Why “Just Doing AI” Is Becoming a Liability
Edited by Clare Muscutt, Founder and CEO, Women in CX
Across the Women in CX community, one theme continues to surface in conversations around AI transformation: many organisations are moving faster to implement AI than they are to redesign the experiences surrounding it.
From customer operations and service design to governance, automation and employee experience, leaders across the community are describing growing pressure to “just do AI” — often without the organisational readiness, operational clarity or human-centred architecture required to make it successful.
The result is a growing disconnect between what organisations hoped AI would solve and what many teams are now experiencing operationally. Fragmented journeys. Failed pilots. Employees are spending increasing amounts of time correcting automation. Customers are trapped in chatbot loops designed to contain demand rather than resolve problems.
This article brings together perspectives from contributors across the WiCX community exploring some of the counter-intuitive realities emerging from the current wave of AI adoption, and why the future of customer experience may depend less on how quickly organisations deploy AI — and more on how intentionally they design around it.
AI Does Not Fix Broken Experiences. It Amplifies Them.
One of the strongest themes emerging from community discussion is the misconception that automation can somehow repair poor service design by itself. In reality, AI often behaves more like an amplification layer than a correction layer.
If processes are fragmented, journeys unclear, ownership inconsistent or operational friction already exists, automation rarely removes those problems. More often, it scales them faster and more visibly.
As Stacy Dye describes:
“AI is a loudspeaker — it amplifies whatever you build into it. If you skip understanding what ‘good’ looks like in real customer interactions, you’re just guessing at scale.”
Across the community, contributors repeatedly highlighted that many AI initiatives are still being treated primarily as technology deployments rather than experience transformation programmes.
That distinction matters.
Because when organisations automate without first defining what a good customer interaction actually looks and feels like, the outcome may technically become faster — but not necessarily better.
Responses become efficient but emotionally flat. Journeys become automated but disconnected. Small operational issues suddenly become systemic because they are now being delivered at scale.
Several contributors reflected on the same underlying tension: many organisations are not yet automating excellence — they are automating existing organisational complexity.
And speed without structure rarely creates better experiences.
“AI is a loudspeaker — it amplifies whatever you build into it. If you skip understanding what ‘good’ looks like in real customer interactions, you’re just guessing at scale.”
— Stacy Dye
The Hidden Productivity Cost of Poor AI Implementation
Another major concern raised across discussions was the growing operational burden created by rushed or poorly governed AI deployments.
While AI is frequently positioned as an efficiency driver, some leaders are now seeing the opposite effect internally: employees spending increasing amounts of time correcting AI-generated outputs, managing hallucinations, resolving customer confusion or repairing failed automated interactions.
Lara Khouri describes this as “negative productivity.”
Forrester estimates that employees may now spend approximately 30% of their time on corrective activities linked to AI-generated work. When attached to real operational costs, the scale of that inefficiency becomes difficult to ignore.
An employee earning £50,000 annually, losing 30% of their productivity, equates to roughly 624 hours each year spent fixing problems introduced by the system itself. In larger operational teams, those costs multiply quickly.
But contributors also highlighted a less visible impact emerging beneath the financial one: emotional fatigue.
Customer-facing employees are increasingly being asked to absorb frustration created by systems they neither designed nor fully trust. In some organisations, frontline teams are becoming recovery layers for automation failures rather than being empowered to deliver meaningful customer value.
Ironically, the people AI was supposed to “free up” are often carrying the operational burden of making it work.
The Industry’s Definition of “Efficiency” May Be Too Narrow
Another recurring tension within the discussion centred around how organisations currently define efficiency itself. Across many AI programmes, success metrics still heavily prioritise containment rates, handle time reduction, cost savings and ticket deflection.
But customers rarely experience efficiency through operational dashboards. They experience it through outcomes. A fast interaction that leaves somebody frustrated, confused or unsupported is not necessarily efficient from the customer’s perspective. It is simply unresolved friction delivered more quickly.
Amandine Le Doze described the emergence of “chatbot loops” — experiences where customers become trapped inside automated journeys with no meaningful escalation path available.
In these scenarios, the organisation may technically achieve a lower cost-to-serve while simultaneously damaging trust, increasing effort and eroding confidence in the experience. Several contributors argued that human handoffs should not be viewed as signs of automation failure. Instead, they should be intentionally designed as part of the experience architecture itself.
Because the real opportunity of AI may not be removing humans from experiences entirely, but removing the repetitive, transactional work that prevents humans from contributing where they matter most: empathy, judgement, creativity, reassurance and complex problem-solving.
