Who Decided How Your Service Desk Works? 

Most organisations have never actually chosen how demand reaches their service desk. It just accumulated. Here is what that costs and how to change it. 

A few weeks had passed since Gary's last visit. In between, I had been in two meetings that I could not stop thinking about. 

The first was with the CIO of a law firm. Intelligent, commercially sharp, running a desk that by any standard metric looked expensive. She knew it. But every time the conversation moved toward efficiency, she pulled it back. Her managing partners did not wait. Her fee earners did not queue. That was not a failure of her service desk. That was a deliberate decision, made at board level, about what the firm’s most valuable people should and should not spend their time doing. White glove service for the people who needed it. 

The second was with a CIO at a financial services firm. Different conversation entirely. He had a platform change coming, a board mandate to reduce operational cost, and an appetite for automation that most of the organisations I speak to do not have. He wanted to move fast and he wanted to know what good looked like at the end of it. 

Same model. Completely different answers. I wanted to understand why. 

Gary arrived with cake again. Different bakery, same ritual. He opened his laptop before I had even filled the kettle. 

Gary: You have been thinking about the law firm. 

Keyvan: Is it that obvious? 

Gary: You get a particular look. What is the question? 

Keyvan: The ratio tells you whether you are getting more efficient. But she does not want to get more efficient. At least not in the way we have been talking about it. She wants her lawyers looked after. So what does the model say to her? 

Gary: It says she is making a choice. And that is fine. But she needs to make it deliberately and know what it costs her. Most organisations in her position have never actually decided how demand reaches the desk. It just accumulated. Channel by channel, year by year, nobody ever sat down and designed it. 

That landed. Contact strategy. The way demand reaches the desk in the first place, and what happens to it when it gets there. Gary pulled up a chart. Phone was still the dominant channel for most organisations, carrying nearly half of all ticket volume. Highest cost per contact. Longest resolution times. And in most cases, not a deliberate choice. Just a habit nobody had examined. 

Gary: The law firm CIO is not wrong to protect her fee earners. But there is a difference between protecting them and giving everyone in the organisation the same level of service because nobody ever drew a line. You can have a white glove tier for the people who genuinely need it, across whatever channels they use. And you can have a completely different model for everyone else. Most firms have never separated the two. They run one service model for three thousand people — and in purely operational terms, that is the more efficient choice. It just does not fit an organisation where some of those people cannot afford to wait. 

Keyvan: And that is where the cost sits. 

Gary: Phone and unstructured email are the two biggest drains on agent time in almost every organisation we look at. Chat — whether that is Teams or any other platform — is more efficient than either, but only when it is properly set up and managed. Self-service is more efficient still, and conversational AI means self-service can now live directly inside those chat channels rather than sitting behind a separate portal. But you have to design for it. You have to decide deliberately who gets what, and why, and build the channels around that. Otherwise you are just managing whatever turns up. 

I thought about the financial services CIO. He had no white glove constraint. He had a board that wanted numbers to move and a platform ready to support it. His question was different. Not who gets what service, but how fast can we change the ratio. 

Keyvan: What do you tell him? 

Gary: The ratio is what you are trying to move. Everything you do either moves it or it does not. So you need to know where it is before you start. 

Keyvan: And that is harder than it sounds. 

Gary: It is harder than most organisations expect. You need ticket volumes by channel, by category, by resolution type. You need agent numbers, properly accounted for. You need to know what is currently deflected and what is not. And you need to know where automation already exists and where it does not. Most of that data exists. It just has never been pulled together and read as a single picture. 

This is the point where most conversations about service desk transformation quietly lose their nerve. The data exists. Most organisations have it sitting in their ITSM tools. The problem is that the reporting infrastructure built around it — SQL-based, structured, rigid — often cannot surface it accurately. The numbers either do not come out or they come out wrong. That is where AI analysis changes the picture: not by finding new data, but by making the data that is already there actually readable. But nobody has ever looked at it all at once and asked the honest question: where are we actually starting from? 

What Gary described next was the model we use when we go into an organisation seriously. Not a discovery workshop with a lot of sticky notes. A structured analysis, running across the data that already exists, benchmarked against four hundred KPIs across forty industry comparisons, producing a scorecard that shows exactly where the floor is soft and exactly which beams need work. Within five working days you have something real to act on. 

Gary: The scorecard is not the answer. It is the starting point. You cannot move the ratio if you do not know what is driving it. And you cannot know what is driving it if you are looking at three numbers in three different systems and calling that a picture. 

Keyvan: And the six month cycle? 

Gary: That is where it becomes something you can actually manage. You baseline, you prioritise, you act, you measure. Six months later you baseline again. The ratio tells you whether anything moved. And if it did not, the scorecard tells you why. 

Gary closed his laptop. The cake was gone. Outside it was one of those grey mornings that had been grey since before anyone got up, the kind that makes the kitchen feel like the only warm place in the world. 

I thought about the law firm CIO and the financial services CIO on the drive in later that morning. Their desks look nothing alike. Their targets look nothing alike. But the question underneath is the same one. Do you actually know what your floor is made of, and do you know if it is holding weight? 

Most organisations think they do. The data usually says otherwise. 

This is the first of six beams we are going to look at properly in this series. The Modern Service Desk is where the cost lives and where the measurement starts. On my travels, the next conversation is about self-service — not the portal nobody uses, but what it actually takes to make deflection work, and why most organisations are still waiting for the returns they were promised. 

Same house. Different room.