What’s happened to Spotify?

Artificial Intelligence doesn’t know what it wants, and it might be exactly what Customer Experience needs.

What does Artificial Intelligence [AI] mean for Customer Experience [CX] and vice versa?

Historically, during any transformational change, two things happen: first, current work and processes are made more efficient. Second, most of these processes are made redundant as new ones replace them (1).

With the AI-fication of CX we are currently in the first phase: We are solving simple problems by making ineffective instruments slightly more efficient (e.g. research, models or measurement).

This builds on the common misconception that the future is in front of us and will only be a faster or cheaper version of what we are already doing. In reality, most things change — a lot.

The future of customer experience is not directly in front of us; it’s outside. It’s the result of merging CX with other fields of expertise in order to see new combinations we didn’t know existed five minutes ago.

This is where CX needs AI, and AI needs CX. Given AI’s preference for ignoring humans and delight at solving simple problems (2)(3). While CX is somewhat stuck manufacturing solutions to problems made in the 1990's.

The future of CX will probably / hopefully be nothing like what it is today.

What does the future hold?

Kevin Kelly argues in his book: The Inevitable (4) that it is impossible to predict the development, uptake and use of future technologies. But we can predict the forces that influence them. E.g. moving from fixed products to always updating ones, making everything much smarter using cheap AI etc.

It is equally impossible to predict the future of CX, but it is possible to discuss what forces have influence and how these might affect CX’s future.

Let us look at the following forces to illustrate how CX might change with AI:

1. Diverse and ethical practice
2. Remembering the human in the age of the algorithm
3. Human-centered design
4. What the work is
5. The outcome economy

1 . Diverse and ethical practice

A technological revolution is not driven by technology, but by a) customer demand and b) business models and practices (5).

As we are moving into an AI’ed world we need to make sure that businesses aren’t using its new superpowers to harvest data and hurt their customers as recent examples shows that it can: extremism (6), genocide (7), excessive consumption (8) and psychological harm (9).

CX can be critical in the support of:

A. Implementing ethical practice
CX can help make sure that the data included in and the power of the predictive models are used to serve customers and users, not hurt them?

B. Incentive models
CX should contribute to an incentive structure motivating what data we use, how we make sense of it and what experiences we are designing to protect and support our customers. We need to be aware of and responsive to the direct and indirect effects of the choices we make.

C. Data diversity and inclusion
AI-training data is known to be heavily biased (10). Lush in misogyni and racism (11). Our models reflect our current biases and sterotypes (12). Left unchecked a data-driven company will only amplify the behaviors and attitudes already inherent in the organization(13).

In other words: a product-centric organization won’t magically become customer centric by using more data. In fact, on its own the only thing the data will do is make the organization more product-centric.

As the organization transforms towards the data, CX needs to help steer the organization in the direction of its customers.

CX’s responsibility is to help our engineers and designers understand what makes a whole, diverse and inclusive population and human being. To use this awareness to influence what is included and prioritized when we are training our talent, models and engines.

2 . Remembering the human in the age of the algorithm

There is a customer data gap (14) in our organizations. We are relying on spurious proxies (13) to make decisions, but these are often unrelated to our customers, they ignore what makes us human (15) and misdirect our teams (16) towards what is easy to measure (channels, products and technologies) instead of what’s important to understand (human motivation and needs)(17).

Where can Customer Experience have influence?

D. Product and channel bias
Customer Experience is the second most important focus area of AI-investments today, but these investments are also producing the 4th most disappointing results (18). One of the reasons might be that a lot of customer-focused work just doesn’t include the customer, but rather uses product or channel data as proxies for the customer (16).

Becoming data driven on its own doesn’t change a company’s (19) mental models (20) or priorities. As an organization becomes data driven unchecked data can reinforce biases, as noted by O’Neill (21) and Blauw (22). Customer experts should support the organization to balance both context-less “thin” data (e.g. behavioral data) and contextual/subjective “thick” data (23).

