Utbenat: Friction

Introduction

Sebastian:
Picture yourself in front of the computer at half past eleven on a Tuesday night. You should have gone to bed an hour ago. But you're stuck. You've found the perfect office chair for your home office. Good-looking, good reviews, reasonable price. You've added it to the cart and gone to checkout.

And then it starts.

Create an account. Enter your email. Choose a password. Confirm the password. First name. Last name. Address. Postcode. City. Phone number. "Would you like our newsletter?" Absolutely not. Now the payment. Card number. Expiry date. CVC – hold on, where's my wallet? Then: "We've sent a verification code to your phone." Right, it's dead, where's the charger? And finally: "Confirm your purchase." At last. You click.

"Something went wrong. Please try again."

You close the tab. You go to bed. The chair stays unsold.

At that very moment, a couple of kilometres away, someone else is buying a near-identical office chair. That person isn't met by a dozen form fields. That person clicks "Pay with Klarna", enters their email address and postcode. Nine seconds, that's all. No account. No fields. No … friction.

Today we're going to talk about what happened during those nine seconds, and what didn't happen for the rest of us. We're going to talk about friction. That little bit of extra resistance that makes customers abandon their carts, makes employees work around the system, and makes quoting processes drag on. And we're going to talk about why the best solution to friction sometimes has nothing whatsoever to do with AI – and sometimes has everything to do with AI.

Sand in the machinery

Sebastian:
Friction, in a business context, is exactly what it is in physics: the thing that makes energy disappear along the way. Every extra click. Every unnecessary field to fill in. Every time a customer has to wait for "someone" to get back to them. Every internal process where Anna has to email Johan, who has to approve something in one system, which Petra then keys manually into another system.

It's sand in the machinery. And here's the tricky part: unlike a physical machine, which starts making noise when it seizes up, business friction is almost always completely silent. No alert report says: "Warning, you're losing hours every week to redundant process steps." The friction is just … there. And we've got used to it.

Resistance that has become normal

Sebastian:
What makes friction so insidious is precisely this: when something has always been slow, we stop perceiving it as slow. It becomes normal.

"That's how long a quote takes." "That's how many steps it takes to onboard a new supplier." "Well of course onboarding a new hire takes two months, there's a lot to get in place."

We stop questioning friction because we don't notice it, and don't think of it, as friction. And this is where AI enters the picture. Because right now there's an expectation that "AI will solve everything". And sure, AI can be phenomenally good at handling certain kinds of friction. An AI chatbot that answers customer questions around the clock instead of the customer waiting for an email reply on Monday morning. An AI that reads and summarises a request for quotation, matches it against the right products and generates a first draft in an instant. That's real.

But – and this is the crucial part – throwing AI at a problem without first understanding why the problem exists is like gutting and rebuilding the entire office kitchen because the coffee tastes bad. New lighting, new worktops, a four-thousand-pound espresso machine. And the coffee still tastes awful. Because the problem was never the kitchen or the machine. The problem was that not even the best machine can turn bad coffee into good coffee.

The Klarna paradox

Sebastian:
Let's stay with Klarna for a moment. And Qliro. And Swish. And BankID. Together, these companies and solutions have eliminated an astronomical amount of friction from everyday digital life in Sweden. Going from digging out a card, keying in the card number, expiry date and CVC and verifying with the bank using a physical code generator, to simply entering your email and using face recognition – that's one of the biggest leaps in user experience we've seen. Decades of entrenched online payment friction, gone.

But here's the thing: almost none of that is AI-driven. Klarna's core – handling the payment flow between customer and merchant – is a smart combination of UX design, risk models and integrations. Not a large language model. Not real-time machine learning on every transaction.

I call this the technology paradox: the biggest wins in reducing friction often don't come from the most advanced technology. They come from someone actually sitting down, looking at how a human being experiences a purchase, and daring to ask: "Do we really need to ask for all of this?" And yet at the same time, it's often technology that gets us to challenge the status quo in the first place.

At the same time, AI applied carelessly can create new friction. Have you ever been stuck in a chatbot loop where it doesn't understand what you mean, and you can't reach a human no matter how many times you type "I want to speak to a real person"? That's AI adding friction. A well-meaning AI assistant that pops up asking "How can I help you?" at the exact moment you've found what you were looking for on the site. Friction. An automated system that miscategorises your support ticket so you end up in the wrong queue. Friction.

The right tool for the right friction

Sebastian:
So instead of asking "How can we use AI to reduce friction?", the question ought to be: "Where is our friction, and what's the right tool to remove that particular friction?"

Let me give you a concrete example. Think about the process of onboarding a new product from a new supplier. It often looks something like this:

Someone in purchasing gets an email from an interesting supplier. That person forwards it to the product manager. The product manager books a meeting. After the meeting the supplier has to fill in a form, often a document emailed back and forth. When the form comes in, it has to be reviewed. Comments come up, which are emailed back. When everything is done, the supplier has to be set up in the ERP system, possibly manually, field by field. The product has to be registered with an article number, price, description, image. Someone has to rewrite the product description from the supplier's format into "our format". And somewhere along the way, finance has to approve the payment terms and the logistics have to be sorted out.

