Why 70% of AI Implementations Fail and What HR Leaders Can Do About It

There’s a number that keeps surfacing in every boardroom and strategy session. Everyone points to it: Harvard, McKinsey, BCG. But the real question isn’t who cites it. It’s what leaders actually do about it.

Seventy percent of change initiatives fail. 

Not because the strategy was flawed. Not because the technology didn’t work. Because the people inside the organization never truly believed in what they were being asked to do.
 
This challenge is not isolated; it is unfolding right now across thousands of organizations attempting to roll out AI. As Ted Wolf, founder of Guidewise, points out, most are heading straight into that same wall for the same reasons, making the same avoidable mistakes.

Ted knows what it takes. Over 18 years, he built a technology staffing company from a $1,000 startup into a 650-employee organization with 13 offices, eventually leading it to an acquisition by a Fortune 1000 company. Along the way, he oversaw hundreds of technology rollouts and began noticing the same patterns emerging again and again. He later spent years overseeing more than 300 ERP and technology implementations, further reinforcing those insights.

  • The technology worked.
  • The business case was there.
  • The people didn’t change.

That experience became the foundation for Guidewise and its Changeworks platform, a system built to measure what most organizations have never thought to measure: what employees actually believe, feel, and make meaning of during a change initiative. Not project milestones. The inner game.

🎧 LISTEN: Hear Ted’s Full Story 

Want to hear the complete conversation? Listen to Ted Wolf’s episode on Courage to Advance, where he walks through real examples of pilot purgatory in action, how the Changeworks platform measures employee beliefs and emotions during an AI rollout, and his advice for HR leaders who want a seat at the table before the plan is already built. [Listen to Episode →] 

AI Is Not an IT Project. Stop Treating It Like One.

Ted identifies a critical, costly error right away: assigning the AI rollout solely to IT and treating it as just another tech project is unacceptable.

IT teams are indispensable. Their expertise ensures system stability. But they are not equipped to address the real driver of AI success: how people adapt to changes in the way they work.

Why AI Is Different from Other Tech Rollouts

Unlike a standard software implementation, AI touches your organization in ways most leaders haven’t fully reckoned with:

  • It doesn’t just change one team’s workflow. It connects to every department’s data.
  • It puts information in everyone’s hands and breaks down the silos you might not even see.
  • It changes reporting structures, daily tasks, and in many cases, people’s sense of what their role even is.
  • If those changes come from the top without real preparation or openness, trust breaks down quickly.

“When trust begins to erode, that means leadership did not communicate very well why the change is necessary and what the impact is going to be on workflow and day-to-day tasks. And once your trust starts to erode, it’s a long, difficult road to come back.”

What the Trust Study Tells Us

SparkEffect’s own Trust Study directly reinforces this. AI-driven system changes are the number one disruptor organizations face right now. And yet, they are not the biggest trust-breaker. Leadership behavior during the rollout is.

The gap between what leaders announce and what employees actually experience is where trust breaks down. By the time most organizations notice it, the damage is already done.

What Is Pilot Purgatory and Are You Already In It?

Ted coined the term “pilot purgatory,” and once you hear it, you’ll recognize it immediately.

The Pattern That Keeps Playing Out

Pilot purgatory is what happens when an organization runs multiple AI pilots across departments, none of them tied to a shared business objective:

  • Sales runs its own pilot.
  • Marketing runs its own.
  • Operations has something going on in the background.
  • Six months pass, and there’s still no clear result.  

The core problem isn’t the pilots. It’s the absence of a clear answer to what Ted calls the fundamental questions of change: the who, what, where, when, why, and how. Without those answers communicated consistently from senior leadership to the frontline, employees have no way to process what’s happening. So they don’t. They work around it, wait it out, or disengage entirely.

Three Things That Break the Cycle

Getting out of pilot purgatory starts here:

  1. A clear business objective
  2. Shared metrics that tell you whether you’re succeeding
  3. An actual timeline with milestones, not a vague “we’ll revisit this in Q3.”

People can orient themselves around milestones. They can’t orient themselves around ambiguity.

The Inner Game: The Variable Nobody Is Measuring

Here is the part of the AI conversation that few are having.

Every employee navigating a major change initiative is running a quiet internal calculation. It shapes engagement levels, productivity, and adoption rates more directly than any training program or communication strategy. Ted calls it the inner game.

What the Inner Game Is Made Of 

The inner game sits at the intersection of three things:

  • Beliefs: What do employees actually believe about what leadership is telling them?
  • Emotions: What are they feeling day-to-day as the rollout unfolds? 
  • Meaning: What does this change say about their future, their career, their value here?

“They’re telling me why it’s important, but do I believe them? Do I trust what they’re telling me? Do they have my best interests? Are they telling me the truth? Or are they just telling me this so they can replace me next year?”