E. Design for better data
We make the best decisions we can based on the information closest to us (24) and the decisions we make reflect the data we collect.

Let’s use an example: Sophie, an accountant knows what her speciality is, her skill level, her assistants’ skill level, the type of clients she has, the nature of her relationships with them and their businesses. She knows how she earns money and the nature of her network. With all this information Sophie comes to a website looking for answers tailored to her. Now what information does the website collect to understand and serve her experience? Click-through rates, time-on-site, pixel depth and completion rate? All nice metrics, but nowhere is the site doing anything to better understand Sophie, nor explore how to serve her.

Life is much easier when the only data we have to serve our decisions is simple data concerning a controllable resource with limited variation and influence. But making data simple doesn’t make it any better. In fact it will only lead the team to imagine a world filled with problems to solve, where none of them really exists (25).

We get the data, the decisions and the customers we design for.

Humans are slow at decision making. We make upward of 35.000 decisions in a day (26), but compare that to a computer which can make millions of decisions every second.

When humans are making decisions the impact of the wrong data is going to be slow and minimal, but when we accelerate those decisions to a computer any bias or error will be amplified exponentially.

As data quality becomes increasingly important we need to rethink how we engage with our customers in order to produce the best data.

F. Personalization
In 2019 Gartner predicted that 80% of personalization efforts would be abandoned by 2025 (27).

My assumption is that this will happen because these were never personalization efforts in the first place. They were data- and technology led projects that biased companies to focus on simple channel-performance proxies instead of human beings (28).

These so-called personalization efforts in fact made a mockery of our customers, equating them to behavioral patterns we found convenient to use in order to make life less complicated for our content automation engines.

Personalization is another way of saying that we want our customers to identify with our experience. Which means we need to understand and act on their identity which demands that we listen to them as individuals.

Companies are more prone to focus their individualization efforts on their content, not their customers.

Customer Experience needs to take on a proactive role at the forefront of our data and insights processes. Connecting the customer to the business, upskilling and maturing our mental models, questions and sensemaking. With the right tools and practices to help the organization articulate how it wants to show up, partner with and serve their customers’ needs and jobs.

3. AI-to-Human-Centered-Design

Our current design processes are based on the un-scalability of the human designer: a human is slow, expensive, limited, not working continuously nor in perpetuity.

Humans, as we know don’t scale very well on their own, but can scale wonderfully through the right partnerships with technology.

What is the role of AI when it comes to design?

G. Experimentation
It has been suggested that we know about 1% of what is knowable (29). In other words we are just scratching the surface of what is possible. But too often we are content, happy thinking we know all there is to know. Not asking what we don’t know we don’t know (30) which is where most of the available knowledge is.

Experimentation keeps us exploring. It helps us ask questions leading to the discovery of new questions and new ideas. It keeps us surprised, it fuels our curiosity. But manual experimentation has its limitations. It is costly and slow. We can only learn a few things.

With semi-automated experimentation, similar to what you can already find at e.g. booking,com (31), skyscanner.com (32) or Spotify (33) you can automate experimentation where the human sets the rules while the computer probes the alternatives.

With costly experiments an organization starts prioritizing what to learn often heavily biased towards what it already knows. The cheaper it can run its experiments the more it can ask questions it doesn’t know the answer to. It will lean into bets that create more surprises, challenge the status quo and brings an increased opportunity to have a significant impact.

With AI-aided experimentation, a team can greatly accelerate their learning process and ability to discover impactful things we didn’t know, helping us see markets through new perspectives and opening up new solution spaces for us to explore.

H. Complexity
There is a big difference between what is simple and what is complex. According to Dave Snowden (34) these are two very different decision-making domains with different problem-seeking and problem-solving approaches.

How we ask questions and how we answer them needs to be in accordance with the domain where we are looking for solutions.

In a simple domain questions have one answer, people agree and there is a best practice. In a complex domain there are multiple answers, they keep changing all the time and there is no best practice.

