Global Services Director · Softcat plc Now booking — Q3 & Q4 2026 LinkedIn ↗
The radar — an 18-month horizon, held in public

What hits service next — and where to stand.

Six signals I am tracking from the operating floor — each with the consensus take the industry will comfort itself with, and the position I would actually defend. Held in public so you can mark my homework later.

Last revised June 2026 · Updated as the picture changes
01Already here

Your next user is not a person. It is their agent.

The signal

AI agents are starting to act on the user's behalf — diagnosing, escalating, transacting, resolving. The best support experiences already feel like this: a fault confirmed, a replacement shipped, before the conversation has finished.

The lazy take

"Chatbots, version two. Bolt one onto the portal, count the deflections, report the saving."

The position

Agents are a new class of customer. Your catalogue, your entitlements and your service contracts must become machine-readable and machine-actionable. Experience design now includes users who never see your interface.

First move Inventory which of your services an agent could request, consume and verify today without a human translating. That number is your real digital maturity score.

From the essays: Smarter systems, not more people

0212 months

The face becomes the endpoint.

The signal

Meta is selling AI Ray-Bans at consumer prices. Gucci is making smart glasses with Google. A keynote has already been delivered from a pair — no notes, no visible script, five hundred delegates none the wiser.

The lazy take

"Niche gadgetry. We'll revisit wearables when there's enterprise demand."

The position

Within eighteen months a meaningful slice of your workforce wears cameras, microphones and an assistant on their face. Support becomes ambient; the distance between question and answer collapses. How do you write a BYOD policy for someone's face? Slowly, badly — or now.

First move Draft the wearables position paper before the first pair walks into your office, not after. Privacy, consent, support model, meeting etiquette.

From the essays: The wearable that delivered a keynote

03Already here

The operating model becomes software.

The signal

Provisioning in code, assurance by telemetry, resolution by agents. The routine layer of service is being absorbed into software faster than most organisations can redraw their org charts.

The lazy take

"We're doing AI — we've bought the platform and stood up a centre of excellence."

The position

64% of UK organisations say they use AI; roughly a quarter are anywhere near mature adoption. The gap is not technology, it is organisational design. AI lands or stalls in the operating model — and most were built for tickets, not flow. Stop buying tools to manage queues. Build services that do not queue.

First move Take one high-volume service end to end and rebuild it software-first: sensed, resolved, improved in code — humans only at the moments that deserve them.

From the manifesto: The Human Operating Model

0412–18 months

The smartphone quietly loses the centre.

The signal

Glasses on faces, intelligence in earpieces, watches losing their jobs to ambient assistants. The phone is still in the pocket — it is just no longer where the action is.

The lazy take

"Mobile-first remains the strategy. We've only just finished the app."

The position

The post-smartphone workplace arrives unevenly — but every service designed around "open the app, log the ticket" inherits a dying interaction model. Design for intent expressed anywhere: spoken, glanced, delegated to an agent.

First move Map your top ten user journeys and ask which survive when nobody opens an app to start them. Redesign the ones that don't.

From the essays: Have we already moved on from the smartphone?

05Already here

The workforce tears up the service contract.

The signal

A generation raised on instant, ambient, self-evidently competent service is now your workforce. They will not queue. They will not log a ticket and wait for a callback.

The lazy take

"Satisfaction scores are holding up. The desk is hitting its SLAs."

The position

The risk is not poor satisfaction scores. It is irrelevance — workarounds, shadow tools, a quiet spreadsheet called "things IT can't fix". A 98% SLA next to that spreadsheet is not success; it is a brilliantly organised description of yesterday.

First move Find the shadow channels people actually use when they skip your desk. That map — not your dashboard — is the truth about your service.

From the essays: Why Gen Z won't wait for your service desk

0618+ months

Competence becomes a commodity. Judgement becomes the market.

The signal

Every provider, every team, every individual gains the same competent baseline at roughly the same time. The average deliverable is now free.

The lazy take

"Automate the existing model, bank the efficiency, publish the AI strategy."

The position

Most AI strategies are a detailed plan to become average, faster. The floor rises for everyone; the ceiling stays human — judgement, taste, context, the ability to pick the right problem. Spend your best people there, or spend them making yesterday cheaper while the game moves.

First move Audit where your scarcest resource — senior judgement — is actually spent. If it is mostly supervising the routine, your AI strategy is a cost programme wearing a strategy costume.

From the essays: Average, faster

Disagree?

Good. Bring it to the room.

Every position on this page is held in public and argued live on stage. If your board is wrestling with any of them, that is a conversation worth having properly.

Book Mark to speak