AI Reality Checks: Patience For AI ROI, Blame For Contact Center Agents

“AI Reality Checks” are all the rage in the CX space lately. While I’m glad to see it, I can’t help but notice most of the conversation is geared towards executives, management, and boards: adjusting and aligning stakeholder expectations, extending ROI timelines, rightsizing implementation budgets, educating C-Suite leadership, resetting KPI frameworks, creating change management protocols and guardrails.

The list goes on, with one recurring theme: be patient, temper your expectations - and never, ever insinuate that the AI technology itself is the problem.

Patience For Technology, But Not For The People Powering It

While executives (and, to a certain extent, investors) seem uncharacteristically patient when it comes to AI-led transformation at scale, they don’t extend that same patience to the people actually implementing, using, and monitoring those fancy AI-powered tools: frontline agents.

Why don’t any of these AI Reality Checks advocate for recalibrating executive expectations that agents should become "AI Whisperers" overnight with little or no training? Where's the reality check about what it actually feels like to be monitored, coached, and scored in real time by three different AI systems, all while trying to help a customer who is furious because the AI agents couldn’t solve their problem?

Why don’t these AI Reality Checks mention the need for leadership to adjust agent performance metrics to account for the new job requirements that AI has created? Where’s the timeline extension for agents that have to learn how to use a new AI tool every week? Why are agents punished for extended handle times when AI performance monitoring tools force them to follow nonsensical call scripts, or when Agent Assist provides incorrect advice?

Where’s the reality check when customer service reps are dinged for correcting mistakes that AI made, or for being less than enthusiastic about being forced to train the virtual agents that may one day replace them? Why are the human agents being blamed for falling CSAT scores or declining customer loyalty when it’s really the fault of poor AI implementation?

The problem is clear: these reality check conversations insist that executives must be patient with AI transformation, but not with agents.

Agents as Scapegoats for Failed AI Strategy

This validates the idea that employees unable to instantly “harness the power of AI” to quadruple their productivity and make every customer happy either have an anti-AI agenda or are simply incompetent. But they ignore the responsibility that leaders have to develop a strategy for agent upskilling in the AI era–one that is led by the agents themselves.

Sure, some of these “AI Reality Checks” discuss the importance of employee upskilling, but no clear and realistic employee upskilling roadmaps or timelines are ever included. It calls to mind the noticeable lack of AI implementation guides from vendors during the “Mass and Immediate AI Adoption” phase that defined 2025 – and look where that got us.

CX leaders must tell enterprises and executives the whole truth about AI adoption: agents bearing the operational and emotional cost of half-baked AI deployments while leadership figures out their "transformation narrative." We have a responsibility to insist that agents are included in conversations about AI in customer support because their perspective is so valuable: they’re the most directly impacted by AI technology and have the most hands-on experience with it. Agents know what is and isn’t working better than the boardroom ever could. They don’t need intelligent analytics to tell them what’s broken about the customer experience - they see it every day.

If leadership genuinely listened to and acted on agent feedback about their latest AI tools, perhaps their expectations of what AI is actually capable of doing right now, not in 5 years, would have been realistic from the start.

Maybe executives could have started their AI journey with a strong change management strategy and tech infrastructure already in place, instead of having to work backwards or restart from scratch now.

Why Agents Are Really Left Out of AI Conversations

Making room for the frontline agent perspective seems so obvious, but even we as CX analysts are guilty of leaving them out. Yes, part of the reason is because, for the first time in a long time, executives themselves are an active part of technology discussions and decision-making. It’s a natural impulse to cater to executive expectations exclusively - I’ve certainly done it.

But perpetuating the “It’s not the technology, it’s the people behind it” narrative is unhelpful at best and harmful at worst. The truth is that agents aren’t included in these conversations for two main reasons. One, because frontline agents are the perfect scapegoat for leadership failures (as the rampant rise in AI Washing tells us.) Two, because many executives still plan to get rid of them as soon as the AI technology is “good enough” to replace them.

Unfortunately, it turns out that AI technology is taking much longer to get “good enough” to replace agents than a lot of executives – and investors – expected.

Many companies – Klarna, Salesforce, and Duolingo among them – went all in on mass layoffs before their AI agents had even partially been deployed successfully. And while some of those companies have rehired human agents after getting predictably disastrous results from AI-only customer support, 2026’s obsession with “AI-First CX” isn’t exactly comforting.

The truth is that human agents aren't going anywhere anytime soon, largely because the AI that was supposed to replace them by now can't handle the complexity, nuance, and emotional labor that customer service actually requires. 

Companies that treated agents as disposable while betting everything on AI are now scrambling to rehire. However, the ones that invested in their people, while thoughtfully integrating AI, are seeing better outcomes for everyone: customers, agents, and yes, even the bottom line.

What Do Real “AI Reality Checks” Look Like?

We’ve arrived at a moment where the gap between what AI was supposed to do and what it's actually doing has become impossible to ignore. While the "AI Reality Checks" are finally happening, they're still incomplete. 

More holistic AI Reality Checks for contact centers must take the agent experience into consideration just as much as the customer experience. They should include: 

• Pausing new AI tool rollouts until agents are actually trained on existing ones
• Adjusting performance metrics to account for the extra work AI creates
• Differentiating employee monitoring from employee surveillance
• Creating feedback loops for agents, then actually acting on that feedback
• Extending the same patience and long-term thinking to human workers that executives extend to AI investments
• Being honest about whether the goal is to augment agents or eliminate them
• Developing a continuous, agent-led employee training and upskilling strategy

The Real CX Conversation We Need To Have

CX leaders and analysts like myself have an uncomfortable, but necessary task ahead of us in 2026: 

We need to advocate - loudly - for the people who have the least power in these conversations but the most insight into what's actually happening. 

That means pushing back when executives want to blame "change resistance" for failed AI deployments and instead telling them the truths they don’t want to hear. It means insisting that agent feedback isn't just collected, but acted on. It also means being able to admit fault about the parts of our “AI for CX” predictions that were incorrect or unhelpful. 

And for executives: if you genuinely believe AI is the future of customer service, then invest in the humans who are building that future alongside the machines. Not as a PR move, or as a temporary measure until the tech improves enough to cut headcount - but as a fundamental operating principle.

We have a choice. We can keep having AI Reality Checks that comfort executives while frontline workers absorb the chaos. Or we can have the reality check that actually matters: the one that centers the humans who never stopped doing the real work of customer service, even while being told they were obsolete.

The question is whether we're brave enough to have that conversation.