A while back I sat across from a business owner who spent twenty minutes telling me about all the people who had let him down. Suppliers who didn’t deliver. Staff who didn’t care. Clients who didn’t pay. Every story had the same shape: he was the reasonable one, and everyone else was difficult. By the end I found myself thinking something I didn’t say out loud — if everyone you meet is the problem, at some point you have to look at the common thread. And the common thread is you.

That’s an uncomfortable idea, so most of us avoid it. It’s far easier to collect grievances than to ask what we might be contributing. But the pattern is the tell. One bad supplier is bad luck. A run of them, year after year, is a signal worth reading.

The pattern is data you already own

Here’s the part people miss. You don’t have to guess at your own patterns anymore — the evidence is sitting in your inbox and your chat history. Every difficult conversation you’ve ever had is recorded somewhere in Microsoft 365. The question is whether you’re willing to look at it honestly.

I’ve started using Copilot in Outlook for this. Not to draft replies, but to hold up a mirror. Ask Copilot to summarise the tone of a thread you found frustrating, and it will describe what’s actually on the page rather than the version you’ve rehearsed in your head. Sometimes the reply that felt firm reads as cold. Sometimes the message you thought was clear was three questions with no answers. That’s a confronting read, but it’s a useful one.

You can go further. Ask Copilot to look across a month of your sent items and describe how you tend to respond when you’re under pressure. It won’t judge you. It just reflects. And what it reflects is the behaviour you bring into the room before anyone else has said a word.

Responsibility is quieter than blame

The point isn’t to beat yourself up. It’s to notice the bits that are yours to change. Maybe you leave replies too long, so relationships arrive at every conversation already frayed. Maybe you set expectations loosely and then feel let down when people fill the gaps their own way. Maybe your Teams messages land harder than you intend because you write them at 11pm.

None of that makes the other person a saint. But it’s the only part you can actually do something about. When I stopped asking “why do I keep dealing with difficult people” and started asking “what am I doing that makes these interactions harder”, the quality of the people around me appeared to improve. They hadn’t changed. My side of it had.

Copilot in Teams is quietly good for this in the moment. Before I fire off something sharp, I’ll ask it to redraft the message so the intent stays but the edge comes off. Same point, less friction. Over time, fewer sparks means fewer fires to put out.

Watch the thread, not the person

I’m not suggesting you’re responsible for every bad actor you meet. Some people genuinely are difficult, and no amount of self-reflection changes that. But if the same story keeps repeating with different faces, the honest move is to check your own part first.

The tools now make that check easy. What’s harder — and always will be — is being willing to read what they show you. That’s the bit I’m still working on, and I suspect I always will be.

A few weeks back I watched someone paste a Copilot answer straight into a client proposal without reading it. Not skimming it — actually reading it. The words went from the screen to the document to the send button in about forty seconds. When I asked why, the answer was telling: “It’s usually right.” Usually. That word did a lot of heavy lifting, and it’s the reason I keep coming back to a simple idea. You need you more than you need the tool.

The tool is fast. You are accountable.

Copilot has changed how quickly I can move. I can sit in Outlook, ask it to draft a reply to a tangled client email, and have something workable in seconds. That’s genuinely useful. But the draft isn’t the decision. The decision is still mine — the tone I want, the thing I choose to leave unsaid, the commitment I’m willing to make on behalf of my business. Copilot can hand me the first version. It can’t hand me the judgement about whether that version is the right thing to send.

I think a lot of people are quietly outsourcing that judgement without noticing. The speed is seductive. When an answer arrives fully formed and reads confidently, the temptation is to assume the thinking has already been done. It hasn’t. The thinking is the part only you can do, and it’s the part that matters most.

Use it to sharpen your view, not replace it

The way I work now, Copilot is a sparring partner, not an oracle. If I’m shaping a position on something, I’ll ask it to argue the other side. I’ll open a messy Excel sheet and ask Copilot to find the pattern I’m missing, then decide for myself whether that pattern actually means anything. In Teams, I’ll have it summarise a meeting I half-listened to, and I still go back and check the bit that felt important. The tool widens the field of view. I’m the one who decides where to look.

