Stop Holding People Accountable. Start Building Momentum.

There’s a conversation I’ve watched play out a hundred times, and it always goes the same way. Someone agreed last week to get a thing done. This week, they didn’t. So the check-in begins. Did you finish it? No. Why not? And then comes the explanation — the kids were sick, a client blew up, the week got away from them. By the end they feel small, the task gets shuffled forward, and everyone agrees to try again next time.

Then we run it back. Same script, same shrinking feeling. And here’s the part nobody admits out loud: the simplest way to make that bad feeling stop is to walk away from the whole arrangement entirely. People don’t quit because they’ve failed. They quit because the relationship has quietly become a source of guilt.

I see this with MSP owners constantly. They build accountability structures for their teams, their clients, even themselves — weekly stand-ups, status trackers, a long list of overdue commitments staring back at them. And the structure, the thing meant to drive progress, becomes the thing people learn to dread.

Accountability is a low ceiling

The uncomfortable truth is that accountability only works on people who don’t really need it — and even then, only for a while. It’s a tool built around the assumption that someone has to be watched or they’ll drift. Chase the high performers in your business with that energy and watch how fast they disengage. The good ones already hold themselves to a higher standard than you ever could. They don’t want a minder checking their homework. They want to get somewhere, faster.

So I stopped asking whether things got done. When I sit down with someone now — whether it’s a one-week review or a full six-week stretch — I ask two questions only. What went well? And what did this teach you? That’s it. Everything else from the past is noise. The misses, the excuses, the guilt — none of it moves you an inch forward. The wins tell you what to repeat. The lessons tell you what to change. The rest is dead weight you’re dragging into next week for no reason.

From rear-view to road ahead

Once those two things are on the table, the conversation flips direction. We’re no longer auditing the past — we’re using it. We find the one obstacle actually slowing the person down, we clear it out of the path, and we decide the very next move. Look back only long enough to learn, then point everything forward.

This is where the tooling matters more than people expect, because most of the friction that kills momentum is small, dumb, and repeatable. Someone “forgot” to follow up because the note lived in their head instead of anywhere useful. A commitment vanished because it was never written down where the team could see it. That’s not a discipline problem. That’s a system problem, and systems are fixable.

Inside Microsoft 365, this gets a lot easier. When I close out a planning session in Teams, I’ll ask Copilot to pull the wins and the agreed next moves straight out of the meeting transcript, so nobody’s relying on memory by Friday. The actual commitments go into Planner as real tasks with owners and dates — visible, shared, alive — not buried in someone’s inbox. A Loop component pinned in the channel becomes the running scoreboard the whole team can glance at. And first thing in the morning, Copilot in Outlook surfaces what’s drifted and what’s due, so the next move is obvious before the day even starts. None of that is about catching people out. It’s about removing the friction that quietly steals forward motion.

Speed, not surveillance

This is the shift I’d put to any MSP leader still running their team on guilt. The higher someone performs, the less they need policing and the more they crave progress. So give them progress. Clear the road, name the next step, and get out of the way. Momentum compounds — one clean win makes the next one easier, and before long the person isn’t being pushed at all. They’re pulling.

If what you want is someone standing over your shoulder reminding you of everything you haven’t done yet, I’m honestly not the right person for that. But if you want the obstacles gone, the path clear, and the next move obvious so you can win quicker — that’s the whole job. Stop counting what’s overdue. Start building what’s next.

The Business Doubled When I Started Cutting, Not Adding

For years I ran my business trying to make it look impressive. Impressive to peers, to prospects, to whoever happened to be watching at a networking event. Every new tool, every new process, every clever workaround was another thing I could point to and say, look how sophisticated this is. The problem is that nobody scales a business off how it looks from the outside. They scale it off how it actually runs on a Tuesday afternoon when three clients call at once.

Somewhere along the way I stopped building for the audience and started building for me. For how the work felt to do. That single shift is the closest thing I have to a real explanation for how we went from eight million a year to twenty.

The mess was self-inflicted

Here’s the uncomfortable part. The complexity that was strangling the business wasn’t forced on me by clients or by Microsoft or by the market. I built it. Every duct-taped process was a decision I made at some point to patch a problem rather than fix it. A spreadsheet here to track what a proper system should have tracked. A manual checklist there because nobody trusted the automation. A Teams channel for this, a separate one for that, a third one nobody remembered the purpose of.

