First the user, then the process

Two years ago, we already saw a shift coming: more and more work is being driven from a prompt rather than via fixed screens and menus. Major software vendors are now moving more emphatically in that direction.

That raises a question I find interesting. If users increasingly determine for themselves how they view information and work with processes, what is the value of a standard interface?

I expect personalized work environments to become more common: an environment that adapts to someone's role, tasks, and needs—a kind of digital tailor-made suit. That is an expectation, not a fact. Fixed interfaces also retain their value. They provide an overview, they are recognizable, and a team is looking at the same thing. I expect a hybrid form, and the ratio varies by organization.

You can already see a small version of this at Claude. Ask for a simple tool, such as a calculation aid or a form, and you will have a working application as an artifact within minutes. It is not a business system, but it shows how low the barrier to development is becoming.

What I have learned

I recently completed the Zoho AI Accelerator training, and MFORZ can now call itself an AI Enabled Zoho Partner. The training confirmed the direction we already saw. It added a sharper way to start: with the user. By user, I mean both the employee working with the system and the customer interacting with it. Who is this person, what does their day look like, and what do they need to get done?

Looking at users was not new to us. What is changing is how early that question comes and how much it determines the outcome.

In the training, the connection to the backend came last in the build sequence. I would like to add a caveat to that. You can build the connection itself later, but it is better to investigate early on whether the data is available, who is allowed to view and modify it, and whether the connection is technically feasible. MCP is one way to connect systems. It does not answer those questions automatically.

What we are starting with now

For years, my initial questions regarding a new project revolved around processes and data. What information is available? How does the process run? What can we automate? Those questions remain. But now I tend to ask a different one: how does someone want to work, and what do we need to make available for that? The answer reveals which processes and which data matter.

A pilot: from photo to quotation preparation

We are running a pilot to investigate how a photo can help prepare a quote and reduce manual work. It is a pilot, and we do not have any results to share yet.

A photo does not contain everything a quotation requires. Information is missing, for example, things only the customer or the person on site know, so someone needs to supplement or confirm this. A person also checks the quotation before it is sent out. We are looking at how subsequent steps, such as the order confirmation, the invoice, and the work order, can build upon the same information. Whether we develop this further depends on how complete the information proves to be, how much correction is needed, and whether the people working with it notice that it saves time.

For the employee on site, the action is simple: taking a photo. The work behind it is not.

The work behind that one photo

A simple action only works if the work behind it is in order. That means clear processes, reliable data, defined authorizations, checks at the right moments, and a plan for when something goes wrong. Security is part of that. The guidelines from the training were practical: grant only the necessary rights, treat a server URL like a password, and test in a sandbox before going to production.

Attention is also needed after delivery. A system that is live is not yet in use. The first few weeks count, and onboarding is part of the job.

What this means for you

The first conversation is about how you currently work and how you want to work, not about modules and screens. From there, we look at what data is available, what people are allowed to do, and where someone needs to check or make decisions.

AI makes building easier. Determining what to build, how to securely connect systems, and who controls what is becoming more important, and you need to build that control in from the start.

AI builds. Humans remain in control.

Would you like to explore what this might look like for your team? Schedule a meeting.