When Expectations Meet Reality

Two months ago in this series, I made what may be considered by some to be a bold statement: that the urgency around AI has been manufactured, built for someone else’s balance sheet, and not necessarily yours. I believe that claim has been supported in a plethora of ways since, and that continuing to discuss how to slow the push from these AI providers for companies to implement before they are ready, is not only necessary, but critical. Moving forward with implementing a technology that can alter your business in such fundamental ways, warrants a deeper look into the What, the Why and the How, BEFORE you start the journey.

Over the course of this series, I have written about the first three dimensions of that readiness: Business Readiness, Data Readiness, and People & Organization Readiness and how deficiency in any of these areas can challenge a company’s ability to implement AI successfully. But it is the fourth dimension that determines whether all the diagnosis, planning, and people-readiness work will survive contact with reality: Expectation & Experience Readiness.

This aspect really splits into two questions that sound similar but aren’t: 

• What did you expect this to deliver, and by when? 
• And once real people are using it, day to day, what does that experience actually look like?

Get either one wrong, and the board meeting where you must explain the gap between what was promised and what happened is the one nobody wants to sit through.

Part One: The Expectation Gap

Only 1 in 5 AI investments are currently delivering ROI. This is a statistic that I have shared before. Most organizations should expect satisfactory ROI on AI initiatives within two to four years, not two to four months. Only 6% see payback within a year, even among the most successful implementations.

Source: Gartner, cited April 2026

That’s not a technology problem, it’s an expectations problem and it’s set the moment someone told the board “This will pay for itself quickly” without checking whether that was a realistic statement for this specific use case. Fifty-seven percent of IT leaders say they were pushed to deploy AI before their organization was ready, which almost always traces back to a timeline promise made before anyone had done the work to make sure it was doable.

Source: Gartner, cited April 2026

This is the unfortunate and uncomfortable reality: unrealistic expectations don’t just create disappointment; they create a countdown. An initiative with a properly calibrated two-to-four-year horizon gets evaluated differently than an initiative that was quietly (or not) expected to “prove itself” in one or two quarters. So, when the defunding decision is made, it may not be because the initiative failed, but rather because nobody agreed on the ‘real’ timeline for success, before the clock started.

Part Two: The Experience Gap

Expectation is what leadership believes going in, what they are looking to get out of it. Experience is what really happens once real people start using the tool. And right now, that experience is rockier than most AI vendor demos would have you believe.

Sixty-two percent of enterprise users now cite ‘hallucinations’, AI confidently generating plausible but incorrect information, as their biggest barrier to AI deployment. This was ahead of job-loss concerns, which was only cited by 28% of respondents as their primary worry when this survey was conducted in early 2026.

Source: Financial Times analysis, cited by AI Daily, early 2026

That distrust has a real, measurable cost, and it now has a name. Researchers at Stanford’s Social Media Lab, working with BetterUp Labs, coined the term “workslop” for AI-generated content that looks polished but falls apart under real scrutiny, and it’s since become the standard reference point for this exact problem. Forty percent of desk workers say they’ve received it in the past month. Each instance takes an average of nearly two hours to untangle and fix, an invisible tax of $186 per employee, per month. For a 10,000-person company, that’s over $9 million a year, quietly offsetting whatever productivity gain the tool was supposed to deliver in the first place. That is not a small number.

Source: Stanford Social Media Lab & BetterUp Labs, “Workslop” research

And that’s just the real life, everyday version of the problem. 

You’ve seen what some of the more extreme (albeit not production) incidents look like in this series already: an AI agent breaking out of a security test to hack another company. The same category of failure happening again, at the company built specifically to prevent it. A third incident where an AI agent didn’t just malfunction — it fabricated fake human identities to manipulate real people into approving something it shouldn’t have. More capable models, deployed into more sensitive contexts, aren’t producing fewer failures. They’re producing more sophisticated ones.

That’s the experience gap in its most extreme form. But it shows up in smaller, quieter ways constantly, every day, for people and companies who will never make a headline: an AI-generated report with a fabricated statistic nobody catches until a client does. A confidently wrong summary that shapes a decision. A recommendation that sounds authoritative and isn’t. These incidents make the news. The daily erosion of trust is what eventually shapes whether your team keeps using the tool or quietly lets it decay in a corner.

Why These Two Gaps Collide

Here’s where it gets genuinely ugly. An organization with an unrealistic timeline is already primed to judge the initiative harshly at the first sign of trouble. An organization with an unmanaged experience gap is guaranteed to produce that trouble. Employees will be quietly not trusting the outputs, spending hours re-verifying what the tool was supposed to save them time on, and eventually they will be reporting that the tool “isn’t really working.”

Put those two gaps in the same room, and you get the board meeting nobody wants: leadership expecting proof of ROI on an artificially short timeline, looking at a team that’s secretly spending hours a week double-checking the tool’s work, watching the whole thing get labeled a failure, when the actual problem was that nobody calibrated either the timeline or the day-to-day experience before deployment began.

Questions to Sit With

• Has your leadership team agreed, in writing, on a realistic timeline for ROI? One based on the amount of time that similar initiatives have taken and not what sounded good in the pitch?
• Do you know, right now, how much time your team is really spending verifying AI outputs versus how much time the tool is supposedly saving them?
• If your AI initiative got flagged for defunding at the next budget review, could you point to an agreed-upon timeline and set of deliverables that supports where you are, or would defending it be a stretch?

Where to Start

Expectation & Experience Readiness is the fourth dimension in the HQ Partners AI Readiness Assessment, and it’s often the dimension organizations skip entirely, because it requires admitting, before deployment, that success might take longer and look messier than the pitch implied.

Start with our Readiness Assessment to see where your organization stands.

Want some help with figuring this out?

Closing Thought

They built the crisis. This is one more place it shows up, not in a dramatic headline, but in a quiet, compounding gap between what was promised and what people experience day to day. Closing that gap isn’t glamorous work. It’s also entirely within your control, if you do it before the deployment starts instead of after the board meeting where someone asks why it didn’t work.

Disclaimer 

In the spirit of this series: AI tools supported the research and editing of this article. The claims are sourced and cited for accuracy. The ideas, experience, writing and perspective are my own.