Adobe Study Finds 61% of Consumer Goods Firms Stuck in AI Limbo

A promising artificial intelligence pilot can produce content, audience segments or sales forecasts—and still fail to deliver lasting value once it leaves the team that built it. That is the central warning in a new Adobe and Incisiv study of AI readiness, with implications for small consumer-goods companies that often have fewer staff, tighter budgets and less control over retail partners.

Adobe and Incisiv surveyed 5,633 executives across 12 industries and 11 markets, including 521 respondents from consumer goods. The research found that 61% of consumer-goods respondents are in what it calls the “Frozen Middle”: AI is operating in individual departments, but companies have not built the coordination needed to scale those experiments.

“The constraint is rarely the model. It’s what happens to a good use case the moment it needs something outside the team that built it,” Adobe said in the release.

For a small brand, that problem can surface quickly. A marketing employee might use generative AI to create product descriptions or ad variations. A sales team might build a retail media audience. But the project may require customer data, legal review, retailer approval, new software integrations or a way to measure sales impact. If no one owns those decisions, the pilot can stall before it becomes part of daily operations.

Key takeaways

  • Only 16% of consumer-goods respondents have a formal AI owner with a title and budget authority.
  • Out of every 100 proposed AI use cases, just 15 reach production.
  • Seventy-six percent still manage business and IT through siloed, transactional or consultative relationships.
  • Only 12% measure AI’s return comprehensively.
  • The research points to ownership, early measurement, cross-functional teamwork and internal champions as more important to scale than simply buying new technology.

The approval process shows where companies lose momentum. The largest absolute drop occurs during IT evaluation, while the steepest proportional losses come later, during integration, budgeting and production-readiness reviews. Adobe said the study does not explain why every use case was rejected, but the pattern suggests that organizations often ask basic feasibility questions too late.

That finding matters for small businesses because a failed pilot can consume a disproportionate share of available time and money. Before committing to an AI project, an owner or leadership team should identify what data the system needs, who can access it, which outside partners must approve the work and how success will be measured. Those checks do not eliminate risk, but they can prevent a company from spending months on a project that cannot legally, technically or commercially move forward.

The same issue applies to retailer relationships. A retail-media initiative can involve brand leaders, ecommerce staff, sales teams, technology providers, media publishers, retail partners, distributors and marketplaces. A small company may not control the data-sharing terms, approval schedule or technical requirements imposed by a retailer. An AI workflow that works inside the business can therefore stop when it reaches an external partner.

In practical terms, small businesses may benefit from starting with tightly defined use cases. Automating repetitive product-content work, summarizing customer feedback or improving demand-planning analysis may require fewer outside dependencies than building a retail-media system. Businesses should still review privacy, accuracy and brand-safety risks, particularly when AI generates customer-facing material or uses personal information.

Measurement also needs to start before launch. Tracking the number of marketing assets produced or hours saved can help establish operational value, but those figures do not necessarily show whether AI increased purchases, margins or customer retention. Adobe’s research found that companies often measure what is easiest to count rather than the business outcome they intended to change.

A small company could establish a baseline for conversion rates, average order value, repeat purchases or production costs before introducing an AI workflow. It can then compare results against that baseline while accounting for promotions, price changes, distribution shifts and seasonal demand. That approach will not produce perfect attribution, but it can support a more credible investment decision.

The study’s most mature group, 342 “Leaders” representing the top 6% of active respondents across all 12 industries, provides a benchmark rather than a guarantee of success. These organizations were 5.8 times more likely than consumer-goods companies to have a formal AI owner. Forty-four percent measured return comprehensively, compared with 12% in consumer goods, and 81% operated business and IT at a collaborative or integrated level, compared with 24% in consumer packaged goods.

The Leaders also reported fewer destructive failure patterns and placed significant value on internal champions. Forty-two percent identified finding and backing internal champions as the most impactful cultural step they had taken.

For a small business, that champion might be a marketing manager, operations lead or salesperson who can test a tool, document the results and train colleagues. Giving that person clear authority, time and a defined budget may be more practical than forming a large committee. Leadership should also make clear when human review remains mandatory.

The broader message is that AI readiness is not limited to selecting a model or software subscription. Small consumer-goods companies that define ownership, involve technology and legal advisers early, agree on measurements and account for partner dependencies will be better positioned to turn isolated experiments into repeatable capabilities.

Adobe’s full report is available here.

Image via Google Gemini

This article, “Adobe Study Finds 61% of Consumer Goods Firms Stuck in AI Limbo” was first published on Small Business Trends

SHARE IT: