Your best people are doing two jobs.
One you’re paying for. The second you’re not.
The second job is everything they do to get the first job done. It’s all the inefficiencies and the lack of clarity — the work redone because someone else had already solved it, the handoff that should have been a handshake. The second job gets heavier the further the work travels from where it started.
~ Phil Verghis, founder of Verghis Group
I’ve been watching this since 1994 — what gets lost in the handoff, what breaks open in the handshake.
AI will only make things messier, faster — unless we set the guidelines and guardrails right.
The three engagements that follow are ones I’ve spoken about publicly with each of these companies, jointly and separately, at industry events over the years.
Akamai was where I first saw it at full scale. SaaS wasn’t a category yet. Cloud didn’t have its current meaning. Customer success wasn’t a function. Swarming wasn’t a model. We were running early versions of all of it because that’s what the work demanded. What held any of it together was knowledge moving in real time across groups that didn’t share vocabulary, measures, or rhythms. That’s where I started seeing the patterns underneath everything I’ve worked on since.
We started as a public company with 1,475 servers across 55 networks. Two years later we were running 13,500 servers across more than a thousand networks in sixty-six countries.
Two years in, we lost Danny Lewin on the first plane hijacked on September 11th. He was a co-founder and the mathematical mind behind the algorithms. He was also the kind of charismatic genius a young company organizes itself around. The economy was imploding around us — dotcom bust, post-9/11 shock, customers going under, telco partners going under. As the company downsized, I was asked to take on Global Network, Operations, and IT on top of Service Delivery.
The network team was given a hard goal — cut what we were spending on the networks we ran on. Bandwidth costs were enormous, and the telcos themselves were collapsing one after another.
Akamai’s network ran on top of those telcos. When a telco went under, our presence inside it went with it. The team raced to relocate before each collapse. Sometimes they made it. Sometimes the network failed first, and customers’ traffic went down. Tens of thousands of end users felt it.
The cross-functional team — network, service delivery, customer-facing leads working it together, not handing it off — figured out a different answer. One of the linchpins: set up our biggest customers’ traffic so a single network failure underneath didn’t matter, because switching to another would happen automatically.
That’s the telco hotel — a neutral facility where many networks met in one place. Move a customer’s endpoint (the point where their traffic connected to ours) into one, and any single telco failure underneath could be routed around in real time. No manual scramble. No customer disruption.
The counterintuitive piece: each endpoint at a telco hotel cost more per unit than the same endpoint on a single network. We absorbed the extra cost out of my team’s budget. Customer effort dropped. Employee effort dropped. The company saved thirty-three million dollars that year.
What I saw at Akamai was one part of the pattern. I’d be tracking it ever since. The answer comes from groups across the work owning the same outcome together — seeing where the cost lands and building the next move from there. Each move taught us how to read the next.
Akamai continued to grow and became the multi-billion-dollar platform behind much of how the modern internet works.
Red Hat was a different kind of company. Built on open source — and open source meant defaulting to open across everything: hyper-local decision making, every voice in the room, willingness to challenge from any level. Their new support executive heard me speak at an event and brought me in for a facilitated offsite. That became years of work together.
The culture shaped how the work had to be done. I spent many hours interviewing their people across the world. The information was powerful and the conversations gave people a sense of their voice being heard, but it wasn’t scalable. That was the genesis of ServiceDNA: there had to be a way to read the conditions underneath the work without putting that many hours into every engagement.
Red Hat’s culture extended even to the measures themselves — usually the part of a company nobody is allowed to look at. Red Hat and I looked together. Following the measures across the flow — from when the customer entered the system to when the work was delivered — we saw many groups pulling against each other, the cost living in the gaps between them. We worked across functions to align the measures around what mattered to the customer at the end of the line. The teams could finally see what their numbers meant to the people next to them. The results were dramatic, with clarity from one end of the work to the other.
This was one of a number of powerful initiatives — the collective work Red Hat and I did together. It earned its own Ivey Publishing case study, the business school press that sits second only to Harvard in case study distribution, available to business schools and companies around the world to teach from.
HPE was a different kind of company again. Multi-billion-dollar global enterprise, structured to scale. Their knowledge team was already excellent — respected across the industry, with competitors coming to them to learn how it was done. I’d been speaking at a conference and their leadership was in the audience. They brought me in for a workshop. The workshop became the engagement.
But they had a goal they couldn’t reach from inside their own excellence: the TSIA Star Award for Best Practices in Knowledge Management — the industry’s peer-judged recognition, and the hardest to win in this field. They’d done everything they knew how to do and still couldn’t close the gap.
HPE’s scale — thousands of people across the support organization — meant hundreds of hours of conversations weren’t on the table. ServiceDNA was. By then it had matured into an instrument that produced a behavioral fingerprint of the operation: where compensating work was concentrating, where the cost between groups was piling up, and the order to act — what to take on first, what would come next.
HPE won the TSIA Star Award a year into the engagement. The number that came with it — blessed by the HPE CFO, which is how you know it’s real — was $58 million in bottom line savings. Most knowledge programs can’t produce a CFO-signoff number. The causal chain is too fuzzy. This one could.
The more durable proof came later. The HPE President’s Quality Award is an internal competition, judged by top leadership, chosen by the CEO, inside a multi-billion-dollar company where thousands of groups produce excellent work. Winning it once is meaningful. HPE’s knowledge team won it twice — once during our engagement, once a few years after I left, on the same foundation. Winning it twice means clearing a bar that got higher.
Three engagements, in outline. Before, during, and after them — many more, across industries, sizes, and stages of company life. I kept learning, and applying what I learned.
Selected Engagements
What I hadn’t done — across all those years of work — was run an operation myself, in a domain new to me.
A friend and colleague from my Akamai days had co-founded a 3D-printing platform behind medical-grade dental devices, and I came in first as a consultant, then transitioned in. What used to take weeks of waiting on a dental lab — printing models, thermoforming plastic, mailing results back and forth — became two hours of work in the same appointment. Time to Smile, literally.
Working with a talented team and an innovative CEO, I chose the AI service underneath the platform and the AI software that spanned engineering, sales, and operations.
Coming back out of LuxCreo and into consulting, I could see what the whole arc had been pointing at. Operations across domains that don’t look alike on the surface — support, manufacturing, healthcare, software, services — turn out to run on the same underlying mechanics. Now AI gives me tools I didn’t have before.
This is the moment I’ve been working toward. I can’t wait.