
Vibe coding at companies: the prototype takes an afternoon, the hard part comes after
When anyone can build an app, the question changes.
Services
AI Solutions DevelopmentCustom AI TrainingsAI Workspace SetupIndustries
ManufacturingEcommerceHR AgenciesForeign TradeAccounting FirmsMarketing AgenciesInsuranceHealth ServicesLegal & NotaryReal EstateFinancial ServicesLogisticsThe plans that try to change everything at once tend to be the ones that change the least.

In almost every company with some history there comes a moment when someone says "we need to do digital transformation." What usually follows is a big plan, with many areas, many systems and a horizon of years. In a traditional company, what delivers the most is the opposite: pick a process that hurts the business today, change it end to end in a few weeks and use that first result to decide what comes next.
The phrase is used for almost anything, from buying an ERP to launching an online store. In a company that has been around for years, in practice it means something more concrete: that everyday work stops depending on paper, loose spreadsheets, double data entry and phone calls to find out the status of each thing.
That kind of company has an advantage new ones lack. It already knows how its business works, knows its customers and has processes that, with their quirks, run. The transformation starts from all of that.
According to a regional report cited by Revista Mercado, only one in three digital transformations in Latin America achieves what it set out to do. Among the most frequent causes are two that have little to do with technology: projects that start without a clear goal, and changes nobody supported within the team.
Behind those numbers there is a familiar pattern. A broad plan is defined, a tool is chosen, months go into configuring it and, when the time comes to use it, the way of working stays the same. The new system lives alongside the old spreadsheet and, before long, the spreadsheet wins.
In companies with small IT teams, or with people who have been there for many years, there is one more brake: everything goes through the same few people, who also keep operations running. Projects drag on, priorities shift and the initiative stalls halfway.
A criterion that tends to work is looking for the process where the cost of doing it by hand is most visible. Some signs:
There is double data entry. The same data is written on paper, then in a spreadsheet, then in the system.
Nobody knows the status without asking. To know whether an order shipped or a complaint was solved, someone has to make a call.
It depends on a single person. If that person goes on vacation, the process stops.
It directly affects customers or cash. Quotes that take too long, lost orders, late collections.
Once the process is chosen, it pays to change it completely, even if it is small: from the moment the data comes in until someone uses it to make a decision. One improvement that goes all the way teaches much more than ten half-done improvements across different areas.
In Ambolt's projects, the first step is to interview several key people about how work really gets done, which rarely matches the org chart or the manual. That is how the first process is chosen, and a first piece goes into use within a few weeks, built around the systems that already exist.
AI changed two things for a traditional company. The first is cost: building a custom tool, connecting systems or automating a repetitive task now takes a fraction of the time it took a few years ago. Many projects that did not add up before now do.
The second is the kind of work that can be digitized. Reading invoices and delivery notes, sorting emails, answering frequent questions or summarizing scattered information used to depend on a person. Now much of it can be automated, with someone reviewing the doubtful cases.
That speed also brings a risk: if building is easy, it is easy to build too much. What is scarce now is the judgment to choose what to change, in what order and how to make it secure and cheap to run.
A new tool only transforms something if people use it. That is why what happens after building matters so much:
In a traditional company, digital transformation moves forward one process at a time. Starting with something concrete, taking it all the way and measuring whether day-to-day work changed gives something no general plan gives: first-hand evidence of what works in that organization. With each process solved, the next one comes faster, because the team has seen that change is possible and starts suggesting where to go next.
It is the process of changing how work gets done so that tasks, data and decisions stop depending on paper, loose spreadsheets and repetitive manual work, and rely instead on connected systems.
By choosing a process that has a visible cost today, such as double data entry, delays with customers or dependence on a single person, and changing it end to end in a few weeks before moving on to the next one.
The most frequent causes are projects without a clear goal, plans so broad that they take too long to deliver results, and a lack of support for the team to adopt the new way of working.
Digitizing order taking, connecting the management system to a customer or supplier portal, automating invoice entry, replacing shared spreadsheets with an internal tool or building dashboards that bring together data from several systems.
It makes building tools cheaper and faster and allows automating tasks that used to need a person, such as reading documents or sorting inquiries. Deciding what to change first and how to sustain adoption is still up to the team.
As a whole process, years. But a well-chosen first process can be running in a few weeks, and that first result is what gives momentum to the rest.
About Ambolt. Ambolt is a product agency that helps companies in Latin America decide what to build and put it into use fast, with no minimum-commitment contracts.
We specialize in finding areas to optimize your company's operations, making them leaner, faster and more productive.
Talk to us →