If you’ve skimmed McKinsey’s State of AI 2025, you’ve probably seen the headline: almost every company says they’re using AI, but only a handful are actually getting value out of it. Everyone’s experimenting. Few are transforming.
That gap is especially clear in customer experience. Contact centers, service teams, and support organizations are full of pilot projects, half-baked automations, and “AI-powered” tools that never quite deliver. And if you’re running a CX or service team at a smaller or mid-size enterprise, you’ve probably wondered if this AI wave is something you can realistically ride, or if it’s just for the billion-dollar players.
The truth is you don’t need a massive budget or a dedicated data science team to win with AI. You just need focus, structure, and a clear plan. Let’s talk about what that means in plain language.
The Reality Check
McKinsey found that 88 percent of organizations use AI somewhere in their business. Sounds great until you realize two-thirds of them haven’t moved past pilots. AI is in everyone’s vocabulary but not yet in everyone’s workflow.
Most of the activity is happening in IT, knowledge management, and customer service. This is our world. It’s where AI can actually make a difference for both the customer and the agent.
So why are results still mixed? Because too many companies treat AI like a side project instead of a core part of how they work.
What This Means for CX Leaders
If you lead a contact center or CX team, here’s what the research really means for you.
1. Don’t start with the tool, start with the problem.
Every week there’s a new AI product that promises the world, but if you don’t know what problem you’re solving, it’s just another login screen. Focus on the friction first, then use AI where it clears real roadblocks. Are handle times creeping up? Are customers bouncing between queues? Are agents wasting time searching for answers? That’s where AI can help right now.
2. Fix your workflows before you add tech.
If your process is broken, AI will only break it faster. Take a breath and trace how work really moves across your people, your systems, and your channels. You’ll find the bottlenecks pretty quickly. Clean up what’s messy, streamline what’s clunky, and make a few simple process tweaks before layering in any new tools. Then use AI where it actually reduces effort or improves accuracy.
As I’ve said many times, you cannot innovate your way out of bad business process.
3. Focus on growth, not just cutting costs.
Too many companies treat AI like a cost-saver instead of a growth enabler. The ones seeing real upside are using it to make experiences better by predicting churn, personalizing outreach, or giving agents smarter insights in the moment. When customers feel the difference, the savings follow naturally. Lead with value creation and efficiency will take care of itself.
Why Smaller Teams Struggle and How to Beat the Odds
Let’s be honest, smaller CX organizations don’t have the same resources as the big guys. You’re probably fighting with old tech, limited budgets, and a dozen priorities competing for attention. McKinsey’s data backs it up. Larger enterprises scale faster because they have people and capital to throw at AI.
But smaller companies have something big enterprises don’t: speed. You can move faster, test ideas without ten approvals, and actually see the impact up close. The key is to stay practical and disciplined.
Here’s what usually holds teams back, and how to move past it:
- Your data is scattered. Calls live in one system, chats in another, and surveys in a spreadsheet. Start by connecting what you can. You don’t need a perfect data warehouse to get started. Even one clean connection can unlock new insights.
- You don’t have AI experts. That’s fine. You need context, not code. Bring in partners who understand CX operations, not just algorithms.
- You trust vendors too much. Don’t buy every “AI add-on” you’re pitched. Build your roadmap first, then choose technology that supports it.
- You’re stuck in pilot mode. Pick one area, measure it, and scale it when it works. You’ll learn more from one full rollout than from five half-finished tests.
Making AI Work in the Real World
AI doesn’t need to be complex to be effective. It just needs to work where your people work.
Start small but measure everything.
If you can’t tie a project to a real metric like CSAT, handle time, or first-contact resolution, don’t do it.
The more seamless the experience, the better.
Put AI inside the tools your teams already use such as CRM, WEM, WFM, or ticketing. If they have to jump between screens or open another tab, adoption will drop fast. The less your agents have to think about using AI, the more they’ll actually use it.
Train and involve your people early.
Agents need to understand what AI is doing and how it helps. When they see it making their jobs easier, adoption takes care of itself.
Build trust before you scale.
More than half the companies McKinsey surveyed have already run into AI issues such as wrong answers, biased results, or compliance headaches. Don’t rush. Validate first, then scale.
How to Build Trust Without Slowing Down
You don’t need an AI ethics committee to be responsible. You just need a few simple rules that keep everyone safe and confident.
Try this:
- Keep humans in the loop and set checkpoints where supervisors or agents verify AI outputs before they go live.
- Decide what data you’ll use. Know which customer data can leave your system and which can’t. Write it down and stick to it.
- Test accuracy with real call data and transcripts in a small pilot before rolling out.
- Show your work. If AI scores a call or recommends an action, make it visible why. Transparency builds trust.
High performers don’t avoid risk, they plan for it. Smaller CX organizations can do the same and move just as confidently, even without big-company budgets.
Where a Partner Adds Value
Most smaller teams don’t have time to design frameworks or monitor AI models. That’s where a CX services partner helps, not by selling more software but by helping you make sense of what you already have.
A good partner will:
- Help you pinpoint where AI fits and where it doesn’t.
- Integrate it with your existing tools so your team actually uses it.
- Set up guardrails for privacy and accuracy so you stay compliant and confident.
- Train your staff and track outcomes so adoption sticks.
- Monitor and tune performance as things evolve.
The right partner doesn’t replace your team. They help your team get better.
The Bottom Line
McKinsey’s message is simple: AI potential is huge, but execution is hard. That’s the part most people miss.
Smaller and mid-size CX organizations can absolutely lead the way if they focus on the fundamentals. You already have the advantage of speed, agility, and proximity to your customers. Use it.
Start small. Fix your workflows. Build trust. Measure impact. Then scale what works.
AI success isn’t about how much data you have or how big your company is. It’s about how well you use what you’ve got. The winners won’t be the ones who buy the most AI, they’ll be the ones who operationalize it and make it part of everyday work.
Quick Takeaways
- Most companies are stuck in pilot mode. Scaling is where the value is.
- You don’t need a big budget to succeed, just discipline and focus.
- Fix broken workflows before adding tech.
- Trust beats hype. Build it through transparency and validation.
- AI done right isn’t about tools, it’s about making work and experiences better.
(Source: McKinsey & Company – The State of AI in 2025)
