Workplace 2030 MSP: a business question, not a technical one
AuthorMSP Late NightPublished on:The workplace 2030 MSP discussion is usually held as if it is about screens, headsets and agents. A director of a 35 person MSP described how his clients had been asking for Copilot for months. His first question back: is your data in order? Silence on the other end. That is where the real conversation sits, and that is where this story starts.
The average MSP still runs on a billing model built around seats and users. Three year contract, five year contract, monthly invoice, an SLA for when something breaks. That works as long as the number of workplaces stays stable and as long as technology is the scarce commodity. Both assumptions are evaporating.
Clients walk in through the back door with Claude, with Cloud Coworker, with Perplexity. They connect their Google Drive and their Microsoft environment without anyone on the service desk noticing. At the same time the seat count at the client falls as soon as those agents start taking on serious work. The MSP that calmly said in 2024 “we do managed services, that is not going anywhere” sits in 2027 with a revenue curve bending the wrong way and a team trained to clear tickets, not to talk to a director about his EBITDA.
And then there is the uncomfortable part: most MSP owners know this. They talk about it by the coffee machine. But the seniors are dealing with today's challenge, yesterday's problem and tomorrow's opportunity all at once. There is no time to rebuild your own proposition while the work keeps coming in.
What an MSP sells in 2030 is no longer a workplace
What works in practice is a shift in the first conversation. Do not arrive as the IT party bringing a quote for 80 workplaces. Arrive with the question: what are you actually here to do as a business? What output do you expect from your people, and what sits in between? Only then does the technical layer come into view, and suddenly it serves the conversation instead of leading it.
The MSPs making this move have three things in order before they touch it at a client. Their own data is classified and made available in one place. Their own workflows are partly automated, so they know where it hurts. And their own team is trained on more than Copilot, with open router, Claude, Gemini and whatever arrives next month. Only then are you a reference for your client, and only then can you hold the business value conversation without it turning into marketing talk.
The three real questions that decide the workplace of 2030
The technology is no longer the interesting part. It has to work, it is about productivity, flexibility and connection. Full stop. Underneath sit three questions most MSPs have no clear answer to, and that the client does want to discuss as soon as you raise them.
What function does the office still have?
Everyone needs an office. The question is whether the office has to be there for you at all times, and you for the office. Staff have built their own cocoon at home with large screens and good light. Then they come into an office with a white ceiling and grey light and are expected to be productive. Occupancy of property becomes a cost line almost nobody looks at seriously. An MSP that discusses with a client how many people will still be in the office in three years, because of or in spite of AI, is suddenly sitting at a very different table from the service desk.
“Look at the occupancy of your property.”
In concrete terms the workplace conversation shifts from “how many devices” to “which modes of work do you facilitate”. Deep work at home, collaboration in the office, client contact hybrid. Each of those needs a different piece of technology, a different piece of furniture and a different agreement. The MSP that puts this on the table together with HR and facilities suddenly becomes a strategic partner instead of a supplier. That is a fundamentally different sales conversation, and it produces fundamentally different margins too.
How do you make money when seats fall?
The Copilot licence itself earns little. The work sits at the front: classifying data, applying role based access, guiding adoption. And at the back: managing those agents once they are in place. A client case we hear back often: 200 users only get Copilot once they accept stricter security policies, in exchange for access to more company data. That is a payoff model the MSP designed, not Microsoft. That is where the margin sits. The MSP that clings to price per seat will not earn from AI in 2030, it will earn in spite of AI.
Run the numbers on your own client base. Say a client with 100 employees has 100 seats today. In three years that may be 70, because agents take over the work of a few juniors. At the same time 30 agents arrive that also need management, security and governance. If you still charge per seat, your revenue halves. If you charge per automated process, per agent under management or per business outcome, your revenue rises. The sum is not complicated, but it does ask you to revise the commercial model today.
Who does the education, if school does not?
Education is behind. Children are still taught at school how to use a search engine while they work with Claude at home. China, South Korea and Japan are on a different track. For the Dutch MSP that means something unpleasant and something interesting at the same time. Unpleasant: young people arrive insufficiently prepared, so you train them yourself. Interesting: if school does not do it and government does not do it, one party is left that can bring the client up to speed on AI. That is the MSP. This is not an extra service, it is the reason you exist for the next five years. The MSP still sitting in the consciously incompetent phase cannot deliver that, and will be overtaken within two years by a 20 year old on the client's team.
“The education, with all respect, is us.”
Practically: this means your internal training budget is no longer a cost line but your most important R&D investment. Who on your team uses three different LLMs every day? Who can build an agent without code? Who can explain to a director why one workflow is suitable for automation and another is not? If the answer to those questions is “nobody” or “one person”, you know where the next six months are going. Education is not a one day seminar, it is a weekly discipline where your own people have to run literally ahead of the pack.
The workplace of 2030 is no longer a technical product. It is a mix of productivity, security, education and human connection, in which the MSP decides whether it sits at the board table or in the utilities corner next to gas, water and electricity. Anyone still waiting for the seats to come back by themselves will be looking in three years at a team trained for a market that no longer exists. Anyone starting now with their own data, their own workflows and their own training will have a story in three years that clients cannot build themselves. Start this week: pick one client, have one different conversation.