Wired to automate

⁠Long before AI demanded the mindset to automate tasks and free up time for human energy, Sam Godfrey was already living it.
⁠
⁠When Sam Godfrey was working as an IT technician at a hospital in the early 2000s, he quietly automated most of his own job. Not as part of some grand plan. He just built systems to stop problems from recurring, kept refining them, and eventually compressed his working week to one or two days. The job was mostly running itself.
⁠
⁠That was twenty years before the industry decided that automating repetitive work and freeing up human time was a revolution.

⁠Sam Godfrey, founder of UK-based MSP Task Group, has been AI-ready since long before AI arrived. AI is now shaking up the IT services and managed services space, teaching the industry to use technology to free up time, remove repetitive work and focus human energy where it actually counts. It's something Sam has been doing for around 15 years.

⁠So when AI showed up, it didn't ask him to change how he thinks. It handed him a better toolkit.

⁠Sam’s tryst with tech began when his parents brought home a ZX Spectrum 48k in the mid-1980s. The Spectrum was a home computer released by Sinclair Research in 1982, one of the first affordable machines to land in UK living rooms. It came with no hand-holding. There was no internet, no tutorials and no forum to consult. If something broke, you figured it out yourself or it stayed broken.

⁠"Those were days when you had to understand how a computer did things," Sam says. "A lot of pulling things apart and fixing what I broke."

⁠Over the years, he became the person everyone around him called when something was not working. By 2001 he had a first-class honors degree in computing. The year after, he helped a close friend get a first in his final year project and took away a lesson he would carry into everything that followed: what you can accomplish when you distribute responsibility rather than hold it all yourself.

⁠A few standard jobs came and went, what Sam calls his "real jobs," before he landed in onsite IT support at hospitals. The instinct that had been forming for years turned into deliberate practice here. His focus was preventing problems before they started. He built scripts, automated what he could and gradually got his working week down to one or two days.

⁠Brigitte, his partner and eventual co-founder of Task Group, pointed out that the freed-up time could go somewhere useful. They started offering IT services to local home users on the side of the day job, and ran that for about a year until one of those home users asked if they could help with their business.

⁠That was 17 years ago. And was the entry that revealed a gaping hole that was hard to ignore.

⁠Kernels of the managed services business

⁠Sam noticed a pattern with IT support for  small businesses. This is what it looked like: something breaks, you call someone, you wait, someone shows up, fixes the immediate problem and leaves. Nothing changes about the underlying system, so the same problems come back.

⁠
⁠Essentially, small businesses were being underserved and overcharged. Most IT providers were reactive and hard to deal with. And small businesses were a lot easier to work with than home users.

⁠Task Group built its model around fixing all three: one point of contact for anything technology-related, priced as a fixed monthly service, focused on preventing problems rather than just responding to them.  It’s what Sam calls "all you can eat" support, a single relationship covering everything technology-related, with no meter running every time something went wrong.

⁠Sam and Brigitte were still working their day jobs, which gave them the freedom to build at their own pace and bring on the right clients without the pressure of chasing revenue to cover overheads. After the first year they had employed two staff to handle first-line support. 

⁠Eventually the volume of work made the decision for them. "We took the big scary step of employing ourselves," Sam says.

⁠Fifteen years on, Task Group is a team of seven. The founding logic still runs straight through it, but the company looks nothing like it did at the start. 

⁠In the beginning, cyber security meant antivirus software and a decent router. Online backup was expensive and a hard sell. The job was largely about keeping things running. 

⁠Over the years, all of that changed. Security became non-negotiable, baked into every client's stack. Backup became standard. Monitoring, alerting and reporting replaced gut feel. Automation started handling work that used to take engineers hours. Antivirus is still part of the picture, but it now plays only a small role in a much larger security stack.

⁠Where the company once showed up to fix things, it now builds systems where those things don't break in the first place. And where it once billed for the hours spent doing that, it now bills for the outcomes it delivers.

⁠The clients changed too. For most of Task Group's early years, the brief from a small business was simple: fix my IT, keep things running.

⁠Then, somewhere in the last five years, the questions shifted. Help me reduce risk. Help me make decisions. Help me scale. Help me understand AI without the hype. Business owners who would not have raised security or strategy in an IT conversation were raising them now. The role of the IT partner had quietly grown from keeping the lights on to helping businesses think about where they were going.

⁠That shift was already well underway when AI arrived.

⁠IT services in the AI era

⁠Task Group started experimenting with AI without hesitation. It wasn’t a big transformative rollout.

⁠They started small and kept it internal, starting with using AI for basic tasks like documentation drafting, ticket summaries, scripting assistance, and knowledge base cleanup. From there they moved into automations in workflows, client-facing improvements introduced carefully, and internal GPTs for engineers. The sequence mattered. Internal first. Tested and proven before it touched anything client-facing. The same logic Sam applied in the hospital nearly two decades ago applied directly here.

⁠But the speed at which things could go wrong without the right checks in place caught him off guard.

⁠"AI doesn't fail loudly. It fails confidently, trying its hardest to convince you it's right."

⁠Output that looks right but isn't is a much harder problem to manage than one that throws an obvious error. So Sam built validation layers into the system from day one and stayed hypercritical of every output. It was the same discipline he had applied to every system he had ever built. Understand it, find where it breaks, and build accordingly.

⁠Right from the get go, Sam also made sure the company put guardrails around the use of AI. 

⁠Today, Task Group evaluates every AI tool the same way it evaluates any external supplier it brings into the business. Before anything gets approved, it goes through a vendor risk assessment. There is a defined list of tools the team is permitted to use, and anything outside that list does not get used until it has been properly vetted. No client data goes into publicly accessible AI models without controls in place. Data classification rules govern what can go where. Usage gets logged where possible.