Customers are people and people are complex. Most often solving customer problems are done in the complex domain. But since we are so used to the nuts and bolts of our internal operations we tend to pretend that customers are simple and controllable (35) .. which is why only 50% of advertising works (36), companies keep solving problems that don’t exist (25) and innovation has an abysmal success rating (37).

The machine is here to save us. But do we want it to? And are we asking it to do the right job?

I. Value
There are two types of innovation: the type that tries to produce the same value only cheaper, and the type that tries to deliver new value.

We are currently wasting too much of our AI-capabilities on the first one — making simple stuff more efficient (38).

Sure it’s nice and saves peoples time and money, but that was not what they promised us.

More of the same only cheaper to make was not why we got enthusiastic, invested or cared.

Cheaper is not what our customers desire or are willing to pay a premium for (39). It is not what will keep us competitive or profitable. Once you start focusing on your costs more than customer needs and value you have reached the last phase of your business’ life cycle (40).

Customer Experience needs to lead with questions that matter.

The explorative partnership between people who care deeply about the customer and people who are creative and brilliant when it comes to data might be one of the most valuable relationships to nurture in our future organizations.

J. Our tools

“Men have become the tools of their tools” — Walden, Henry David Thoreau (1854) (41)

Tools influence how we see, understand and make decisions about the world. Are we challenging our tools, or too happy just seeing that they produce outputs?

Outputs aren’t necessarily any measure of success, and tools should be scrutinized for what they are not doing.

The slide ruler was a wonderful instrument helping us solve complicated calculations. But it was also very limited. And because of its limitations we made bridges that where to heavy, houses where the walls were to thick and inefficient airplanes (42).

If we choose simple tools we will get simple answers and with a complex customer those answers might ignore what’s most valuable, desirable or relevant to our customers.

K.Automated design systems
Design Systems are the heart of the organizations user and brand experience.

With these systems we can set rules and automate the design and content of our experiences. We can manage complexity with the customer at the center and limit the interruption of any slow human process from the chain of interaction between two computers (43).

The customer will benefit greatly from automation, adaptability and responsiveness. Humans set rules and design models. Computers enhance our ability to design for use we cannot even imagine (44).

l. Real-time experience
People show up to get jobs done in given situations. They are influenced by e.g. where they are (context), who/what they are with (relationships) what they are trying to achieve (progress) and how they measure value (outcomes).

In addition people jump between different contexts all the time where the progress they are trying to achieve, the rituals, habits and technologies connected to them are entirely different.

e.g. when we travel to the same hotel for business or pleasure the needs and measures of value might be entirely different.

Collecting data about one person in one context can be irrelevant insights if we are trying to understand the same person, but in a different context.

Therefore understanding a whole human being might be unnecessarily complex, we only need to understand the context they are in right now in order to serve them.

e.g. as on online retailer I don’t need to know the name of your mother, dogs and high-school. Nor how much money you make, your age, ten closest friends, where you work and how you commute there. I only need to know if right now you are more interested in a coffee maker or a tripod.

As we learn why the presumptive customer is engaging with us (e.g. on our website) we need to figure out what they are trying to achieve as fast as possible and start serving motivation, influence and content to that end.

A human designer can’t do this on their own. They would have to design pre-determined paths based on what they can imagine.

But a human with a computer can collaborate to create an environment and a system of interaction which responds to the customer nuances and whims in real time.

4 — What the work is

Think bigger. Using AI to improve how we work means incremental improvement while the work stays the same. We have to change what the work is!

“The important questions are: ‘why’ and ‘what’ the work is?” — Michelle Gilboy

Asking people to do the same work for the same outcomes measured by the same incentives, but changing the entire portfolio of tools and processes for how they do it only produces frustration and disconnected workers.

If we want to introduce AI to work we need to see the bigger picture. We need to change what the work is. Which means a change in focus from output to outcomes, new incentives and measures of value, different mental models and agreements on what the situation we are operating in is and why we are needed.

k. Meaningful work
Turning human workers into computers started long before the computer was even a thing.

