New Alteryx research of 1,400 leaders shows AI investment is rising, but weak data, low trust, and governance gaps are slowing pilots from reaching production.
Even as corporate spending on artificial intelligence accelerates, most companies remain stuck in experimentation mode.
New research released Tuesday by Alteryx, the data and analytics software provider, suggests that trust — not ambition — is the chief obstacle preventing AI pilots from becoming operational at scale.
The survey of 1,400 business and IT leaders across the Americas, EMEA and Asia-Pacific regions found that fewer than one in four AI pilots successfully make the leap into production. Despite rising budgets and mounting executive pressure, organizations continue to struggle with low confidence in AI outputs, poor data quality and legacy systems ill-suited for scale.
The findings underscore a widening gap between AI aspiration and measurable impact.
Trust Lags Behind Investment
While nearly half of respondents said they trust AI to automate repetitive tasks, draft content and monitor systems, confidence drops sharply when higher-stakes decisions are involved. Only 28 percent trust AI to support decision-making, and just 27 percent trust it to assist with forecasting or planning.
In other words, AI is welcomed as a co-pilot — but not yet as a strategist.
At the same time, investment shows no signs of slowing. Nearly half of leaders said they plan to increase spending on AI infrastructure and tools, and 89 percent expect to maintain or grow their AI budgets in 2026. AI platforms, which accounted for roughly a third of data stacks in 2024, are projected to comprise more than half of data stacks within three years.
The tension is clear: companies are buying more AI, but not yet trusting it with consequential decisions.
Also Read: AI Is a Leadership Test, Not a Tech Rollout
The Data Problem
Nearly half of respondents — 49 percent — identified high-quality, accessible and well-governed data as the single most important factor for realizing the promise of agentic AI systems. Without clean inputs and clear lineage, AI systems risk producing inconsistent outputs, hallucinations and results that vary from one query to the next.
Many organizations, the report notes, are layering generative AI directly on top of raw data sources without sufficient business logic or guardrails. The result is diminished confidence — and stalled initiatives.
As a corrective, 28 percent of leaders said they plan to prioritize data governance improvements. The report argues that enterprises must combine generative AI’s creative capacity with deterministic rules, defined metrics and workflows that can be adapted quickly as business needs evolve.
Trust, in other words, must be engineered.
“AI adoption is accelerating fast,” said Andy MacMillan, chief executive of Alteryx. “Compared to a year ago, two-thirds of business and IT leaders are using AI more in their roles.”
He added that AI ownership is also shifting. Over the next three years, leaders expect responsibility for AI workflows to move away from centralized teams and toward individual lines of business — rising from 22 percent today to 33 percent by 2028.
The companies furthest along, MacMillan said, are investing in data quality while embedding AI deeper into day-to-day operations.
Also Read: Cloudflare’s Data Chief on the Internet’s Fragile Future
From Pilot to Production
The survey, conducted by Coleman Parkes between August and September 2025, included respondents from banking, insurance, manufacturing, retail and the public sector.
Its message is less about technological limits than organizational readiness. Enterprises are eager to experiment. They are less prepared to institutionalize.
The report concludes that scaling AI will require more than capital. It will demand governed data, operational clarity and confidence that systems can deliver consistent, explainable results.
Until then, the AI revolution may remain largely confined to the lab.