⁠"In practice, it's not a document. It's controls," Sam says.

⁠The criteria for approving a tool are consistent. It needs a clear policy on how it handles data, full GDPR compliance, proper access controls, outputs that are repeatable and reliable, and the ability to connect with the tools the team already uses. Any tool with vague language around data usage gets rejected immediately. So does anything that cannot be properly audited. And if a tool does not save at least 15 to 20 percent of effort on a given task, it is not worth the disruption of rolling it out.

⁠"If we wouldn't trust them with backups, we don't trust them with data. Security isn't a bolt-on here. It's the entry requirement."

⁠That same deliberate approach shaped how Sam brought the rest of the team along.

⁠Adopting AI across the board

⁠Major technology changes rarely go down well with the people expected to use them. Resistance, skepticism and surface-level compliance are the norm. Sam approached it differently.

⁠AI was never mandated, or presented as something that would replace work. He showed how it would take the worst parts of the job off people's plates. Staff were asked to list their most boring, repetitive tasks, and AI was tested against those specific jobs. Tools were optional to begin with. Real time savings were shown rather than promised. Wins got shared every week.

⁠Adoption was never forced. "We put controlled systems in place for them to use and people came along because it worked, not because they were told to use it."

⁠People who choose to use a tool because it makes their day easier are fundamentally different from people who use it because they have been told to. The first group improves the tool. The second group tolerates it. That difference shows up in how well the tools actually get used, and ultimately in the quality of the service clients receive.

⁠Extending AI to customers

⁠With the tools embedded and the team behind them, the conversation with clients followed naturally. For Sam, it has never been a difficult one.

⁠Task Group tells clients exactly where AI plays a role in their service and where it does not, what data it touches and what controls are in place. No softening, no vague reassurances. "We have always found that honesty and clarity is what most people want."

⁠What’s important to note is that clients do not need convincing that AI is safe in the abstract. They need to know that accountability has not moved. Every decision, every action and every outcome at Task Group still sits with a named engineer. Clients deal with the same people they always have, get advice grounded in their specific business, and have someone who takes ownership when something goes wrong. AI runs in the background. As Sam puts it, it is normally something clients never see.

⁠What Task Group does that AI cannot is also what clients rely on most. Understanding the quirks of how a particular business runs. Making judgment calls where there is no obvious right answer. Taking responsibility for outcomes rather than just outputs. Building a working relationship over years that means a client never has to explain their business from scratch. "AI can assist with speed and efficiency, but it can't take ownership, and it can't build trust."

⁠The work and the people behind it

⁠Much of what makes all of this feel effortless from the outside is happening in the background every day without clients ever seeing it.

⁠The team monitors hundreds of data points across every client's systems. Automated scripts fix common problems before anyone needs to raise a ticket. Patches get rolled out across every server and device on schedule. Backups get tested not just to confirm they ran, but to confirm the data can actually be recovered. Security incidents get reviewed in real time. When a client's systems start approaching their limits, capacity planning kicks in before anything slows down.

⁠"If we're doing it right, clients think nothing ever goes wrong."

⁠That is the goal, but it comes with a commercial risk attached. If clients have no window into any of this, they have no sense of what they are paying for. Sam is deliberate about managing that gap. The quiet is not an accident, and clients need to understand that they are the reason it exists.

⁠Keeping that standard up as technology changes around it is something Sam handles as a team effort. Different people gravitate toward different areas. When someone finds something worth paying attention to, it gets shared. Others test it in sandbox environments before it goes anywhere near a client. New tools get tried in real scenarios first. Ideas that hold up get adopted. Those that don't get set aside. No one person carries the burden of keeping up, and nothing reaches clients before it has been properly tried.

⁠The people doing that work have also shifted over time. The early priority was generalists who could handle anything thrown at them. Now Task Group looks for structured thinkers who follow process and know how to work alongside automation. When a gap in the team appears, the instinct is always to develop someone already there rather than bring someone in above them. "Our main focus is about continuing to upskill our current staff and enable them to expand what they can do. We don't want new employees coming in above existing ones where we can train up current staff to meet those roles."

⁠Sam's measure of success has shifted too. In the early years it was about building something that could support a team, and eventually support him and Brigitte. Now it is about stability, predictability and quality of service. Low noise, low chaos. And personally, it is about not being the person the buck always stops with. The more responsibility spreads across the team, the better the business runs, and the more everyone in it benefits from its growth.

⁠Keeping the human in the loop

⁠All of it traces back to a decision that Task Group made long before AI became an industry conversation.

⁠"Human" has been a named core value at the company for years. As AI has become more embedded in how the team works, that value has grown more important, not less. The volume of automated work has increased, but what clients experience has not changed. They still get a fast response from a real person. They still get advice specific to their situation. They still have someone consistent and accountable on the other end. AI handles more behind the scenes, but the parts clients actually notice are still entirely human.

⁠Too many MSPs, Sam says, sell fear, overpromise and under-explain. The ones that will pull ahead are the ones doing the opposite. Treating the human layer as a differentiator rather than an overhead to reduce. Standardizing heavily, using automation for what it is genuinely good at, treating security as a foundation rather than an add-on, and keeping the focus firmly on what clients actually need.

⁠"AI isn't the disruption. Clarity is." The MSPs who can clearly explain what they do, how they do it and why it matters are the ones that will grow, he says.

⁠Sam has been building toward that answer for a long time. The tools have just gotten faster.