A computer runs on algorithms (rules, instructions and sequences of action) which is what Scientific Management (45) introduced to work more than 100 years ago.

Scientific Management turned workers into machines with limited and repetitive tasks subordinate to a prompt (manager/leadership) with control of our responsibilities and actions.

So when we ask today: why is it so easy to replace workers with machines?

It’s because when it comes to work we’ve been machines for almost a century.

According to Kevin Kelly efficiency (which seems to be one of the premier drivers of modern management) is for robots (46) while Cory Doctorow helps us realize that too many of us are already being treated like prompting engines at work (47).

“They [movie executives] already treat us [writers] like a prompting engine.” Like, make me ET [the 1980’s Spielberg movie], but the hero is a dog. And there’s a car chase in the second act.” And then they come back and they’re like, “Can you add a love interest?” Right? And this is just how you prompt a language model, right?’ — Cory Doctorow

Turning everyone into machines hasn’t made anyone happy: Gallup suggests between 69–87% of employees are not engaged or are actively disengaged at work (48).

Which leads us to ask: is work really that great in the first place? And can we use AI to make it better?

Scientific Management was not the only idea exploring the future of work. Another approach was promoted by Mary Parker Follett who suggested a much more inclusive and collaborative way-of-working where orders were extracted from a shared understanding of the situation to be solved.

“One person should not give orders to another person, but both should agree to [discover and] take their orders from the situation” — M.P.Follett (49)

Follett’s work influenced Management by Objectives (50) and eventually Objectives and Key Results (51). It carries a different approach to work and management that is much more human and where work isn’t designed to be taken over by machines, but where there is a collaborative environment growing of the different strengths of its diverse participants (even computer participants).

We need to recognize that ‘work’ can be many things. And realize that we’ve chosen the one where we are already unhappily being treated as machines and as long as efficiency is the overarching goal ‘resistance is futile’..

But we are standing at a crossroads where we can choose to fight a loosing battle of efficiency, or pick up a different model where computers can only serve part of the need.

Mary Parket Follett tried to warn us, and can still lead the way.

m.Automation and waste

“30% of all human work in finance is replacement for bad middleware” — startup bank CEO, 2014

Emily Gorcenski (43) points out that we are suffering from human beings often being the slowest part in a chain between two computers.

A common scenario would be where a computer collects the data and presents it, a human looks at the output makes a decision and asks a different computer to take action. Why is there a human in that chain?

If a human is reacting on some fixed rules or common experience, then that human could be making models for computers to do the work, instead of doing the work herself.

And there are plenty of opportunities for improvement to choose from. When it comes to work we are currently wasting plenty of time every day on mundane, repetitive tasks (52)(53) that slow down the system we are a part of.

I’m not advocating for the replacement of humans where it makes less sense, I’m advocating for humans to make rules and models, and for computers to do the work that makes us miserable and unhappy.

We are spending too much time conjuring up ideas for how people can add new routines, tasks and processes to work that is already half meaningless.

With AI we should spend more time thinking about how to do less compared to layering on more.

f. work is for computers, humans make models
The environments we are in are becoming so complex that humans alone are not fit for the work that is needed.

Like the tiny human sitting in a big mechanical crane lifting enormous weights far up into the sky, we also need these same superpower machines for thinking.

Imagine the following scenario:

1. You are individually tracking 65.000 customers
2. All are at different places in their customer journey, with different profiles, behaviors and history, coming from somewhere unique going somewhere unique
3. They each have individual needs requiring tailored motivation
4. And any interaction with them could be your first interaction
5. Patterns keep changing as forces of influence keep changing
6. and by the way .. you have five seconds to react

Solving a common problem like this would be impossible for a human alone, while a computer needs to know what problem they are solving, how to understand it and what their role is, if it is working and how to improve (or change).

We can only get to the future if we work together; humans and AI. As with the industrial revolution where our muscles and physical capacity got super powered, this time its the same revolution but for the speed and complexity of our thinking.

