Skip to content

Every board meeting, every budget review, every conversation with your CEO eventually circles back to the same question: “What’s our AI strategy?” 

AI promises to revolutionize how we work by boosting productivity, accelerating innovation, and giving us a competitive edge in a crowded market. But here’s what nobody wants to talk about, especially with their CEO … most corporate AI initiatives are failing to deliver. 

MIT’s latest research found that despite $30–40 billion in enterprise investment, 95% of generative AI initiatives are producing zero return. That’s not a typo. Zero. 

So what’s going wrong? Our recent survey of 700+ CIOs and CFOs reveals the answer. 

The missing foundation 

The problem isn’t the AI itself, but what it’s built upon. 

According to our research, 84% of IT and finance leaders agree that automation is a necessary first step before successfully implementing AI in business processes. And when you look at outcomes, the numbers tell a compelling story: 

Think about that for a minute: AI by itself delivers results for barely one in ten organizations. But pair it with automation, and suddenly, nearly two-thirds are seeing a significant or transformational impact. 

What you need to break free from AI pilot purgatory 

If you’re like most IT leaders, you’ve probably got a few AI pilots running right now. Maybe they’re showing promise. Maybe they’re generating excitement. But can you scale them? 

This is where most organizations get stuck — in what we call “pilot purgatory.” The tech works great in a controlled environment with dedicated resources, but when you try to industrialize it across departments and workflows, everything falls apart. 

If you’re trying to layer AI onto fundamentally manual, unstructured processes, you’re setting yourself and your organization up for failure. It’s like trying to build a skyscraper on quicksand. The technology might be impressive, but the foundation just can’t support it. 

To escape pilot purgatory and scale AI throughout the business, you need to: 

  1. Build process orchestration first to eliminate bottlenecks and fragmented workflows 
  2. Establish governance within those orchestrated processes 
  3. Use automation to augment existing talent and close skill gaps 
  4. Demonstrate value through scaled deployment to justify continued investment 

Our survey data backs this up: Organizations with advanced automation maturity (58% of respondents) saw tangible value from AI, while those in earlier automation phases (42%) struggled to move beyond moderate impact. 

Why automation matters so much 

At its core, AI is essentially decision-making intelligence. But without automation to execute those decisions across teams, systems, and workflows, AI remains isolated — powerful in theory, but limited in practice. 

Here’s what automation provides that AI desperately needs: 

Standardized processes: AI needs consistent, repeatable workflows to operate effectively. Without automation establishing those patterns, AI is trying to optimize chaos. 

Clean data infrastructure: Nearly half (48%) of survey respondents cite automation’s role in establishing the technical infrastructure needed for scalable AI. Another 47% point to the clean data foundation that automation creates. You know the principle: garbage in, garbage out. Automation ensures the “in” part isn’t garbage. 

Governance guardrails: AI operating without automated guardrails is a compliance nightmare waiting to happen. When AI runs within automated processes, however, you get built-in audit trails, embedded compliance rules, and mandatory human-in-the-loop checkpoints where they matter most. 

Measurable baselines: How do you prove AI’s value if you don’t know what performance looked like before? Automation creates the baseline metrics that make ROI calculations possible. 

One respondent in our survey captured this perfectly: “By combining automation with AI, we reduced administrative workload by 5–7 hours per employee per week and cut compliance reporting time by 30%. Productivity rose 12%, while accuracy improved 15%.” 

Those are the kinds of numbers that make CFOs happy and board members take notice. 

It isn’t too late: Your path forward 

So where does this leave you as an IT leader trying to navigate the AI revolution? 

When we asked our survey respondents about their biggest AI implementation regrets, they gave two major pieces of advice we’d like to pass along to you:  

  1. Don’t underestimate the foundational requirements of successful AI —specifically automation. 
  2. Don’t try to move too quickly, or without proper governance and risk planning. 

While the AI urgency (and pressure from your CEO) is real, it’s riskier to move too quickly and waste time, money, and resources on AI that can’t scale than to pause and make sure you’re doing things the right way.  

Here’s how your peers are doing that now:  

64% of IT and finance leaders say they intend to embed or consolidate AI into their automation strategy if they hadn’t already. They’re moving away from the sprawl of disconnected tools and toward unified platforms that bring everything together. 

This is what Nintex calls agentic business orchestration — a single platform that unites process intelligence, workflow automation, and AI capabilities under one governance framework. Instead of AI being a bolt-on experiment prone to failure, it works as the brain to automation’s muscle, upleveling both technologies to drive measurable transformation. 

The organizations getting this right aren’t treating AI as a standalone initiative. They’re recognizing that AI’s potential can only be realized when it’s built on a foundation of automation that provides structure, governance, and scalability. 

The bottom line 

AI is transformative technology. There’s no question about that. But transformation doesn’t happen in a vacuum. It requires infrastructure, governance, standardized processes, and clean data. In other words, it requires automation. 

The survey results couldn’t be clearer: 95% of organizations combining automation and AI report that their initiatives are effectively delivering on business outcomes. That’s not incremental improvement. That’s a complete reversal of the failure rates seen with AI-only approaches. 

The future of AI in your organization depends not on how quickly you can deploy the latest models, but on how well you can build the automated foundation those models need to thrive. 

The question isn’t whether to invest in AI. It’s whether you’re willing to do the foundational work first — the automation that turns AI experiments into enterprise transformation. 

Want to see all the data and insights from our survey of 700+ IT and finance leaders? Download the full AI UNLESS report to guide your automation and AI strategy.