None of it was wrong on the day I added it. It was all reasonable in isolation. But reasonable decisions stacked on top of each other for five years become a structure that no one person can hold in their head. And when no one can hold it in their head, everything slows down. People stop deciding and start asking. The business gets heavier with every fix.

Why adding more is the instinct, and the trap

When things feel chaotic, the instinct is to reach for another tool. New ticketing platform. New project board. Another layer of approval to stop the mistakes. It feels like progress because you’re doing something. But you’re usually just adding another joint to a structure that already has too many.

I had to break that habit in myself first. The question I started asking wasn’t “what can I add to fix this” but “what can I remove so this stops happening at all.” Removal is harder. It means admitting that something you built no longer earns its keep. It means killing the spreadsheet you were quietly proud of.

This is where I leaned on the Microsoft 365 stack properly rather than around it. We were running four overlapping trackers, so I had Copilot in Excel pull them apart and show me where the same data lived in three places. Then we cut it to one source of truth in SharePoint and let Copilot answer the questions people used to dig through the others to find. The chat channels got the same treatment. Half of them went. The ones that stayed got a clear job, and when someone asked “where does this go,” the answer was finally obvious.

Simplicity is a discipline, not a milestone

The thing nobody warns you about is that simplicity doesn’t stay. It’s not a state you reach and then relax. Complexity creeps back the moment you stop watching, because every small patch feels harmless in the moment. So now I run a regular review where the only acceptable change is a subtraction. What process can we retire. What approval step is just fear wearing a hat. What report does nobody actually read. I use Copilot to summarise where time is going across the team’s Outlook and Teams activity, and more often than not it points straight at something we could stop doing entirely.

That review is the most valuable hour in my month. Not because it adds capability, but because it protects the lightness that let us grow in the first place.

The takeaway

If your business feels stuck and heavy, resist the urge to bolt on one more thing. You almost certainly don’t have a tooling gap. You have an accumulation problem, and you built the accumulation yourself, one sensible patch at a time. Growth didn’t come to me from being more elaborate. It came from being willing to cut, to trust the simpler version, and to stop caring whether it looked clever to anyone else.

The hard part isn’t knowing what to remove. It’s having the nerve to actually pick up the scissors.

The Answer Is Never "More"

Every business owner I talk to has the same instinct when something isn’t working. Add something. Not enough leads? Launch a new offer. Service slipping? Hire another person. Growth flat? Spin up a new program, a new campaign, a new line of work. Adding feels good. It feels like progress. It costs almost nothing in the moment, and it gives you something to point at when someone asks what you’re doing about the problem.

The trouble is that adding is the cheap part. Keeping the thing alive afterwards is where the real bill lands.

Adding is light, owning is heavy

A new offer takes an afternoon to design and an email to announce. But now it needs pricing, delivery, support, a place in your onboarding, a line in your reporting, and a person who owns it when it breaks. The new hire fills a gap on Monday and becomes a salary, a review cycle, and a management responsibility by Friday. None of that weight shows up on the day you make the decision. It accumulates quietly, one addition at a time, until you’re carrying a load you never consciously agreed to.

I see this constantly in the MSP world, because our whole industry is built on saying yes. A client asks for one more thing, and it’s easier to absorb it than to have the awkward conversation. Multiply that across a few years and a few dozen clients, and you end up with a service catalogue nobody can fully explain and a delivery model held together by the memory of whoever happens to still work there.

How the tangle gets built

Nobody sets out to build a mess. It happens in sensible-looking steps. A bolt-on tool here. A side agreement there. A special process for one customer that quietly becomes the process for everyone. A department that exists because two years ago a problem needed an owner. Each layer made sense on its own. Stacked together, they form something heavy and strangely fragile, and at some point you realise you’ve become the load-bearing wall. Pull you out and the whole thing sways.

That’s the part that catches people off guard. You set out to build a business that runs without you, and instead you’ve built one that can’t run at all unless you’re sitting in the middle of it, holding the seams together.

Subtraction is a skill worth practising

The harder, more valuable move is taking things away. Killing the offer that earns little and costs plenty. Retiring the program that three people use. Saying no to the request that doesn’t fit, even when yes would be easier today. Subtraction rarely feels like progress in the moment, which is exactly why so few owners do it. But a smaller, sharper business is far easier to run, sell, and live inside than a sprawling one.

