The Undervalued Risk In AI Product Development: Go-To-Market

Rob Versaw, Product Strategy at Dynatrace, bridging innovation and business impact.

getty

​Roughly 25% of new consumer products are off the shelf within 12 months, and 40% are gone by the end of year two. In nearly every post-mortem I've sat through, the cause of death gets recorded as a product-market fit (PMF) problem. It usually wasn't.

It was a go-to-market (GTM) problem that got reclassified as one because the build was visible and the GTM wasn't. The product is fine. The sales motion was never tested. The buyer wasn't who the deck said it was. The channel didn't fit the product.

We've gotten very good at building things. We have not gotten better at selling them.

The result is a class of products that look successful on the engineering scoreboard, but never become a business. And the cost compounds. The build investment is sunk. The next launch inherits the same blind spot. The team learns to build faster while the gap they couldn't close gets wider.

How Wide Is The Gap Getting?

The toolkit on the build side has compounded for two decades. Engineering rigor, UX research, product-market hypotheses, structured development processes, observability. Every craft on the product side has its own discipline, its own KPIs and its own books. The default org chart reflects that.

GTM has never received the same treatment. In most companies it is still treated as something that happens after the build is "ready." We'll figure that out closer to launch. That assumption was already expensive in 2021. AI makes it more so.

A single engineer with Cursor and Claude Code now ships in a morning what a team of six shipped in a sprint. GTM capability does not scale at the same rate. The gap between what we can build and what we can sell is widening, and most organizations have not noticed.

Having A Commercial Team Is Not Having GTM Capability

The most common pushback from leadership is some version of, "but we have a commercial team." Org chart is not capability. You do not have GTM capability until you can consistently position the product, repeatably sell it, reliably onboard the customer and support it at scale.

Until all four are true, the product is still an experiment. Experiments are how you learn. They aren't assured revenue, and leadership tends to confuse the two. I have sat in product reviews where engineering has 30 slides on technical risk and the GTM section is one paragraph of confidence. That asymmetry is the failure mode, not a byproduct of it. The instinct is to push harder on the build, and no one has the capacity left to ask the GTM questions.

What Actually Breaks

The cracks tend to show up at the same time, the way unbalanced loads expose a weak frame.

Sellers don't know how to sell the product: wrong buyer, wrong message, wrong motion. Incentives reward the legacy product, so the new thing loses every quarterly priority call. Support wasn't prepared, so the first customers get a worse experience than the demo.

Or, the channel doesn't fit the product—a developer tool pushed through enterprise field reps, or a self-serve product gated behind a sales call.

Each one looks small on its own. Together, they are why a launch flatlines after the first 30 logos and then gets reclassified as something to revisit next fiscal year. By then, the team has moved on, and the lesson is lost.

How To Fix It

Treat GTM the way you treat product. Test it—early, deliberately and continuously. Every layer should get the same iteration loop the engineering team uses on the build side. Messaging and positioning. Ideal customer profile and buyer-versus-user. Pricing and packaging. Sales motion and conversion. Onboarding and support readiness.

SafetyCulture is the version of this most operators outside Australia have never heard of. Luke Anear started it in a Townsville garage in 2004, selling printed safety compliance documents to small businesses. He spent seven years working the GTM problem before he had a SaaS product to sell. Which buyers, what message, which channel, what price, what the product actually needed to do.

When smartphones became ubiquitous, he had the customers, the industry knowledge and the distribution to launch iAuditor on top of that base in 2011. The SaaS product found PMF quickly because the GTM was already solved. SafetyCulture is now valued at $2.5 billion on $197.6 million in 2025 revenue. The product story is the part that gets written about. The GTM story is the part that mattered.

I ran a smaller version of the same play at Fortive. We sold a SaaS product before we had anything past a PowerPoint; pricing, ideal customer, buyer, objection patterns derisked first, then we figured out what to build. The play won an internal innovation award and moved our business unit from industrial into industrial tech.

The order felt backward. It wasn't. GTM was always going to be the harder, slower thing. Prove that, and the build compresses.

No one cares that we paid for the tokens to build it if we can't sell it.

The Diagnostic

Every new product introduction carries two questions. Can we build it? Can we sell it?

The companies that prove most successful de-risk both with the same rigor, the same gate criteria, the same willingness to kill the project if the GTM experiments fail.

Three gates worth running before greenlighting the build:

• Buyer Evidence: Can we name the buyer, the budget line and the trigger event, and have we sold a paid pilot or pre-order to at least three of them?

• Motion Fit: Is the channel matched to the product economics? Self-serve, inside sales, field, partner: Pick one and prove it converts before scaling spend.

• Support Readiness: Will the customer onboarding, success and support model be ready at launch, or are we shipping a demo into a service gap?

A product that can't be sold at scale isn't a product. It's a prototype.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?