That distinction sounds small, but it changes everything about how the work feels. When you treat Copilot as the answer, you shrink. When you treat it as a way to test your own answer, you grow. The people who’ll do best over the next few years aren’t the ones who lean on it hardest. They’re the ones who bring more of themselves to it — more context, more standards, more of a point of view worth defending.

Bring more of yourself, not less

There’s a version of this technology that quietly makes people passive, and there’s a version that makes them sharper. The difference isn’t in the tool. It’s in the person sitting in front of it. Copilot will happily fill any silence you leave. If you show up with nothing, you’ll get something generic and forgettable. If you show up with a real opinion, it becomes an amplifier for that opinion.

So I keep reminding myself, and the people I work with: the tool is a multiplier, and a multiplier does nothing to zero. Your taste, your standards, your read on the situation — that’s the number it multiplies.

What I’m watching now is whether people use this moment to think more or think less. Because the tools will only get better at answering. The rare thing, the thing that stays yours, is knowing which answer is actually worth having.

I have sat through meetings where the proposal looked excellent. The pricing was sensible. The scope seemed clear. Everyone left sounding positive.

Yet something still felt wrong.

I have learned to pay attention to that feeling. Commercial arrangements rarely fail because the Excel formula was slightly imperfect. They fail because one party avoids difficult conversations, changes the meaning of an agreement after the fact, or treats trust as something to spend rather than protect.

A contract matters. Character matters more.

Pressure Reveals the Relationship

I do not really know a business partner when everything is going well. I learn who they are when a project slips, a client becomes frustrated, or an unexpected cost appears.

That is when behaviour becomes evidence.

A trustworthy person raises the issue early, owns their part and helps work through the options, even when the answer is uncomfortable. Someone less trustworthy delays, obscures, redirects blame, or searches for wording that lets them escape responsibility.

I see this in MSP relationships. A new client may appear commercially attractive, but the warning signs often arrive before the agreement is signed. They speak dismissively about their previous provider. They want undocumented exceptions. They push for access before approvals are complete. They treat every boundary as an obstacle to negotiate away.

I used to think a strong process could contain that behaviour. Now I think process can expose it, but it cannot repair it.

Technology Records Decisions, Not Intentions

Microsoft 365 gives me ways to make work visible. I can capture meeting actions in Teams, keep agreed documents in SharePoint, use Planner to assign responsibilities, and ask Copilot to summarise a discussion or draft a follow-up in Outlook.

That visibility reduces ambiguity and gives everyone a shared record.

But none of those tools can supply integrity.

For example, after a client meeting I might ask Copilot in Teams to summarise the decisions, then send the agreed actions through Outlook. A good partner reads that record and corrects anything that is unclear. A bad one may stay silent, then challenge the same agreement later when it becomes inconvenient.

The technology did its job. The problem was never documentation. The problem was whether the people involved respected what the documentation represented.

This is why I do not treat governance as a substitute for judgement. Good governance supports good people. It also limits the damage caused by poor behaviour. It should never become an excuse to ignore obvious warning signs.

The Best Return Is Fewer Regrets

I still examine margin, risk, scope, payment terms and delivery capacity. I would not replace due diligence with instinct. However, I now give equal weight to how people behave during the sales process.

Do they answer direct questions directly? Do they keep small commitments? Do they respect the people who are not in the room? Can they disagree without becoming personal? Do their actions match the story they tell?

Those signals influence the future more than a polished proposal does.

I have found that the right person can help a difficult arrangement recover. They communicate, adapt and work towards a fair result. The wrong person can turn even an attractive opportunity into months of distraction, resentment and avoidable cost.

So I try to choose the relationship before I choose the transaction.

The numbers still need to work. The scope still needs to be clear. The agreement still needs to be written down. But before I commit, I ask myself a simpler question: if circumstances become difficult, do I trust this person to remain honest, reasonable and accountable?

If the answer is no, the spreadsheet is not going to save me.