You can’t subtract what you can’t see, though, and most owners genuinely don’t know what they’re carrying. This is where I’ve found Microsoft 365 quietly useful — not as a fix, but as a way to make the weight visible. Ask Copilot in Teams to summarise every service, project, and recurring commitment mentioned across your channels over the last quarter, and you’ll get an honest inventory of what your business is actually maintaining versus what you think it is. Point Copilot at your shared mailbox and ask which client requests keep recurring and which offers nobody has touched in months. The answers are usually uncomfortable, and that’s the point.

From there, a single SharePoint page or a Loop component listing every active offer, tool, and program — with an owner and a last-reviewed date against each — turns “we should clean this up someday” into something you can actually work through. I’ve had owners use Copilot in Excel to model what dropping their two worst-performing services would do to revenue and to workload, and the workload relief almost always dwarfs the revenue hit. That’s the number that frees you to cut.

The lighter business wins

The goal isn’t a bigger business. It’s a business light enough that you can still lift it. Every time you reach for “more” as the answer, it’s worth pausing to ask whether you’re solving the problem or just burying it under another layer you’ll have to carry next year.

The best thing I’ve removed from my own business cost me nothing to delete and gave me back hours every week. Adding will always feel easier. Removing is what actually sets you free.

AI Doesn’t Care About Your MSP Business Model

I’ve been thinking about a question that might make some MSPs uncomfortable.

Should the next technology revolution be guided by the same people who have spent the last twenty years building businesses that depend on things staying exactly the way they are?

That sounds harsh, but hear me out.

Every major technology shift creates winners and losers. Cloud changed the value of servers. Microsoft 365 changed the value of maintaining Exchange on-premises. Remote monitoring and management changed the need for technicians driving between customer sites. AI is now doing the same thing to knowledge work.

Yet when I talk to some MSPs about AI, I often hear the same response.

“It’s not ready.”

“Our clients aren’t asking for it.”

“We’ll wait and see.”

“They still need us to do the basics.”

Maybe. But that’s exactly what people said about cloud computing.

The Incentive Problem

The challenge isn’t that MSPs don’t understand technology. Most understand it very well.

The challenge is incentives.

If you’ve built a successful business around selling licences, managing devices, responding to support tickets and charging for technical expertise, then AI creates a problem. The more capable AI becomes, the more it starts consuming activities that have traditionally generated revenue.

If a Microsoft 365 Copilot prompt can analyse a meeting, draft a proposal and create an action plan in minutes, what happens to the hours previously spent doing that work?

If an AI agent can resolve basic support requests, summarise conversations and retrieve information from SharePoint or Teams, what happens to the service model built around handling those requests?

It’s understandable that some MSPs look at this and feel uncomfortable.

But technology has never cared about existing business models.

History Doesn’t Reward Defenders

I’ve seen this pattern before.

Businesses often spend more energy trying to protect the old revenue stream than understanding the new opportunity. They look at disruption as something to resist rather than something to harness.

The problem is that customers rarely share that loyalty to the old model.

Business owners don’t wake up in the morning hoping to buy more support hours. They want outcomes. They want faster decisions. They want less administration. They want their staff spending more time serving customers and less time managing information.

Today, a growing number of those outcomes can be delivered through AI.

Whether MSPs like it or not.

The Risk of Becoming the Blocker

One of the dangers for traditional MSPs is becoming the organisation that says no.

No, you shouldn’t use AI yet.

No, you shouldn’t change your process.

No, your existing approach is good enough.

Meanwhile, another adviser walks in and shows the client how Microsoft 365 Copilot can reduce the time spent preparing for meetings in Teams, drafting emails in Outlook and analysing information in Excel.

Guess which adviser the client sees as helping them move forward?

I’ve always believed that our role as technology professionals is not to preserve the status quo. It’s to help customers navigate change safely and effectively.

That doesn’t mean blindly embracing every new feature that appears. Governance still matters. Security still matters. Data quality still matters. AI without preparation can create as many problems as it solves.

But that’s very different from pretending the change isn’t happening.

The MSP Opportunity

Ironically, AI may create one of the biggest opportunities MSPs have seen in years.

Most businesses have no idea how to prepare their data, secure their environment or establish the governance needed to use AI safely. They need guidance. They need strategy. They need trusted advisers.

That’s where MSPs can create enormous value.

But only if they’re willing to evolve.

The conversation can’t stay focused on devices, licences and tickets. It needs to move towards business processes, information management, knowledge discovery and AI readiness.

That’s where the real opportunity sits.

Final Thoughts

I don’t think AI should be entrusted solely to old-school MSPs who want everything to stay the same. Equally, I don’t think businesses should rush headlong into AI without experienced technology advisers.