This whole carousel takes, if everything runs smoothly, maybe four weeks. And nobody questions it. Because "that's how you onboard a new supplier".

But if we stop and look at each step: how much of this is pure administrative friction? How many of those emails are really just a human being manually moving information from one place to another? This is where the real question comes in. Not "can AI solve this?" but: which of these steps actually add anything?

Maybe the solution is to replace the form emailed back and forth with a web form that validates data on the spot, flags errors immediately, and pushes the data straight into the systems once everything is correct. That isn't AI. That's an integration that would have been possible ten years ago. But nobody built it, because nobody questioned the process.

And once you've removed the unnecessary friction, that's when AI can come in and do the things that genuinely require intelligence: ingesting and structuring the supplier's product information, suggesting categorisation, generating a first draft of the product description in your format, flagging deviations from your quality requirements. But that's only possible once you've made sure the AI has any data to work with at all.

So the answer to a friction problem can be AI. But the answer can also be removing one form field. Changing the text on a button. Letting the customer pay with Swish instead of a card. Giving an employee the mandate to approve an order directly, instead of emailing their line manager and waiting until the next day. It's rarely the most expensive solutions that have the biggest effect. It's the most precise ones.

The friction hunt: three ways to find it

Sebastian:
So how do we find the friction? Because that really is the hard part. Not fixing it, but seeing it. Here are three concrete ways.

First: do it yourself. Walk through your own customer journey. Not as a PowerPoint exercise, but for real. Sit down one evening and try to become a customer of your own company. Buy your cheapest product. Request a quote. Raise a support ticket. Count the clicks. Notice how long it takes. Think about when your customers would be doing this. Pay attention to where you get irritated. That's where the friction lives.

Second: ask the people who know. Not managers or middle managers. Ask the person in customer service who hears the same complaint over and over. Ask the salesperson who has to tell customers every week to "just fill this in as well". They know exactly where the friction is, but they're rarely asked.

Third: look at the drop-off. Where in your flows are you losing people? Is it between quote and close? Is it when someone has to upload a file? It isn't chance that decides whether people bail out. And it's often down to friction.

Only once you've mapped where the friction actually is – only then is it time to ask: what's the right tool here?

The door nobody thought to oil

Sebastian:
Picture an office. In this office there's a door out to the lunchroom. That door has been sticking for three years. Everyone going out has to put a shoulder into it. Everyone coming in has to give it a good yank. Nobody thinks about it any more. New hires quickly learn from the others: "Oh yes, that door, you just have to give it a shove."

One day a summer temp arrives. On the first day, after lunch, they say: "Why has nobody oiled the hinges?"

Everyone stops. Oiled the hinges. It was such a completely obvious thing. A tiny bit of effort. A couple of drops of oil. But nobody had done it. Because everyone had stopped seeing the door. They'd got used to the resistance.

This is what it looks like in almost every organisation I walk into. There are sticking doors everywhere. Processes where everyone has learned to "push a bit harder". And now, with today's toolbox – which holds AI, automation, smart APIs and, not least, a new way of thinking about user experience – we suddenly have more lubricant than ever. But it doesn't help if we don't see the door first.

What to do on Monday

Sebastian:
Pick one single journey. It could be the customer's journey, from "I'm interested" to "I've paid". It could be an internal process, like that "new product from a new supplier" we talked about earlier: from an email landing in purchasing to the product actually being live in the systems, with the right price, the right description and a green light from everyone involved.

Walk through that journey step by step. Not in a spreadsheet or a PowerPoint, but for real. Every step. At each stage, ask yourselves: "Does this step need to exist? Does it add value for the customer, or for us?"

Write down everything that sticks. Every unnecessary field. Every waiting period. Every "let me just check with…". And only once you have that list, ask yourselves: what tools do we need to fix this?

Maybe the answer is an AI assistant that can support you in responding to requests for quotation. Maybe the answer is removing a mandatory field whose data nobody actually uses. Maybe the answer is a payment solution that's already there, ready to integrate.

Because the beautiful thing about friction is that it's almost always cheaper to remove it than to live with it. A drop of oil costs next to nothing. A lost customer costs a great deal.

Summary

Sebastian:
To sum up: friction is the silent resistance that makes customers abandon their carts and colleagues work around the systems. AI can be a powerful tool for reducing friction, but only if we first identify where the friction actually is. The biggest wins rarely come from the most expensive technology, but from the most precise action. Start by finding your sticking doors. Oil the hinges.

Thanks for listening to today's episode of Utbenat! If you'd like to talk more about today's topic, or just talk data, get in touch – you'll find us on LinkedIn. Or contact us through our website. Talk soon, and take care out there!

You've been listening to Syntra, a podcast from Fiwe.