If you believed the hype around AI in call centers, live agents were supposed to be gone by now. Automation was going to absorb nearly every interaction. The cost of support was going to fall dramatically almost overnight, and outsourcing would be less of a factor.

This didn’t happen.

While we are all glad the hyperbole has toned down, the narrative keeps changing but the temptation remains the same. Many organizations are already making decisions as if the outcome of AI and customer experience (CX) in call centers is certain.

That is the point of this article. It is not about whether AI matters. It does. It is not about whether AI will improve customer experience. It already is.

The concern is that AI is beginning to distort how some brands evaluate outsourcing decisions. In too many cases, AI is not improving decision-making. It is making it less disciplined.

AI should be part of the outsourcing conversation. It should not replace the outsourcing conversation.

Perception vs. Reality

In recent interactions with CX, BPO and sourcing professionals, I noticed a strange phenomenon. AI is no longer just a technology discussion. It has become a perception issue.

Some leaders appear to be planning as if the future of customer support is already settled. They assume AI will rapidly reduce headcount, absorb complexity, and change the economics of CX so dramatically that today’s operational problems may not matter much longer.

That belief is dangerous.

We are still at the starting line of a marathon, but some organizations are behaving as if this is a sprint. The result is unrealistic expectations, misplaced priorities, diverted resources, and in some cases, impetuous decisions.

AI’s positive impact is real, and the full effect of AI on CX will take time to understand. AI can improve quality monitoring, agent assist, knowledge management, forecasting, coaching, self-service, sentiment analysis, and workflow automation. It can help brands and BPOs become more efficient, more consistent, and more responsive.

But AI has also produced highly visible failures. Poorly designed automation can frustrate customers, damage trust, increase repeat contacts, escalate costs, and hurt brand equity. The problem is not AI itself. The problem is assuming AI is a shortcut around the hard work of customer experience management.

AI Cannot Fix a Broken Operation by Itself

Putting aside the importance and ubiquity of AI, many brands and BPOs still have immediate “call center 101” problems that need attention now.

These include missing service levels, workforce management struggles, scale issues, inconsistent quality and performance, insufficient training and coaching, inexperienced frontline leadership, high attrition, unclear escalation processes, incomplete reporting, innovation gaps, reactive vs. proactive approaches and much more.

If the brand’s internal teams and BPO partners aren’t maximizing the customer experience consistently, then does it make sense to be preoccupied with AI in areas where AI alone can’t fix the problem(s)?

That is backwards.

If a brand cannot clearly define the problem that it needs a BPO to solve, then how will AI? Technology can accelerate a strong operation. It can also expose or magnify a weak one.

That is why the fundamentals still matter in outsourcing strategy, partner selection, geographic footprint, cultural alignment, governance, training, workforce management, quality assurance, and operational execution. Some of these will be influenced by technology, but most will still be determined by people, process, leadership, and discipline.

AI does not eliminate the need to get the basics right. It makes getting the basics right even more important.

AI should be part of the outsourcing conversation. It should not replace the outsourcing conversation.

Outsourcing Is Still Favored Over Insourcing

With or without the AI conversation, outsourcing remains a viable and robust strategy for companies of all sizes. I do not see that changing anytime soon.

I will not revisit every benefit of outsourcing here because most CX leaders already understand them (and you can refer to my previous articles on the subject). Demand for outsourcing remains strong because it is a strategic growth tool, no longer viewed only as a cost-reduction lever.

Hence why many brands are looking to outsource higher-value and white glove customer interactions along with HR, finance, IT, quality as a service, WFM as a service, and other back-office functions.

The global BPO industry also continues to grow. Grand View Research projects the market will expand from an estimated $328 billion in 2025 to $695 billion by 2033.

That growth is fueled by rising internal operating costs, more brands embracing outsourcing, soaring demand for nearshore and offshore, and the expansion of outsourcing into more complex business functions

Most brands that utilize a balanced approach in partnering with BPOs are still seeking the same core outcomes: the right strategy, the right partners, the right locations, strong operational delivery, and increasingly, a technology-forward approach.

