Excel and one phone number: what companies really run on

Over the past year, something has shifted. Clients no longer come to us with a specification for the system they need built. They come with a question: "Where can you help us?"

At the first meeting, we then learn that the whole company runs on the assistant's institutional memory and on a former marketing guy's willingness to pick up the phone years after he left. That one department carries binders from office to office. That the owner is the only one who knows what the numbers in the reports actually mean. That the same numbers get entered several times, copied by hand, and reality isn't captured anywhere. That when the company takes on more orders, the operational workload grows so much it's barely worth it. That its customers have requirements for its systems that it can't meet, and it doesn't know how to meet them without going under. That the system is fragile and holds growth back instead of supporting it.

In my view, these problems have always been there. They could just be hidden in the margins. When there's a surplus, inefficiency disappears. But the market is changing, and AI is exposing them faster than we can sweep them under the rug.

That's why we stopped talking to clients about technology a long time ago. We talk about their business. Software is just a tool to improve it.

How things really work at your company

The hardest part comes right at the start: how do things actually work at your company? Maybe you have systems. You have an ERP, Salesforce, maybe even SAP, or you once had a custom system built. You have accounting software, an invoicing system, a quoting system. But things still grind. New spreadsheets and overviews keep appearing, and you only see the numbers long after you've closed the month, quarter or year. Or you copy data from one system to another, build your own views of it using your own logic, the inventory system doesn't match your spreadsheet, the process is full of exceptions and a report takes a week. And that's assuming someone knows how things really stand.

And how it all really works is still known only to a few people (best case scenario). They know how things run in real life, which systems have which gaps, what's done by hand and what work is done twice.

New system, new worries

Even so, introducing new systems runs into (perhaps not so surprising) problems. I don't mean the cases where management pretends the market storm and the age of AI don't concern them, and wants to coast to retirement without major changes to the processes they're used to (which, by the way, is perfectly human and understandable). Far more often, in my conversations with owners, directors and managers, I hear concerns they have for good reason.

Will we lose control over our company data?

You've built your system over many years, and every part of it reflects a need that probably surfaced through some problem. Changing it when you don't know exactly how the system works is hard.

If we streamline the process, will we have to let people go? And what if people resist the change?

Every new system has to take into account and preserve what actually works in the company. There's no point building or buying anything new when the old thing works well. Integrating the system and people's knowledge brings more value than a coat of AI polish. Saving on human work doesn't necessarily mean cutting jobs. If the new system is part of business decision-making, restructuring the work can also open up new ways to use your people's capacity. We can help with that.

Will we get burned again? And what if we want to replace the system later?

Project rescue and vendor lock-in are topics we know very well. In our experience, many clients really have been burned before, and a system they put a lot of money and energy into now works against them, or only half works. So caution is definitely warranted. That's also why we say there's no need to replace the whole system right away. Let's look at its most painful part. It can be how long it takes to put together a quote, inaccurate data, paper-based records, anything that, once fixed, can make a big difference to your work. Small step, proportionate effort, big impact. But above all, building trust and proof that things can work.

Won't we become dependent on AI?

What if it gets more expensive? What if it slows down? What if it disappears? What if it breaks? Anything can happen. That's why AI is a tool, not the foundation of a system, and why vibe coding has such a bad reputation. That's also why we decide at the design stage how the model or service can be swapped out and what the system will do if it goes down. A swap may take some work, but it shouldn't require rebuilding the whole solution. In the hands of professionals with development experience who don't skip the basic principles of software engineering, AI is a very powerful and effective tool. In the hands of people without an IT background, it's a ticking time bomb.

Is our data safe?

It's the question I hear most often these days, and definitely the right one. The answer is that you decide: where the data will be stored, who has access to it and what AI is allowed to read at all. Data can stay in your environment, and AI doesn't need to see more than it needs for one specific task. Security is a series of decisions someone has to make before development starts, and we can guide you through them.

So where do you start?

You don't need to know exactly what you need. Just tell us what problem you're dealing with. "Where can you help us?" is our favorite brief.

There's no need to rush. The train hasn't left yet. Set aside startups and frontier tech, and what we've seen this year tells me the potential of traditional companies is far from exhausted.

If you recognized yourself in any of this, you have enough for a first conversation. You don't have to decide upfront whether you need AI, a new system or to connect the ones you already have. Just show us where work gets stuck, which numbers you don't trust or what you'd like to manage but your current operations won't allow.

That's where we can start. Together with your people, we'll find out what's worth keeping, what wastes their time and where it makes sense to make the first change. You know your business. From us, you'll get a simple proposal you can assess, change or test on your own process.

Keep Excel, it's a good tool. Your company just shouldn't depend on it and one phone number to function.