What do CX leaders say about AI deflection in customer service?

CX leaders converge on two positions. The first, from Jeff Toister of Toister Performance Solutions, says AI deflection works inside a specific scope he calls CRaP: Confident, Routine, Predictable. Outside that scope, route to a human. The second, from Annette Franz of CX Journey Inc., says deflection is the wrong unit of measurement; the right one is whether AI absorbs complexity for the customer or adds new complexity around the customer. Both leaders flag closed-loop AI with no escalation path as the dominant failure mode.

What is the expert consensus on agent augmentation versus replacement?

The consensus across named CX leaders has shifted from agent replacement to agent augmentation since late 2025, driven by burnout data. Jeff Toister cites that 59 percent of contact center agents are at risk of burnout and empowerment is the biggest factor that lowers risk. AI deployed well augments empowerment by handling routine cases. AI deployed for adherence grading removes empowerment and accelerates turnover. One cited company saw an 85 percent reduction in abandoned calls plus a 50 percent improvement in agent retention by using AI for simple transactions.

What do CX leaders say about vendor consolidation in customer experience tooling?

The named platforms (Zendesk, Intercom, Freshdesk, HubSpot, Salesforce, Help Scout) are competing on AI capability, not feature parity. Annette Franz suggests filtering vendors with one question: what becomes more human because of this investment? Bruce Temkin has argued the consolidation cycle follows a predictable shape: fragmented market consolidates around an AI capability, produces a temporary monopoly, the monopoly produces a backlash, a new wave of unbundled best-of-breed tools follows. The risk in 2026 is that AI bundles include a weak knowledge base layer that becomes the closed-loop failure mode.

How are CX leaders changing hiring in 2026 because of AI?

The role change that CX leaders are talking about is not a chatbot manager. It is a partnership. Annette Franz argues the most strategically important partnership a CX leader can build is with the CHRO, because AI changes the agent role, the role change requires a hiring profile change, and the profile change requires HR. The failure pattern is embedded everywhere, owned nowhere: CX functions distributed across product, ops, and support without a single accountable owner. Easy decisions get made and hard ones get punted.

What do CX leaders say AI cannot do for customer service?

The strongest cross-leader consensus, appearing in 11 of 23 verbatim quotes, is that AI inherits the quality of the organization behind it. Annette Franz says AI is making it easier for companies to hide systemic problems instead of fixing them. Jeff Toister says AI is just a tool, shaped by the culture, processes, and training of the people using it. AI cannot fix a broken culture, a missing escalation path, a stale knowledge base, or a closed feedback loop. It scales whatever is there, good or bad.

CX Leaders on AI: 17 Sourced Quotes from 2026

This article collects what customer experience and customer success leaders are actually saying about AI, on the record, in named and linked sources. Every quote below carries a reference link directly beneath it. Nothing here is paraphrased into an anonymous "experts say" line. If a leader is quoted, you can click through and read the source.

We rebuilt this piece for one reason: an honest quote roundup has to be honest about its quotes. The version you are reading covers 17 named leaders. Fifteen are external, drawn from published 2026 industry sources. Two are from HappySupport's own AI in CS interview series, quoted from our transcripts. That is the real number, and it is more useful than an inflated one.

How we sourced these quotes

Three rules governed this roundup. First, every external quote comes from a named, published, linked source: a 2026 trends report, a vendor newsroom, or an on-the-record industry article. The source sits beneath the quote so you can verify it. Second, the two HappySupport interview quotes are taken from our own recorded conversations with Jeff Toister and Annette Franz, linked to the full interviews. Third, where we could not verify a quote against a live source, we left it out rather than guess. That is why the count is 17 and not 30.

The external leaders span customer experience, customer support, and customer success across the SaaS industry, from platform CEOs to independent CS consultants. The synthesis at the end shows where they agree and where they split.

On AI deflection: what works, what fails

The leaders who have shipped AI into live support are the least starry-eyed about it. The recurring theme: buying AI is trivial, making it work is not.

"Launching AI is easy, but transforming with it is not."
Declan Ivory, VP of Customer Support at Intercom

Source: Intercom Customer Service Transformation Report

"The most successful customer-facing AI focuses on automating CRaP: Confident, Routine, Predictable."
Jeff Toister, Toister Performance Solutions

Source: HappySupport AI in CS interview

"The next frontier for AI in customer success is operationalizing it as a front-facing, embedded coach that delivers hyper-contextual adoption pathways."
Naomi Aiken, Techtonic Lift

Source: ChurnZero 2026 Customer Success Trends

On agent augmentation and the role of humans

Nobody serious argues AI removes the human. The argument is about what the human is freed up to do.

"AI-only is a mistake. The best experiences always give customers a way to get to a human agent."
Shep Hyken, Customer Service and CX Expert, bestselling author

Source: NiCE: AI-first CX trends for 2026

"AI should absorb complexity for the customer, not create new complexity around the customer."
Annette Franz, Founder and CEO of CX Journey Inc.

