AI Made Enterprise Software Cheaper To Prove. Here’s How I’m Adapting As A Marketing Leader

Tobias Fischer founded HigherUp Digital (B2B Tech Agency), Sermondo (Enterprise Tech Media), and is VP Marketing at Euvic US (Software Dev).

gettyTwo years ago, almost everyone I talked to expected AI to shrink the software engineering market. Cheaper code would mean fewer engineers and less spend.

It went the other way. The U.S. Bureau of Labor Statistics projects employment of software developers, QA analysts and testers to grow 10% from 2025 to 2035, “much faster than the average for all occupations.” Gartner forecasts worldwide IT spending of $6.37 trillion in 2026, up 14.2%.

This is Jevons paradox. In 1865, economist William Stanley Jevons observed that more efficient steam engines increased Britain’s coal consumption rather than reducing it. When something gets cheaper to use, we use more of it. The question for software is which part actually got cheaper.

I’m vice president of Marketing at Euvic US, a software development and IT services group with more than 6,000 people. Over the past 18 months, the question prospects bring up has changed. They used to ask, “Is this worth building?” Now many arrive with a prototype built in days and ask, “Can you make this real?”

The bottleneck was never writing code. It was the months of committee work needed to justify a build. AI prototyping removed most of that, so projects that sat in the “someday” column for years are getting approved. Software didn’t get much cheaper to finish. It got much cheaper to prove, and every approved idea still has to become a production system.

Software moves through a pipeline: requirements, code, review, testing and release. AI has sped up one stage. Faros “AI Engineering Report 2026,” covering 22,000 developers, revealed that task throughput is up about 34% on teams with high AI adoption. Over the same period, code review time rose fivefold and incidents per change tripled.

Delivery moves only as fast as its slowest stage. That is why DX’s longitudinal study puts the average gain in engineering output at around 10%.

AI also adds work to those slower stages, because every AI-generated spec and pull request needs a human to verify it. On our own projects, we’ve seen a 50-page AI-written spec nobody could check before the build. We’ve also seen a design that cost up to $2,000 in AI spend get thrown out after a late requirements change. Neither was a tool failure. Both were coordination failures.

Security widens the gap. Veracode’s “2025 GenAI Code Security Report“ found that 45% of AI code-generation tasks introduced a vulnerability.

Most companies have added AI to the way they already worked. Each function uses AI for its own step, and work still gets handed from one group to the next. That is where the 10% comes from.

Google’s “2025 DORA Report“ calls AI an amplifier. It strengthens teams with good practices and exposes those without them. The teams getting more from AI put every role in one team, use AI at every step and give one person ownership of quality and AI spend. We’ve seen this in our own delivery work at Euvic US. On one full build, a team organized this way delivered about 50% more than a traditional team with the same scope and headcount, based on our internal project data​​.

Building this capability in-house takes time few companies have. MIT’s “State of AI in Business” research (registration required) found AI initiatives built with specialized partners succeeded about 67% of the time versus about a third as often for internal builds.

But AI also made every vendor look capable, since polished websites and case studies now take an afternoon to produce. And Gartner found 67% of business-to-business (B2B) buyers prefer a rep-free experience. Buyers are judging firms with less contact, just as surface signals mean the least.

As vice president of Marketing at Euvic US, this has changed my role more than any new channel or tool. Prospects no longer need us to prove an idea is worth building. What they want de-risked now is the build.

They want to know we have taken an application from prototype to live, at a scale where millions of people will eventually use it and that it has passed real security reviews along the way.

I now spend more time with our delivery teams (managed teams) as the main source of content and proof points. I sit in on project reviews, talk to the engineers who handled the difficult migrations and ask what nearly went wrong and how they caught it. Those details are what a prospect can’t find anywhere else.

The distribution channels and formats haven’t changed—the level of insights has. Case studies get enriched with interview recordings, sales decks are always personalized and highly relevant.

Our proof points have had to become human-first: named engineers and specific decisions and outcomes we can verify. When every vendor can produce the same polished content, the thing that stands out is evidence that real people have done this work before and can explain exactly how.

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