The answer lies somewhere in the middle.

The MSPs that thrive over the next decade won’t be the ones defending the past. They’ll be the ones helping clients prepare for the future.

Because AI isn’t interested in protecting anyone’s business model.

It’s simply moving forward.

And the question every MSP needs to answer is whether they’ll be leading that change or explaining to clients why they missed it.

The Hard Part Isn’t Adding Anymore

I caught myself last week with eleven browser tabs open, three half-built automations, a new Planner board nobody asked for, and a Teams channel I’d spun up that morning and already forgotten the purpose of. None of it was hard to create. That was the problem. Every one of those things took me about ninety seconds, and ninety seconds is now the entire cost of bringing something new into existence.

For most of my working life, the brake on doing more was effort. You wanted a new report, a new process, a new client deliverable — and the friction of actually building it slowed you down enough to ask whether it was worth it. That friction is mostly gone. Copilot drafts the document, the deck, the email sequence. Power Automate wires the workflow. The wall that used to stop me from acting on every passing idea has quietly come down.

Speed without judgement is just mess

I’ll be honest about the wiring in my own head. I’m wired to react. A thought lands, I want to act on it immediately, and for years that instinct ran straight into the resistance of having to do the work by hand. The work was the filter. It forced a pause.

Now there’s no pause. The reaction and the execution have collapsed into the same moment. I think “we should have a dashboard for that” and Copilot in Excel has one in front of me before the idea’s even finished forming. Multiply that across a week and you don’t get a sharper business. You get a cluttered one. More dashboards nobody reads. More channels nobody checks. More automations quietly firing into inboxes that have stopped paying attention.

The skill that used to matter was production — the ability to make the thing. That’s not scarce anymore. Anyone with a Copilot licence can produce. What’s scarce now is the judgement to know which of the ten things you could build is the one that actually matters, and the discipline to let the other nine die.

Taste is the work now

I keep coming back to the word taste, because I can’t find a better one. It’s the ability to look at everything you’re capable of generating and recognise what’s worth keeping. It’s hearing the difference between a real signal and noise dressed up to look like progress. And it’s having the spine to cut — to delete the document, kill the channel, switch off the automation — even though it cost you almost nothing to make and feels wasteful to throw away.

This is uncomfortable, especially for those of us who measure ourselves by output. Cutting feels like the opposite of productivity. But a business that adds endlessly and never subtracts doesn’t get faster. It gets heavier. Every new thing you create has a maintenance cost, an attention cost, a “what was this for again?” cost that lands months later. The Loop pages multiply. The SharePoint sites pile up. The thing you built in ninety seconds takes a year to quietly rot in front of everyone.

So I’ve started running my week the other way around. Instead of asking Copilot what else I could build, I ask it what I should stop. I’ll point it at a Teams channel and ask for a summary of the last month — and if the honest answer is “three automated posts and no human replies,” that’s my cue to shut it down. I use Copilot to find the dead weight, not just to make more.

The momentum problem

Here’s the part I want to be straight about, because it’s where I think a lot of us are stuck. You can feel that your best work is still in front of you. The ambition is real. But the momentum you’re chasing won’t arrive while the business is carrying this much weight. Every unfinished idea, every half-used tool, every process you bolted on and never removed is drag. You can’t accelerate something this loaded, no matter how fast you’re able to add to it.

The temptation, when you feel slow, is to add more — a new tool, a new system, a new initiative to fix the sluggishness. That instinct is exactly backwards. The way through isn’t more production. It’s subtraction. The lightness comes from what you’re willing to remove.

It’s never been easier to make something. Which means the rare, valuable, genuinely hard skill is no longer making — it’s choosing. Less, but the right less. That’s the whole game now.

The Report You Should Be Checking Every Month

I’ve lost count of how many times I’ve seen organisations proudly tell me they’ve automated patch management, only to discover they have no idea whether the updates actually made it to the devices.

Getting updates deployed is only half the job. Knowing what happened afterwards is where the real value lies.

That’s why I think one of the most overlooked additions to Windows Autopatch is its reporting capability. Microsoft has invested heavily in giving administrators visibility into both quality updates and feature updates, yet many people still seem to view Autopatch as a simple “set and forget” service. It isn’t. It’s a managed update service that still needs oversight. [learn.microsoft.com]

In my experience, the organisations that get the most value from Windows Autopatch are the ones that spend a few minutes each month reviewing the reports rather than assuming everything worked perfectly.