Source: HappySupport AI in CS interview

"By the end of 2026, the average CSM will have 25 to 50 percent more bandwidth."
You Mon Tsang, CEO of ChurnZero

Source: ChurnZero 2026 Customer Success Trends

"AI will not replace CSMs, but it will replace the teams that never learn how to use it."
Anika Zubair, Founder of The Customer Success Pro

Source: ChurnZero 2026 Customer Success Trends

"The job of being a project manager will rapidly fade as AI takes on the coordination work."
Abby Hammer, Chief Customer and Product Officer at ChurnZero

Source: ChurnZero 2026 Customer Success Trends

On hiring and team structure

If AI takes the routine work, the hiring profile changes. Several leaders said the same thing in different words: hire for judgment, not task execution.

"CS leaders will hire less for task execution and more for decision-making under pressure."
Swati Garg, Founder of Melo Associates

Source: ChurnZero 2026 Customer Success Trends

"Leaders will intentionally redefine people's roles around judgment, empathy, and strategic decision-making, elevating employee experience while avoiding over-automation."
De'Edra Williams, Customer Success Revenue Frontier

Source: ChurnZero 2026 Customer Success Trends

"New role titles like CS AI Analyst, Customer Intelligence Lead, and Customer Signal Architect will start to differentiate teams."
Marley Wagner, EverHealth

Source: ChurnZero 2026 Customer Success Trends

"AI fluency will become a leadership requirement in 2026."
Rod Cherkas, Post-sale Growth Consultant and author

Source: ChurnZero 2026 Customer Success Trends

On measurement and accountability

AI changes what gets measured, not just who does the measuring.

"AI will evolve from a reactive data analyst to a real-time interpreter of customer value."
Ejieme Eromosele, Quiq

Source: ChurnZero 2026 Customer Success Trends

"AI models will analyze historical churn to forecast revenue risk and upside months in advance."
Emma Lo, Deepgram

Source: ChurnZero 2026 Customer Success Trends

On vendor consolidation and tool selection

When every vendor ships AI, the AI stops being the reason to buy.

"AI is not the differentiator anymore. How intelligently you apply it is."
Tom Eggemeier, CEO of Zendesk

Source: Zendesk CX Trends 2026

"AI-native companies increasingly push toward consumption-based pricing models."
John Gleeson, Success Venture Partners

Source: ChurnZero 2026 Customer Success Trends

On predictions for 2027 and beyond

The structural prediction most leaders converge on: the org chart itself gets redrawn around AI.

"Leadership will realign customer success orgs around assistive and agentic AI."
David Ellin, Partner at Winning by Design

Source: ChurnZero 2026 Customer Success Trends

The synthesis: what 17 named CX and CS leaders agree on

Read across all 17, and the agreement is sharper than the disagreement. Three points recur.

First, AI removes routine work, not humans. Every leader who addressed the human question, from Shep Hyken to Annette Franz to Anika Zubair, drew the same line: AI takes the predictable load, the human handles judgment, empathy, and the hard cases. Nobody credible predicted a human-free support org.

Second, the hard part is not the AI, it is the operating model around it. Declan Ivory's line, that launching AI is easy and transforming with it is not, is the most-echoed idea in the set. You Mon Tsang quantifies the upside (25 to 50 percent more CSM bandwidth), but Swati Garg, De'Edra Williams, and Rod Cherkas all point at the same precondition: the team, the hiring profile, and the leadership skill set have to change first.

Third, the differentiator moves up a layer. Tom Eggemeier says it plainly: AI itself is not the edge anymore, applying it intelligently is. Jeff Toister's CRaP framing and Naomi Aiken's embedded-coach prediction are both versions of the same point: the value is in what you point the AI at, and how current the knowledge underneath it stays.

Where they split: pace. The ChurnZero contributors lean optimistic on 2026 timelines. The leaders running live support at scale, Declan Ivory most clearly, are more measured about how long real transformation takes. Both positions are in this roundup on purpose.

How HappySupport approaches the gap CX leaders flag

One gap runs underneath almost every quote above. AI removes routine work only when the knowledge it draws from is correct. Jeff Toister's CRaP framing assumes the Confident, Routine, Predictable answers are documented accurately. Declan Ivory's point about transformation being hard is, in large part, a knowledge-base problem: the AI is only as good as the help center it reads. Annette Franz's warning that AI amplifies organizational dysfunction applies directly to documentation that has drifted behind the product.

HappySupport is the help center layer built for that gap. HappyAgent watches the product repository for changes that affect documented user flows and surfaces the affected articles before customers hit the stale answer. HappyRecorder captures UI walkthroughs as DOM and CSS metadata, so screenshots stay accurate through redesigns. The result is a knowledge base that stays current as the product ships, which is the precondition every leader in this roundup is implicitly describing. More on the architecture at how a self-updating help center works and on the underlying problem at the hidden cost of documentation decay.

HappySupport sits beside whichever ticketing system and AI agent you already run. Keep Intercom, Zendesk, or your AI vendor of choice. Swap in HappySupport for the article layer that the AI reads from, so the routine answers it deflects are the right ones.

Discover HappySupport