Until those questions are honestly answered, no project plan, no executive announcement, and no change management methodology will move the needle. People do not change because they are told to. They change when they believe it is safe to do so, when they understand what it means for them, and when they trust the organization asking them to do it.
 

How Guidewise AI Surfaces What’s Hidden

Ted’s Changeworks platform was built to surface exactly this. It generates 40 human-centered metrics that give leaders real-time visibility into employee belief systems, emotional patterns, and meaning-making, something no standard project management tool has ever provided.
 
As Ted puts it: Asana can tell you whether a task is on track. It tells you nothing about whether the person completing the task believes in what they’re doing.

HR’s Seat at the Table Is Not Optional

One of the most important and most commonly missed structural failures in AI implementation is where HR sits in the process.

How It Usually Goes and Why It Fails

In most organizations, the AI conversation starts in the C-suite or IT. By the time HR is informed, the key decisions have already been made. HR is handed a communication plan and asked to manage the rollout of something they had no part in designing.
 
Ted is direct about why this is a problem. AI is not a standard technology project. It touches:
  • Every department’s data
  • Every employee’s sense of career security
  • Every manager’s ability to lead their team through uncertainty

Where HR Needs to Be Instead

The human complexity of an AI implementation requires HR expertise from the beginning, not as a communication function, but as an implementation architect.
 
That means HR is in the room when:
  • The business objective is being defined.
  • Decisions are being made about which departments are involved.
  • Governance structures are being built around AI agent accountability.
  • Policies around who uses which tools are being set
 
These initial decisions are critical. They determine whether an implementation succeeds or quietly breaks apart months later.

The Dual Workforce Is Coming. The Culture Question Is Now.

One of the most important reframes Ted offers is about where all of this is ultimately headed.

What the Future of Work Actually Looks Like

The future of work, in Ted’s view, is a dual workforce: human employees and AI agents working in tandem, with people managing agents the way they currently manage other people.

“You’re not going to lose your job. It’s going to be different. You’re going to manage an agentic workforce. Managing 10 agents is no different than managing 10 employees. You have to check on their work, onboard them properly, have policies they follow.”

Why This Reframe Changes Everything

The narrative that AI is coming to replace jobs is the one that kills trust and tanks adoption. The more accurate and more useful narrative is that AI is changing jobs.
 
The organizations that bring their people along in understanding what it means and how it affects their careers are the ones that come out ahead. And building a culture resilient enough to absorb that transition doesn’t happen during the rollout. It happens:
 
  • In the months and years before
  • In the small decisions about how leaders communicate
  • In how trust is earned and maintained before it’s ever tested
  • In how employees are prepared for change before it arrives at their desk

What Getting It Right Actually Looks Like

Ted defines successful AI implementation as three critical elements working in unison. While most organizations get the first two right, very few accomplish the crucial third element.

The Three-Part Framework

  1.  A clear business objective that the AI is meant to serve
  2. Technology selected specifically to serve that objective
  3. A people adoption plan designed to get employees to actually use it

What a Real Change Adoption Plan Requires

A change adoption plan is more than communication or training. It is a structured way to track employees’ beliefs before, during, and after the rollout. That means:

  • Collect people metrics, not just project metrics- Identify employee beliefs about the change
  • Notice emotions that surface
  • Find out what meaning employees areassigning to the change

Without this data, leaders make change decisions blindly.

Top-Down and Bottom-Up, Together

The macro directive from senior leadership matters. But it only reaches its full effect when matched by genuine engagement at the frontline when managers have the tools and the trust to have honest conversations with their teams, and when employees get a clear answer to the question that matters most: What does this mean for me? 

Download the Complete Research 

Get the full SparkEffect Trust Study with detailed breakdowns by industry, organization size, and leadership level. [Research link] 

Hear Ted’s Full Story

Listen to the complete Courage to Advance conversation, including how Ted navigated real pilot purgatory situations inside organizations, the specific metrics Changeworks uses to measure employee beliefs and emotions during an AI rollout, and his advice for HR leaders who want to get in the room before the plan is already built. [Podcast link] 

Work with SparkEffect 

If your organization is facing the kind of crisis Tammy inherited, or you want to build trust infrastructure before crisis hits, we can help. Schedule a conversation to explore how: [Contact link] 

About SparkEffect 

SparkEffect helps organizations build trust that bends without breaking. Our Trust Performance Index and consulting services give leadership teams the tools to diagnose trust gaps and build the infrastructure required for lasting transformation. 

The Courage to Advance podcast features leaders building the organizations they wish existed, proving that business doesn’t have to be dehumanizing. 

Recommended reading

Episode 12: Featuring Ted Wolf, Co-Founder & CEO, Guidewise Seventy percent of change initiatives fail. Not because the...

Episode 10: Featuring Dr. Suzi Potts, Founder, SP Consulting Picture this: A small business owner walks into a...

Episode 10: Featuring Mitchelle Jeffery, Founder of The Ember Collective Your policies are failing you. You just don’t...