Research from the MIT Initiative on the Digital Economy supports this thinking, finding that thoughtfully designed human-and-AI collaboration consistently outperforms both human-only and technology-only approaches.
Beyond “Responsible AI”: Toward Regenerative Systems
As the conversation evolved, contributors also challenged whether current industry discussions around responsible AI are ambitious enough. Much of today’s discourse remains focused on minimising harm, mitigating bias or managing risk.
But some members argued the more important question may actually be: what kind of organisational systems are we creating through AI in the first place?
Rose Clarkson introduced the concept of Regenerative AI — a framework inspired by living ecosystems where systems actively create the conditions for resilience, replenishment and long-term flourishing. The metaphor resonated strongly across the discussion. In forests, fallen “nurse logs” continue supporting life long after they collapse, nourishing new growth and strengthening the surrounding ecosystem.
Several contributors reflected on whether organisational systems should be designed with similar intentionality: not simply avoiding harm, but actively creating healthier conditions for employees, customers and communities to thrive.
That shift changes the questions leaders ask during design and implementation.
Not simply:
Did we reduce cost?
Did we increase automation?
Did we improve containment?
But also:
Did this system create more trust?
Did it reduce cognitive overload?
Did it strengthen cooperation?
Did it create space for better human contribution?
Did it support people rather than exhaust them?
As AI adoption accelerates, these questions may become increasingly important indicators of long-term organisational health.
The Leadership Shift AI Actually Requires
A final theme emerging consistently across the community was the idea that successful AI transformation is ultimately less about technology maturity and more about leadership maturity.
Elsa Tranquillo warned against top-down implementation approaches where AI is imposed onto teams without meaningful involvement in shaping how the systems will operate day-to-day.
Contributors repeatedly emphasised that the people closest to the work often hold the clearest understanding of where automation genuinely creates value — and where it risks creating friction.
Tiahna McDowell also argued that every AI capability should operate with three distinct layers of ownership:
a business owner accountable for customer and organisational outcomes
a technical owner responsible for system integrity and performance
and a governance owner responsible for oversight, ethics and risk management.
Without clear accountability structures, organisations risk creating systems that nobody fully owns, and everybody eventually inherits responsibility for when things go wrong.
Across the discussion, one point became increasingly clear:
Responsible AI is not simply a governance exercise. It is an experience design challenge.
“AI capability should operate with three distinct layers of ownership:
a business owner,
a technical owner
and a governance owner”
— Tiahna McDowell
Humanising the Machine
Perhaps the strongest conclusion emerging from the discussion is that the future of customer experience will not be defined by which organisations implemented AI fastest.
It will be shaped by which organisations understand where human value mattered most — and designed around it intentionally. Because customers do not build relationships with automation itself. They build relationships with how organisations make them feel.
When implemented thoughtfully, AI absolutely has the potential to create better experiences. It can reduce friction, remove repetitive work and create more space for meaningful human contribution.
But only when organisations stop treating AI as the strategy itself. AI is infrastructure. Experience is the strategy.
And right now, many organisations may be moving so quickly to automate that they have not fully considered what exactly they are amplifying in the process.
Because if AI acts as a loudspeaker for an organisation… what is it currently saying to customers?
Continue The Conversation At WiCX UnConference 2026
These are exactly the conversations shaping the agenda at the WiCX UnConference 2026.
As organisations navigate rapid AI acceleration, rising customer expectations and increasing operational complexity, the industry is being forced to rethink not just how we implement technology — but how we intentionally design experiences, operating models and leadership around it.
This year’s UnConference theme, Accelerating Responsible Action, explores what responsible, human-centred transformation actually looks like inside modern organisations.
Together, we’ll move beyond surface-level AI conversations to explore:
How organisations turn insight into meaningful action
Why journey orchestration is becoming a leadership challenge, not just a CX challenge
What happens when AI starts making decisions on behalf of organisations
How we create systems that scale trust, not just efficiency
Unlike traditional conferences, the WiCX UnConference is intentionally designed as a curated, practitioner-led experience built around honest conversation, collaborative problem-solving and peer-led discussion.
No sales pitches. No passive PowerPoint presentations. No performative thought leadership.
Just the real conversations leaders need to have now.
📍 London: 14–15 July 2026
📍 Dallas: 20–21 October 2026
Join the waitlist to hear announcements first, access early ticket releases, and help shape the future of customer experience alongside the women leading it.
This article brings together perspectives from contributors across the Women in CX community exploring the operational, ethical and human realities shaping AI adoption inside modern organisations. Edited by Clare Muscutt.