The best organizations, however, are not allowing AI perception to dictate the entire sourcing, BPO selection and management process. They are evaluating AI as part of the equation, not as the equation.

The Old Selection Problem Has New AI Layers

After decades in the CX and BPO industry, I have seen nearly every version of the “vendor” selection process. Some brands get it right. They define clear expectations. They invite the right BPOs into the process. Experienced CX leaders guide the decision. The scoring process is fair. The operating model is realistic. The final decision is based on capability, alignment, and evidence.

When that happens, brands and BPOs have a much better chance of building the kind of partnership both sides want.

Other brands continue to make bewildering decisions, suffer the consequences, and then repeat the same pattern in the next sourcing cycle.

The industry has improved in many ways. The BPO selection process itself, however, is still too often flawed.

And AI is making that flawed process even more complicated.

Now, many decision committees also include AI experts. That is not inherently a problem. AI expertise is valuable. If the AI expectations are unclear, speculative, or disconnected from the brand’s current CX challenges, the process can quickly lose focus.

A tech-forward solution should absolutely be part of BPO selection. Any BPO that fails to invest in AI, automation, analytics, and modern CX tools will be left behind. But should AI carry a dominant weighting in an RFP when the brand’s stated priority is to fix core operational problems?

I do not think so.

Do Not Let Shine and Gloss Replace Call Center 101

As I stated, brands are treating AI as if it can solve problems they have not clearly defined. They are asking BPOs to explain their most advanced technology strategies before addressing basic operational issues. They are underweighting the fundamentals that still determine whether an outsourcing relationship succeeds or fails

I was involved in a recent RFP for a large brand that serves as a case study in how not to select a new BPO. We were told the client wanted a BPO to get the basics right: meeting SLAs, improving workforce planning, stabilizing performance, and consistently hitting key metrics.

That is a normal and reasonable objective.

But the RFP itself was vague and poorly written. The decision process shifted midstream. Members of the decision team seemed more enamored with shine and gloss than with the core operational solutions they originally said they needed.

This is not an isolated example. A brand says it wants execution, discipline, stability, and performance improvement, but then rewards the vendor with the most exciting AI narrative, the flashiest demo, or the boldest future-state promise.

That is not strategy. That is distraction.

"Technology can accelerate a strong operation. It can also expose or magnify a weak one."

Before giving AI or other lower priority considerations a disproportionate influence in outsourcing decisions and partner selection, brands should answer several basic questions:

• Who are we and what do we want?

• What problem are we trying to solve right now?

• What is broken in our current operation?

• What must improve in the first 90, 180, and 365 days?

• Which metrics matter most?

• What workstreams are truly ready for automation or augmentation?

• Where do we still need human judgment, empathy, and escalation?

• What does the BPO need to prove operationally before we evaluate its AI roadmap?

• Is our team experienced and knowledgeable enough to make critical BPO decisions?

Those questions bring the conversation back to reality. They also make the AI conversation more useful, because the technology is evaluated against the brand’s actual needs instead of vague future assumptions.

AI Is a Partner, Not a Panacea

Brands and BPOs need to view AI as a partner, not a quick fix, not a wholesale replacement for human capital, and not a panacea for longstanding CX challenges.

AI should help the industry make better decisions. AI should not cause brands and BPOs to lose sight of the problems that need solving right now.

Inconsistent service levels require better planning and workforce strategy. Poor quality scores need better training, coaching, and QA to improve. High attrition must be addressed with better leadership, culture, career pathing, and employee engagement. A broken escalation process does not only need agentic AI. It needs clarity.

The strongest outsourcing partnerships will be built by brands and BPOs in sympatico. They know exactly who they are, what they want, and expectations are reasonable and clear.

The future of AI in CX is exciting. It is also uncertain. We can predict. We can prognosticate. We can make educated assumptions. But we should be honest enough to admit what we do not yet know.

What we do know is this: outsourcing decisions still require operational discipline, experienced leadership, clear priorities, and a grounded understanding of what customers and agents need today.

AI belongs in that conversation. It just should not be allowed to dominate it.

Tech-forward should never mean fundamentals-backward.