In the future of work we would be making models that our computers are using instead of doing the leg work ourselves.

Like our crane driver: lifting the weight is only part of the work, but knowing how to lift it, where from and where to, taking into considerations safety and environmental influence, knowing the timing and communicating with the people on the ground for the perfect fit. The machine can’t do it on their own. It has to be a partnership.

5 — The Outcome Economy

Having access to new information and the ability to see and combine it in new ways enables us to understand and solve problems in new ways.

This includes interpreting and solving for the mechanics and dynamics of how we do business.

AI has the power to drive a business model evolution, where CX can help push that evolution towards a more customer centric approach.

CX helps the company deliver business value through customer value. But this is inherently complex to plan, lead and prove. Which means that most companies have focused on their outputs, not their outcomes. Making the assumptions that their products offer value and that producing more of them for less cost it will increase value to the business.

The customer therefore quickly disappears from the business’ attention span (40). And the company turns its focus towards efficiency and standardization instead of values and needs (54).

This is understandable, until now …

The Outcome Economy’s (55) promise (aprox. 2015) (56) is to identify and measure the outcomes that customers find valuable setting these as measures of success for the company itself.

“Intelligent hardware is bridging the last mile between the digital enterprise and the physical world. .. to give customers what they really want: not more products or services, but more meaningful outcomes. These “digital disrupters” know that getting ahead is no longer about selling things — it’s about selling results.” — Accenture Technology Vision 2015 (57)

e.g. a jet engine manufacturer would use fuel efficiency as the outcome measure, compared to how many jet engines they sold. Because an airline (the customer of jet engines) wouldn’t care about how many jet engines the manufacturer sold, they care about the fuel efficiency of their flights.

g. Getting to better business models
With the use of smart sensors, big data, advanced analytics and artificial intelligence the Outcome Economy is even more available for more businesses and industries.

But the technology on its own won’t get us there. As mentioned previously: technology doesn’t know what it wants (58). It’s up to us to take advantage of the technology in order to get to the business models and practices we need to produce value (59).

With the Outcome Economy the company’s incentive and motivation becomes the same as their customer’s. But where does the company get the insights, sense making, original thinking and solutions that help them combine their desired outcomes with the customer’s? The answer of course is Customer Experience, if CX wants to?

The opportunity is there. Businesses needs a good partner to help them effectively and productively include the customer in their strategic decision making. Companies are investing heavily in technology and data, but as we’ve shown to no effect. Sometimes even making it worse. The door is wide open if CX wants to ride through it.

This requires change, we can’t just bring in old CX. It demands that CX includes business thinking, challenges some of its myopic models and aims to combine business processes and AI more into the maps and models we are creating.

We are already starting, and some brilliant headway is being made. This could become the most valuable opportunity for both business, AI and CX going forward.

In summary

As with every other product, service or process CX also eventually over-shoots the needs and usefulness of its audience.

At the same time AI is sucking all the oxygen out of the room, but seems to have no direction nor purpose other than doing more of the same only faster and cheaper.

That is not how we get to the future.

In this article I’ve attempted to shed a light on some of the opportunities I’ve seen for the future of AI and CX combined. Where CX offers the mindset, tools and practice an organization uses to tame AI, guide it and make it unlock new value for the business and its customers.

AI needs direction and CX needs a disruption. I think this is a match made in heaven.

Sources:

(1). US Now, Clay Shirky, https://www.youtube.com/watch?v=0zUu_uVzEv4

(2). Manage incomplexity don’t simplify it, https://everythingnewisdangerous.medium.com/manage-in-complexity-dont-simplify-it-bcd830e18d14

(3). Using synthetic users for UX research (because who needs humans, anyway… right?), https://www.linkedin.com/pulse/using-synthetic-users-ux-research-because-who-needs-humans-vidakovic-2lezc/

(4). Kevin Kelly, The Inevitable, https://en.wikipedia.org/wiki/The_Inevitable_%28book%29

