.

AI has made the first two faster and more accessible than ever. It has not changed the third, and the third is usually where DIY pricing efforts quietly come apart.