Compliance Is Not the Same as Configuration

When I speak with MSPs and SMBs, I often hear a variation of the same story.

“We’ve got update policies configured in Intune.”

That’s great, but having a policy isn’t proof that devices are patched.

A device can be powered off, have a failed update, miss a reboot, or simply stop checking in. The policy might be configured perfectly, yet the endpoint remains vulnerable. That distinction matters. In fact, it came up recently in a discussion about the importance of validating that updates have actually reached devices rather than relying on configuration alone.

Windows Autopatch reports help bridge that gap by showing what has actually happened on the endpoint rather than what should have happened.

Visibility at Multiple Levels

The quality update reporting in Windows Autopatch provides several different perspectives. There is a summary view that gives an organisational snapshot, a device-level status report that drills into individual machines, and a trending report that shows update progress over time. Microsoft states that these reports are designed to provide insight into readiness, update health, alerts, compliance, and update status trends over the previous 90 days. [learn.microsoft.com]

That combination is important.

A dashboard might tell you that 95% of devices are compliant. Useful information, certainly. But the remaining 5% are often where the interesting conversations happen.

Which devices failed?

Why are they behind?

Have they stopped checking in?

Do they belong to a key executive, a remote worker, or a critical system?

Those are the questions that reduce risk.

Better Conversations with Copilot

One area I think many organisations overlook is how these reports can work alongside Microsoft 365 Copilot.

Imagine exporting your Windows Autopatch status data into Excel and then asking Copilot questions such as:

Which devices have failed their latest quality update?
Summarise update issues by department.
Identify devices that haven’t checked in recently.
Explain the trend in update compliance over the last quarter.

Rather than manually analysing thousands of rows, Copilot can help surface patterns and priorities much faster. The update data becomes more than just a compliance report. It becomes a decision-making tool.

That’s where I see real value emerging. The reporting tells you what happened. Copilot helps you understand what you should do next.

Reports Help During Audits Too

Anyone who’s been through a security assessment, cyber insurance review, or customer audit knows that “we patch our systems” isn’t usually enough.

You’ll often be asked to demonstrate patch status, prove compliance, explain exceptions, and show evidence of remediation efforts.

The Windows Autopatch reporting framework provides exactly the sort of information auditors tend to request, including device status, readiness information, alerts, compliance data, and historical trends. The data can also be exported for further analysis and reporting. [learn.microsoft.com]

That means you’re not scrambling to produce evidence when someone asks the question.

The evidence is already there.

My Recommendation

If you’re already using Windows Autopatch, add a recurring monthly task to your calendar.

Open the reports.

Review the summary dashboard.

Look for failed devices.

Investigate alerts.

Check the trend lines.

Even better, use Copilot in Excel to help analyse the exported data and identify patterns you might otherwise miss.

Patching isn’t finished when Microsoft releases the update. Patching is finished when you can prove the update successfully reached the devices you’re responsible for.

Windows Autopatch helps automate deployment.

The reports tell you whether that automation is actually working.

Need to Know podcast–Episode 368

Latest news and updates from the Microsoft cloud from security, to copilot and beyond. I also share my thoughts on recent changes at Microsoft, Cowork costs, security challenges with agents and my experiences configuring local AI. Always keen to hear your thoughts and feedback on the content in this episode.

Brought to you by www.ciaopspatron.com

you can listen directly to this episode at:

https://ciaops.podbean.com/e/episode-367-goldilocks-gone/

Subscribe via iTunes at:

https://itunes.apple.com/au/podcast/ciaops-need-to-know-podcasts/id406891445?mt=2

or Spotify:

https://open.spotify.com/show/7ejj00cOuw8977GnnE2lPb

Don’t forget to give the show a rating as well as send me any feedback or suggestions you may have for the show

Resources

CIAOPS Need to Know podcast – CIAOPS – Need to Know podcasts | CIAOPS

X – https://www.twitter.com/directorcia

director@ciaops.com

CIAOPS Blog

Join my Teams Shared Channel – CIAOPS

CIAOPS Merch store – CIAOPS

Become a CIAOPS Patron

CIAOPS AI Dojo

CIAOPS weekly news update – CIA Brief – CIAOPS

CIAOPS Labs – The Special Activities Division of the CIAOPS

Support CIAOPS

Get your M365 questions answered via email

Join my email list

A special thanks to the CIAOPS Patron community for making this podcast possible. You can find the benefits of a subscription to the community and become a member at https://www.ciaopspatron.com

Security & Threat Intelligence

Microsoft 365 Copilot & AI

SharePoint & Content Management

Microsoft Corporate News

Microsoft cuts 4,800 employees in new layoffs

The Real Power of Copilot in Excel Isn’t Formulas. It’s Repeatability.

