‘AI And The Illusion of Productivity’ by Lara Khouri
In customer experience and organisational transformation, AI is often positioned as a straightforward accelerator of productivity.
The narrative is compelling: automate tasks, reduce effort and unlock faster outcomes at scale. Yet, as many CX leaders are discovering, the reality of implementation is far more nuanced. Without the right data, preparation, and human support, AI adoption can just as easily slow things down as speed them up.
In this insightful piece, Women in CX Inner Circle member, award-winning people-centric leadership strategist, and CX transformation expert Lara Khouri explores the gap between the promise of AI-driven productivity and the lived reality of implementation.
Drawing on her work as “The CXO Maker”, she examines how poorly supported adoption can create “negative productivity” and unintended organisational friction. She challenges leaders to move beyond hype and focus on embedding AI in ways that truly support people, improve workflows, and deliver sustainable value.
“Just do AI” seems to be the chorus everyone is singing these days.
However, as all CX professionals worth their salt know, “just doing” something without the right data and the human context is a liability and can be a very costly mistake.
In simple business terms: it’s a big risk.
Research from MIT Sloan shows that organisations often experience an initial drop in productivity after introducing AI, which could be due to one of two main reasons: either AI has caused disruptions that people were not prepared for, or people had to take the time to be trained on how to work with AI.
In both cases, workflows that may have been running for years require redesigning. In both cases, there is a time lapse between adoption and return on investment. And, in both cases, because the way most businesses are run is broken, organisations are focused on short-term value creation in the form of increased revenue or decreased costs (eg: quarterly earnings, annual profits) and do not give the investment the time it needs to mature into new standard operating procedures that can impact the bottom line.
So the question isn’t whether AI, or anything new that’s introduced into the workspace and workflow, can improve productivity; it’s whether the people who are working alongside AI are being appropriately supported so that productivity does increase.
“So the question isn’t whether AI, or anything new that’s introduced into the workspace and workflow, can improve productivity; it’s whether the people who are working alongside AI are being appropriately supported so that productivity does increase.”
— Lara Khouri
Perception, Reality, And The Cost Of “Negative Productivity”
On paper, it’s quite straightforward: AI should perform the repetitive tasks, accelerate response times, and, consequently, increase the amount of work that gets done (also, AI doesn’t take lunch or bathroom breaks).
That’s the perception.
However, to paraphrase a line from my website: [leadership’s perception] may not be [their staff’s or their customers’] reality.
In fact, a Deloitte survey from late last year found that while 21% of C-suite respondents believe GenAI is transforming their organisation (presumably positively), only 8% of non-C-suite survey respondents share that belief.
The reality is that because of the way in which AI is being implemented, it contributes to what Forrester calls “negative productivity” – which can eat up to 30% of a person’s time and, in the case of AI, can include correcting mistakes, dealing with customer or colleague fallout, reworking flawed approaches or outputs, and simply struggling to use the technology appropriately.
I have seen this tech struggle first-hand.
In one of my corporate roles, I was involved in a major transformation initiative that promised efficiency gains on a global scale through technology upgrading and adoption. After years of effort by dozens of people and millions of pounds, it was abandoned because the human context – both staff and customer – hadn’t been considered until the technology was being rolled out and started creating roadblocks and hurdles for people.
Let me put a Pound figure to “negative productivity”:
A year has approximately 2,080 working hours. Say someone earning £50,000 a year is spending 10% of their time fixing a mistake or problem that didn’t exist before AI was implemented (I won’t use the full 30% because “negative productivity” isn’t caused by AI alone), that’s 208 hours and £5,000 of “negative productivity” focused on continuously second-guessing the technology to make sure that it’s right, fixing the output if the tech isn’t right, or putting in extra effort to plug any gaps in the output.
Without the necessary support before and during the implementation of AI, how can this situation not result in decreased productivity? Not to mention decreased motivation and increased stress.
