For most of the past three years, the hard part of enterprise AI has come after the demo. Gartner estimateds that at least half of GenAI projects were abandoned after proof of concept by the end of 2025, and it points to familiar causes: poor data quality, token costs that grow quickly once a pilot reaches real users, unclear business value and risk controls added as an afterthought.
The picture is starting to shift, though the gains remain concentrated. BCG’s Applied AI Index 2026 found that nearly half of companies now generate measurable value from AI.
In BCG’s research just the year prior, only 5% were generating substantial value.
Companies are also spending more. AI budgets rose from about 1.7% of revenue in late 2025 to 3.3%, and more than 80% of that spending now sits outside the IT budget.
That gap between ambition and execution has made engineering and services firms central to how AI reaches the enterprise. Model providers have built partner ecosystems around the idea.
In March, Anthropic committed an initial $100 million to its Claude Partner Network, offering partners training, certification and dedicated Applied AI engineers on live customer deals. “Our partners are instrumental in getting enterprises from proof of concept to production with Claude,” said Steve Corfield, Anthropic’s head of global business development and partnerships, at the launch.
Gorilla Logic, an AI-native engineering company that builds and modernizes business-critical technology for enterprises, announced this week that it had named Steve Daly as CTO, days after joining that network as a Select Services Partner.
Steve Daly (article’s featured photo) has spent more than 35 years in technology and consulting leadership. Most recently he was senior vice president of solutions, global digital transformation at New Era Technology, where he led the company’s global AI practice and directed initiatives across AI, cloud transformation and enterprise innovation.
At Gorilla Logic, his remit lines up closely with the problems the Gartner and BCG research describe. He will help lead the company’s work with organizations moving AI from experimentation into production, with a focus on the architecture, data and engineering foundations that make results repeatable in business-critical environments.
Drew Naukam, Gorilla Logic’s CEO, framed the hire around that gap. “AI is changing what companies can build and how quickly they can build it, but enterprise adoption requires much more than access to a model,” he said. “Steve adds the leadership and experience to help us connect those capabilities to the complex systems, data and engineering realities of the enterprise.”
Daly put more weight on the people side. “AI creates tremendous opportunity, but technology alone doesn’t create transformation,” he said, describing the work as “connecting AI to the business, and the people who have to operate it every day.”
Gorilla Logic, he added, “has the engineering depth to help clients move beyond experimentation and build AI systems designed for the realities of the enterprise.”
The company works with global brands, private equity firms and high-growth companies on new product development, platform consolidation, application modernization and ongoing engineering. Much of that work runs through Construct, Gorilla Logic’s portfolio of delivery-tested workflows and modular AI agents. Under Daly, the company plans to expand its AI-native product and platform engineering, applying the same engineering discipline to both.
The appointment reflects a broader change in what enterprises are buying. Capable models are widely available, and the BCG data suggests budgets are moving into business units that want results they can measure. The harder thing to find is the experience to wire those models into legacy systems, messy data and the controls that regulators and boards now expect. Hiring a CTO with decades of large modernization programs behind him is a bet that this is where the next round of enterprise AI work will be won.
If the share of companies abandoning AI after the pilot stage falls and the share reporting real value keeps climbing, much of the credit will belong to the engineering teams doing the integration work.