The first thing most people do when they open Copilot in Excel is ask it for a formula.

That’s understandable. For decades, Excel expertise has often been measured by how quickly someone can build complex formulas, create PivotTables, or untangle messy spreadsheets. Copilot changes that. You can now describe what you want in plain English and let the AI do much of the heavy lifting.

But after looking at the latest capabilities around Copilot Skills in Excel, I think many people are missing the bigger opportunity.

The real value isn’t that Copilot can generate a formula.

It’s that it can help you repeat a proven process over and over again.

In most organisations, especially SMBs and MSPs, there are a handful of people who know how to make Excel sing. They’re the people who understand financial models, reporting structures, forecasting tools, and data analysis techniques. Everyone else tends to rely on them whenever something complicated appears in a workbook.

That creates a bottleneck.

I’ve seen it countless times. Month-end reporting arrives and everyone waits for the same person. Quarterly forecasting needs updating and the same expert gets involved again. A new staff member arrives and spends weeks learning spreadsheet processes that only exist in somebody’s head.

That’s not really an Excel problem.

It’s a knowledge-sharing problem.

Copilot Skills in Excel feel like Microsoft’s attempt to address exactly that issue. Rather than repeatedly explaining the same process, you can package those instructions into a reusable skill that Copilot can invoke when required. According to Microsoft’s documentation, skills allow Copilot in Excel to perform repeatable tasks using predefined instructions, and organisations can create custom skills stored in OneDrive for reuse. [support.mi…rosoft.com], [support.mi…rosoft.com]

That might sound like a small change, but I think it’s significant.

From Spreadsheet Expert to Process Expert

Imagine you’ve spent years refining a monthly reporting workbook.

You know exactly how the raw data is imported. You know which columns need cleaning. You know which calculations matter and which charts management expects to see.

Traditionally, every new employee needed training. Documentation had to be updated. Mistakes inevitably crept in.

Now imagine creating a skill that guides Copilot through that process.

Instead of asking a colleague to remember twenty separate steps, they simply invoke the skill and allow Copilot to perform the work in a consistent manner.

The expertise becomes transferable.

That’s a very different proposition from merely generating formulas.

Excel Is Becoming More Conversational

Something else strikes me here.

For years, becoming good at Excel meant learning Excel’s language. You memorised formulas, syntax, functions, and workarounds.

Copilot flips that model.

Now Excel is increasingly learning your language.

You can ask questions about data. Request analysis. Generate charts. Create reports. Import information. Explain formulas. Build dashboards. And increasingly, define repeatable business processes using natural language instructions. [support.mi…rosoft.com], [support.mi…rosoft.com]

That’s a major shift.

The barrier to entry drops dramatically.

People who previously avoided advanced spreadsheet work now have a capable assistant sitting beside them.

The Opportunity for MSPs

From an MSP perspective, I think this capability will become particularly interesting.

Most advice around AI focuses on content generation, meeting summaries, or email drafting. Those are valuable, but they’re often incremental productivity gains.

Skills have the potential to standardise operations.

Imagine creating repeatable Excel processes for:

Monthly financial reporting
Customer profitability analysis
Service desk trend reporting
Project forecasting
Licence consumption tracking
Security compliance reporting

Rather than documenting procedures in lengthy manuals, organisations can embed that knowledge directly into skills that guide Copilot.

That’s a much more scalable approach.

And importantly, it helps preserve organisational knowledge when staff move on.

The Human Still Matters

Of course, none of this removes the need for human oversight.

One lesson I’ve repeated many times with Microsoft 365 Copilot is that AI works best when it’s treated as a capable assistant, not an autonomous decision maker.

If Copilot analyses data, review the results.

If it creates a forecast, validate the assumptions.

If it generates a report, make sure the conclusions make sense.

The person remains accountable. Copilot simply removes much of the repetitive effort.

Final Thoughts

When people think about AI in Excel, they often focus on saving a few minutes creating formulas or formatting data.

That’s useful.

But I think the more interesting story is the ability to capture expertise and make it reusable.

The organisations that benefit most from AI won’t necessarily be those with the smartest prompts. They’ll be the ones that systematically turn repeatable knowledge into repeatable processes.

Excel Skills look like another step in that direction.

And for many businesses, that could end up being far more valuable than any individual formula.