And that’s just for one person. Imagine if you had an entire department doing that.
Or every department in every location, organisation-wide.
The numbers add up very quickly and they don’t paint a pretty picture; especially when we consider that, when the numbers come in at the end of the year and the C-suite are looking at the AI implementation’s price tag and comparing it to the expected KPIs, they’re likely to be disappointed and, in order to manage the fallout of the ill-informed decision, will look at cutting costs – which usually means cutting jobs.
Aim To Enhance, Not To Replace
Of course, there are many examples of successful AI implementation. Think of the digital assistants that come as standard with so much of our tech, or rideshare and navigation apps.
There are more creative implementations as well. Duolingo integrated GitHub Copilot into its engineering workflow to support its 300+ developers by providing real-time code suggestions and automating repetitive tasks. Alibaba, the world’s largest e-commerce platform, integrated AI into its daily operations so it could access the data necessary to predict consumer purchases. Mattel, of Barbie and Ken fame, supports its product designers with OpenAI’s DALL-E system, which creates realistic images and art based on natural language inputs.
What these examples all have in common is that AI was implemented thoughtfully into the workflow so that it could enhance human effort, which then increased productivity.
The beauty is that it doesn’t have to be an expensive or long process to meaningfully integrate AI into existing workflows.
I’ll share another example from my corporate life to illustrate.
As the non-HR recruitment lead for a very well-known international organisation that had a great reputation and loyal staff, it was my responsibility to go through the hundreds of CVs we received for the rare staff vacancies.
With a day job to do and so many CVs to go through, I simply didn’t have the time to get back to anyone who wasn’t shortlisted. And this killed me because I knew that each CV represented someone’s hope for the opportunity to grow, to provide for their family, to get one step closer to achieving their career goals. While we did say in the job advert that only shortlisted candidates would be contacted, it didn’t help me feel any better about it.
Although this example is going back over 10 years and we didn’t have the tech to help with the sorting back then, today, using tech to help with the shortlisting means that I could probably make the time to write a short email thanking the people who didn’t get shortlisted for their interest, complementing them on their skills, encouraging them to apply for any role they find interesting, and maybe even giving them tips on how to improve their CVs or cover letters or letting them know why they weren’t shortlisted.
That’s showing care.
That’s human.
And that’s what contributes to brand equity and brand positioning in a competitive global talent space.
Productivity isn’t always about doing more by either doing things quickly or reducing costs. Sometimes, productivity is about doing things better.
I recently read an article about how AI is being used in the recruitment process and, a few years ago, I’d mentored a young job seeker who experienced this use of AI. This is the human and brand perception cost of “just [doing] AI”. It’s … not good.
“Productivity isn’t always about doing more by either doing things quickly or reducing costs. Sometimes, productivity is about doing things better.”
— Lara Khouri
Reframe “Just Do AI”
Can AI increase productivity? Yes. Can AI improve people and organisational performance? Yes.
When implementation is properly thought-out based on having the correct data and information, yes, AI can do those things.
However, when that’s not the case, implementing AI can increase “negative productivity”, which decreases productivity and performance, and increases hidden, and not-so-hidden, costs.
So, please, to every CX or EX leader being told to "just do AI", find the information (it’s there!) that will allow you to shift the focus to building hybrid, effective, human-centric, sustainable business operations.
Keep the Conversation Going
Lara’s insights highlight a critical reality for CX and EX leaders: AI is only as effective as the human systems, support, and context around it. Productivity isn’t guaranteed by adoption – it’s shaped by how thoughtfully we design for people, process, and practice.
Inside the WiCX Inner Circle, we create space for exactly these kinds of conversations.
From member-led discussions and expert sessions to webinars and live podcasts, our community brings together women shaping the future of customer experience, leadership, and transformation.
If you’d like to keep exploring how to implement AI responsibly, reduce “negative productivity,” and build truly human-centric ways of working, you can join the WiCX Inner Circle waitlist today.