(5). Shoshana Zuboff, The changing content of capitalism, https://www.180360720.no/?p=5112

(6). Social Media and Political Extremism, https://onlinewilder.vcu.edu/blog/political-extremism/

(7). Tom Miles. U.N. investigators cite Facebook role in Myanmar crisis, https://www.reuters.com/article/idUSKCN1GO2Q4/

(8). Understanding the Excessive Craze of Online Shopping, https://www.psychologs.com/understanding-the-excessive-craze-of-online-shopping/

(9). The Dangerous Experiment on Teen Girls, https://www.theatlantic.com/ideas/archive/2021/11/facebooks-dangerous-experiment-teen-girls/620767/

(10). L. Nicoletti et. al., Humans are biased. Generative AI is even worse, https://www.bloomberg.com/graphics/2023-generative-ai-bias/

(11). These robots were trained on AI. They became sexist and racist, https://www.washingtonpost.com/technology/2022/07/16/racist-robots-ai/

(12). Roberto Torres, How AI learns the biases of its creators, https://www.ciodive.com/news/how-ai-learns-the-biases-of-its-creators/563089/

(13). Cathy O’Neil, “Models are opinions reflected in mathematics” — O’Neil, https://everythingnewisdangerous.medium.com/models-are-opinions-reflected-in-mathematics-oneil-a93ad607f893

(14). The Customer Data Gap, https://everythingnewisdangerous.medium.com/the-customer-data-gap-520cdf695d68

(15). Remembering the human in the age of the algorithm, https://bootcamp.uxdesign.cc/remembering-the-human-in-the-age-of-the-algorithm-1a34333c9956

(16). You can’t measure customers the same way you measure channels, https://uxdesign.cc/you-cant-measure-customers-the-same-way-you-measure-channels-225bcdf609f6

(17). Sensemaking The power of the humanities in the age of the algorithm, https://www.redassociates.com/sensemaking

(18). AI Infrastructure Alliance, Enterprise generative AI adoption, https://go.clear.ml/new-research-report-on-enterprise-generative-ai-adoption

(19). An algorithm doesn’t know what it wantes, https://everythingnewisdangerous.medium.com/an-algorithm-doesnt-know-what-it-wants-7620a1fdfc64

(20). Farnam Street, Mental Models: The Best Way to Make Intelligent Decisions, https://fs.blog/mental-models/

(21). Cathy O’Neil, Weapons of math destruction, https://en.wikipedia.org/wiki/Weapons_of_Math_Destruction

(22). Sanne Blauw, The number bias, https://www.sanneblauw.com/book

(23). Pratibha Kumari J., What is thick data?, https://www.linkedin.com/pulse/what-thick-data-pratibha-kumari-jha/

(24). Keep your products close, but your customers closer, https://bootcamp.uxdesign.cc/keep-your-products-close-but-your-customers-closer-c991c600ef59

(25). Are you solving problems that don’t exist?, https://betterprogramming.pub/are-you-solving-problems-that-dont-exist-c36685ad8c8b

(26). Dr. Joel Hoomans, 35,000 Decisions: The Great Choices of Strategic Leaders, https://go.roberts.edu/leadingedge/the-great-choices-of-strategic-leaders

(27). Gartner Predicts 80% of Marketers Will Abandon Personalization Efforts by 2025, https://www.gartner.com/en/newsroom/press-releases/2019-12-02-gartner-predicts-80–of-marketers-will-abandon-person

(28). Will personalization fail (by 2025)?, https://everythingnewisdangerous.medium.com/will-personalization-fail-by-2025-850f5234feaa

(29). Matthew Addicoat, https://theconversation.com/only-1-of-chemical-compounds-have-been-discovered-heres-how-we-search-for-others-that-could-change-the-world-211302, Only 1% of chemical compounds have been discovered — here’s how we search for others that could change the world

(30). Dinald Rumsfeld, Rumsfeld / knowns, https://www.youtube.com/watch?v=REWeBzGuzCc

