I keep hearing a version of the same story from high-tech marketing leaders.

The pilots are working. Product marketing is using AI for content. Demand generation is applying it to audience targeting. Customer marketing is automating lifecycle programs. Each team can point to a working use case.

Then someone asks: Which of these should we fund next, and how much should we invest? That question is surprisingly difficult to answer.

Each team has measured its work against different goals, using different systems and different definitions of success. There may be strong results inside individual programs, but very little basis for comparing them or deciding where the next dollar should go.

This is where many high-tech companies find themselves today. There is no shortage of capable AI technology. Experimentation is rampant. The challenge is that these efforts are often siloed.

The CMO, VP of demand generation, head of customer marketing, and leaders in marketing operations, sales, and IT may all be moving AI forward, but they work with different priorities, data, and decision-making processes. Individually, there is progress. Collectively, it becomes much harder to turn that progress into a repeatable way of working.

The Frozen Middle in high tech.

Adobe worked with Incisiv to study 5,633 executives across 12 industries, including 521 in high-tech. The research found that 44% of high-tech organizations sit in what we call the Frozen Middle.

These companies have moved beyond experimentation. AI is already active within individual functions, where they struggle to take a successful use case from one team and scale it across the organization.

Content generation, personalization, account intelligence, and ABM may all be running somewhere in the business. But when each is funded, managed, and measured independently, leaders have a difficult time demonstrating ROI beyond an individual campaign, comparing investments, or deciding which use cases deserve more funding.

It’s easy to see how companies get here.

AI tends to begin with one use case at a time. Each team moves at its own pace. Meanwhile, many of the processes connecting marketing, data, and technology were designed around annual planning cycles and quarterly reviews.

AI moves much faster. Teams need to iterate quickly, know who has authority to make decisions, and have shared access to the data required to make those decisions.

Without those conditions, individual deployments accumulate. And the more they accumulate, the harder they become to connect and govern.

The approval data illustrates the problem. Out of every 100 AI initiatives proposed in high tech, only 18 reach production. Integration assessment removes the largest share at the first gate.

Questions about data rights, system ownership, and integration often surface late in the process, even though they could have shaped the use case from the beginning.

What Leaders are doing differently.

The research identified 342 organizations across industries that scored in the top 6% for operating-model maturity. Looking at how these Leaders work, five practices stand out.

1. Leaders have a named AI owner who owns outcomes across functions: Leaders are 2.3 times more likely to have a named AI owner with a defined mandate and budget authority. In the high-tech industry overall, only 40% of organizations have appointed an AI owner, and where ownership exists, it often stops at the functional boundary.

That becomes a real problem once AI workflows cross teams. A lifecycle use case might require product usage data, CRM access, and personalization through the customer experience platform. Someone needs enough authority to resolve those dependencies and keep the work moving.

This doesn't necessarily require another executive or another layer in the organization. A CMO or VP of marketing can play the role. What matters is giving that person enough scope to make decisions across budgets, platforms, data, and teams.

2. Leaders decide how they'll measure value before they launch: 44% of Leaders measure ROI comprehensively, compared with 20% of high-tech organizations. That means defining the outcome, baseline, and attribution approach while the use case is still being designed, with teams such as revOps and customer success aligned on how the results will be interpreted.

This matters because baselines become remarkably difficult to recreate later. Once a campaign, product launch, or customer interaction changes behavior, teams can spend more time debating what AI contributed than learning from the result.

Building measurement in from the beginning gives marketing leaders much stronger evidence when they return to the investment conversation.

3. Marketing and technology share the outcome: 81% of Leaders operate at a collaborative or integrated level, compared with 41% of high-tech organizations. 59% of high-tech still operate through siloed, transactional, or consultative relationships.

Shared objectives and key results (OKRs) are one of the clearest differences.

When “demand generation” and “marketing operations” are accountable for the same commercial outcome, the brief, build, and measurement are much more likely to stay connected. Data and timing issues can be resolved as the work progresses instead of becoming another handoff.

Much of this can happen within existing reporting lines. It doesn't automatically require reorganization.

4. Leaders learn from the things that don't work: This was one of the findings that I found most interesting. 16% of high-tech organizations show destructive failure patterns characterized by blame and repetition. Among Leaders, that figure is under 1%.

High-tech marketing teams work against compressed timelines. When an account-scoring model misfires or AI-generated content loses quality as it scales, there is enormous pressure to fix the immediate problem and move on.

But those failures contain useful information.

An ICP segmentation model that performs poorly can reveal something about the data, assumptions, or process behind it. An ABM sequence that misses the mark can show where orchestration broke down.

Leaders create room in operating reviews to examine those signals and carry the learning into the next use case. That discipline becomes increasingly important as the number of AI initiatives grows.

5. Leaders find and empower internal champions: The culture finding is probably my favorite. 42% of Leaders said finding and empowering internal champions was their single most effective culture action, ahead of executive sponsorship at 16% and formal training at 11%.

That feels very true to what I see in practice.

People become interested in AI when they see someone they know using it to make familiar work better: a demand generation manager improving an ABM brief with stronger intent signals, a content marketer accelerating launch assets, or a lifecycle analyst turning product data into better conversion triggers.

Those practitioners can have an outsized impact. Give them visibility, some resources, and permission to bring what they've learned to another team, and adoption starts to spread through the organization.

From pilots to a way of working.

Taken together, these findings point to where the next phase of AI adoption is heading for the high-tech industry.

Many companies already have the tools. They already have use cases. And they already have people proving that AI can create value.

The next question for marketing leaders is how well their organization can take what works in one team and turn it into something repeatable across many.

That requires clear ownership, common measures of value, shared accountability across marketing and technology, room to learn from failures, and practitioners who can carry successful approaches across the organization.

Those are the conditions that allow AI to move beyond a collection of promising pilots and become part of how the business operates.

Explore the full report to see where high-tech organizations stand and what Leaders are doing differently.

https://business.adobe.com/fragments/resources/cards/thank-you-collections/high-tech