Thoughts

Most teams work with a mix of spreadsheets, emails and one or two systems (usually an accounting software, a CRM, maybe a project tracker).

Work passes from person to person.
Reports are produced regularly.
Numbers pile up.

Excel sits at the centre. Not by accident.
It is fast, reliable, flexible and close to the work.

Data analysis exists, but often arrives after decisions are already made.

As systems multiply, understanding drops.
People cope with judgement, habit and workarounds.

This is not carelessness.
It’s what happens when tools, numbers and decisions fall out of sync.

For instance: Someone asks for last quarter’s numbers by region. You know the data exists somewhere between the accounting export, the sales tracker and another spreadsheet. It takes 45 minutes to piece it together because each source defines ‘region’ slightly differently. By the time you have an answer, they’ve already made the decision.

The pattern described above isn’t neutral. It creates costs that are hard to measure but easy to feel:

⬤ The Friday rush before the board meeting.
⬤ The client question you can’t answer without three phone calls.
⬤ The decision delayed because no one quite trusts the numbers.
⬤ The two months it takes a new hire to understand how things actually work.
⬤ The moment something goes wrong and no one can explain how you got there.

These aren’t technology problems. They’re problems of structure and visibility that technology was supposed to solve but often makes worse.

But there’s an alternative approach. Most businesses already have the technology they need (Excel, existing systems, basic data). What’s missing isn’t another platform or an AI tool. It’s the expertise to connect what already exists to how work actually happens. Custom-built, context-specific solutions that work with your current setup, not against it.

If AI can do the work, why do we still need help?

Many professionals and business owners now use AI daily.

It generates analysis, ideas, writes e-mails, summaries and working logic at a speed that did not exist before. It creates apps and websites, writes complex Excel formulas and advanced code in any language within seconds. It analyzes datasets and produces charts faster than any human could.

That scale and speed is real leverage. But leverage still requires understanding, patience and effort.

In reality?

It produces different answers each time you ask the same question. Sometimes because the system updates. Or simply because it interprets the question differently. Leading to hours spent rephrasing prompts, double-checking outputs, wondering if you can trust it.

All while preparing for client meetings, running the business and deciding what to do next.

When you ask AI for a tech solution, you get something that looks right (and usually it is). It might even work for a few weeks. But what do you do when it breaks in a way you don’t understand, at month-end when you’re already stretched? You suffer the consequences.

So is it really about whether AI can do the work? No. The real question is whether your team has the foundational understanding to use AI effectively, or whether they’ll spend months building beautiful solutions to the wrong problems.

Used badly, AI doesn’t save time. It consumes attention, energy, reputation and skills.
Used well, it can be impactful. It can truly support processes and operations.
But it should never replace understanding.
Or thinking.

How do I know if a tool is actually needed?

Look at where people spend time fixing the same problem repeatedly.

If the same issue appears across different people (for example the sales team, finance and operations all struggling to get a consistent customer list) it points to a structural problem. The data exists in multiple places with no unified logic.

If the same issue appears in the same situation repeatedly (for example every month-end close takes an extra day because two systems don’t align) it points to a process that needs documentation, clearer handover or proper integration.

If people have built elaborate workarounds (like exporting from three systems, running manual checks in Excel, then re-entering data elsewhere) that’s a signal the current setup isn’t working.

The tool is needed when the pattern is clear and repetitive. Not when the problem is vague or one-off.

What would still work if a tool disappeared tomorrow?

Run this test: imagine the software is unavailable for a week.

Can you still:
⬤ Serve your customers?
⬤ Make basic operational decisions?
⬤ Explain to someone how the business works?
⬤ Access the core information you actually need?

If yes, the tool is support. Useful and not a significant operational risk.

If losing access means you can’t function or can’t explain what you’re doing, you’ve outsourced critical thinking. That’s dependency, not support.

The answer should be: everything that matters still works. Maybe slower, maybe manually, maybe less efficiently, but it works.

If the software went offline and you couldn’t make basic decisions or serve customers, you’ve let a tool become a substitute for understanding your own business.

Always have a contingency plan in place.

What should Excel be used for and where does it stop being the right tool?

Excel is not a temporary fix. It is ideal for building the smallest setup a business truly needs.

It works best for:
⬤ keeping critical business logic in-house and transparent
⬤ analysing data, testing assumptions and modelling scenarios
⬤ prototyping ideas and processes before committing to expensive software
⬤ explaining how the business works to other people
⬤ serving as the universal thinking environment for collaborative problem-solving
⬤ being the “lingua franca” that aligns technical and business teams

Excel struggles when:
⬤ data size exceeds a few hundred thousand rows (the file takes over a minute to save or crashes regularly)
⬤ you need data to flow automatically between systems without manual export and import
⬤ multiple people need to edit the same file simultaneously
⬤ knowing who changed what, when and why matters
⬤ different users must have different access levels
⬤ complex quantitative techniques are required

We have data scattered across spreadsheets, emails and systems. Should we buy new software?

Not yet. Start with one weekly task first like a report, an update, a regular process or a simple decision.

Where does its information come from?
How is it checked?
Who depends on it?

If these can be pulled together reliably with what you already have, software is not the issue.

We spend hours producing the same report. Is automation worth it?

Check what happens after the report is sent.

If it sits untouched, don’t automate it. Stop making it.
If it informs action, automation helps.

We have years of data but don’t know what to do with it. Should we hire an analyst?

Only once you know what question you’re trying to answer.

Ask yourself: what decision would I make differently if I understood this data better? If you can name it, like ‘which product line to phase out’ or ‘which customer segment to focus on’ then yes, hire someone. If you can’t, hire someone who can help you name it.

Without clear questions, analysis creates volume. Not value.

We track a lot of numbers. How do we know which ones matter?

Try this: stop updating half of them for a month.

Notice which numbers people actually ask for when something feels wrong.

Those are the ones that matter.

Why do teams avoid tools that were meant to help?

Because the tool adds work before it removes work.

In a busy day, people default to what they already know. Even if imperfect.

Everyone talks about being “data-driven”, but judgement still decides. Why?

Because data only matters when it reaches the people who carry responsibility and understand the business.

Analytics supports judgement. It does not replace it.

Why does some change stick?

Change sticks when it removes a frustration people already feel.

When frustration disappears, people follow without being pushed.

You have a dashboard someone built last year. Fifteen charts, forty metrics.

You glance at it Monday mornings but make decisions based on conversations with your team or your own feel for what’s happening.

That’s not the problem. The dashboard failed because it wasn’t built from how you actually think.

Good analysis would have started with those conversations: what you already pay attention to, what makes you worried, what changes your mind  and worked backward to what should be measured.

Most everyday problems are not caused by missing tools.

They come from unclear ways of working, slow feedback and tasks that are hard to explain or hand over.

Excel-based, custom-built solutions matter because they live inside the work.
They make assumptions visible and trade-offs easier to see.

Data analysis does not remove uncertainty.
Uncertainty is always there.

What good data analysis does is help people make decisions while facing it.

It helps you understand what you actually know, what you’re guessing and what would change your mind.

Before adding new software, three things need to be in place:
⬤ How the work actually flows (who does what, when and what they need)
⬤ What assumptions you’re making (so you can test if they’re true)
⬤ How decisions were made (so you can explain them later or learn from them)

Tools earn their place once things are clear.
They fail when they’re expected to do the thinking.