(31). Evie Brockwell, Lessons from Booking.com experimentation culture, https://www.hustlebadger.com/what-do-product-teams-do/booking-com-experimentation-culture/

(32). Rik Higham, Get Comfortable Breaking Your Product by Rik Higham at Mind the Product London 2018, https://www.youtube.com/watch?v=yJgRzLugU1I

(33). Experiment like Spotify, https://confidence.spotify.com

(34). Dave Snowden, The Cynefin framework, https://youtu.be/N7oz366X0-8?feature=shared

(35). Roger Martin, A plan is not a strategy, https://www.youtube.com/watch?v=iuYlGRnC7J8

(36). Søren Fromberg, Does the rule of “50% of media spend is wasted” still work?, https://annalectnordics.com/do-the-rule-of-50-of-media-spend-is-waste-still-works/

(37). To be great at innovation you need a great innovation process, https://strategyn.com/outcome-driven-innovation-process/

(38). Maciej Marcinowski, Artificial Intelligence or the Ultimate Tool for Conservatism, https://sciendo.com/article/10.2478/danb-2022-0001

(39). Shoshana Zuboff, Creating value in the age of distributed capitalism, https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/creating-value-in-the-age-of-distributed-capitalism

(40). Clayton Christensen, Clayton Christensen (Innovator’s Dilemma) & Marc Andreessen (a16z) | Startup Grind Global ,https://www.youtube.com/watch?v=IkBp1ntD3Zc&t=1812s

(41). Henry David Thoreau, Walden, https://en.wikipedia.org/wiki/Walden

(42). Keith Houston, Empire of the sum, https://99percentinvisible.org/episode/empire-of-the-sum/

(43). Emily F. Gorcenski, https://emilygorcenski.com

(44). Kevin Slavin, Design as participation, https://jods.mitpress.mit.edu/pub/design-as-participation/release/1

(45). Scientific Management, https://en.wikipedia.org/wiki/Scientific_management

(46). Kevin Kelly, How AI can bring on a second Industrial Revolution, https://www.youtube.com/watch?v=IjbTiRbeNpM

(47). Cory Doctorow, The Cory Doctorow Interview — Part I, https://www.gaslitnationpod.com/episodes-transcripts-20/2023/08/23/cory-doctorow-interview-part-1

(48). State of the global workplace, https://www.gallup.com/workplace/349484/state-of-the-global-workplace.aspx

(49). Mary Parker Follett, The giving of orders, https://www.linkedin.com/pulse/mary-parket-follett-giving-orders-1926-helge-tennø/

(50). Management by objectives, https://en.wikipedia.org/wiki/Management_by_objectives

(51). Objectives and key results, https://en.wikipedia.org/wiki/Objectives_and_key_results

(52). https://www.inc.com/david-finkel/new-study-shows-youre-wasting-218-hours-a-week.html

(53). https://di.ku.dk/english/news/2023/even-though-our-computers-are-now-better-than-15-years-ago-they-still-malfunction-between-11-and-20-percent-of-the-time-a-ne_/

(54). Shoshana Zuboff, New logics, new business models, new commercial frameworks, https://everythingnewisdangerous.medium.com/new-logics-new-business-models-new-commercial-frameworks-e133f67db035

(55). The outcome economy the bigger picture on the wall, https://www.virginmediabusiness.co.uk/insights/the-outcome-economy/

(56). The outcome economy a data driven economy, https://www.180360720.no/?p=5246

(57). When atoms meet bits, https://www.accenture.com/us-en/insights/technology/technology-trends-2023

(58). Kevin Kelly, What technology wants, https://en.wikipedia.org/wiki/What_Technology_Wants

(59). Shoshana Zuboff, Disruptions Tragic Flaw, https://www.faz.net/aktuell/feuilleton/debatten/the-digital-debate/shoshana-zuboff-on-the-sharing-economy-13500770.html

---

Why AI’s lack of direction is perfect for CX’s loss of direction was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.