Businesses don't grow by using more technology. They grow when they understand how they work.
That's why Itria doesn't sell just software, automations or tools. First we analyse the business, then we organise the processes, then we measure the results: only then do we bring in the right technology. Technology is the means. Commercial performance is the goal.
Analyse
We study the business's customer journey, its day-to-day operations and its sales processes, based not on assumptions but on how it really runs, day by day. We track where time goes in and where it leaks out, which tasks are repeatable (and can be automated without losing quality) and which need human judgement that can't be replaced. This phase has one aim: to know exactly where the business is losing hours before proposing any solution.
real processes
Organise
We redesign those processes as simple, clear, repeatable workflows: every task follows a defined sequence and no longer depends on improvisation or one person's memory. This is where we build operational continuity: the aim is for the business to keep running when the decision-maker is out of the office, without anything slipping through.
processes
Measure
We turn scattered data and information into decision metrics (KPIs): time to first response, share of leads followed up, no-shows prevented. They're the figures that really matter to that particular business, not a generic dashboard. What isn't measured can't be improved: the aim is for every commercial decision to have a number behind it, not a hunch.
metrics
Improve
Only now do we bring in the right tools: CRM, automations (workflows that handle the repeatable steps on their own), artificial intelligence (where judgement on text or language is needed) and a control dashboard. Not because they're innovative, but because they fix exactly the bottleneck found in phase 1. The aim: a working system within thirty days that the client owns and understands, not a permanent dependence on a supplier.
system
Each phase has to deliver a result before we talk about the next one.
We never bring in a tool before we've understood the process it's meant to serve. A tool on top of a process nobody understands fixes nothing: it just adds one more tool to manage.
The gap between Italian SMEs that have adopted AI and those that have actually built it into their processes has been measured, and it's wide: we wrote about it in 81% of Italian SMEs use AI, one in four has built it in.
A number only counts if it could have gone the other way.
At the end of any month, anyone can find a figure that has improved. The point is to decide in advance which figure to watch and what counts as failure, and then to accept the verdict. There are four steps, and they apply to every piece we build.
We pick one metric, and only one
Not “improve the processes”, but a specific figure: the median time to first reply to a foreign enquiry, the average days to get paid, the share of enquiries left unanswered. One, chosen together and written down, because with three metrics you can always cling to the one that went well.
We measure where things stand today
Using your data from the previous months, before we touch anything. It's the dullest step and the one nobody bothers with: without a baseline, any later improvement is just an impression, well told.
We write down what counts as failure
A value and a date, agreed before we start: if the metric hasn't passed that threshold by that date, the piece has failed. We write it ourselves, and it counts against us.
We make sure it isn't chance
A figure can improve because of the season, a campaign that's running or a trend that had already started. Before taking any credit, we ask whether the improvement beats what would have happened anyway, over periods long enough to mean something: a few weeks don't prove anything, even when the numbers are big.
When the result is uncertain, we say so, instead of rounding it down out of caution or up out of enthusiasm. An improvement that doesn't survive this check isn't an improvement: it's a coincidence with an invoice attached.
The load tests already run on the prototypes, with the figures and how they were calculated, are on the sector pages: food and export, hospitality, restaurants.
We don't install everything on day one. We build one piece, prove it works, then decide on the next.
The first piece of work starts as a pilot: we begin with the most obvious problem, the one already costing the most in time or lost revenue, and build only that piece. Thirty days, a measured result, a number before and after. If the piece has earned back more than it cost, the next one makes sense. If not, we stop there.
It isn't a package sold all at once on day one: it's a system that grows only where it has proved its worth, one automation at a time, even when they're quite different from one another. The business decides whether and when to add the next piece, always with its own numbers in hand.
We don't sell just software, CRM or automations.
We build them when they're needed, connect them to each other, and above all make them pay off. Software shows you the numbers. Itria turns them into decisions.
| What | Typical web agency | Traditional consultancy | Itria |
|---|---|---|---|
| Starting point | A package or standard template, the same for every client. | An analysis, often on slides, with nothing built that actually works. | A real measurement: where the business is losing time or revenue today. |
| Payment | Everything up front, before you see any result. | By the day or by the hour, whatever the outcome. | One piece at a time: each piece pays for itself before the next. |
| Who owns the system | Often stays with the supplier, with paid maintenance. | Not applicable: no system gets built. | The client owns what's been built, with no permanent dependence. |
| After delivery | Ticket-based support, minimal oversight. | A document, not a working system. | Looked after directly by the person who built it, not a ticket number. |
What we don't do.
- ×We don't sell a subscription without measuring first. If we don't know where the business is losing hours or revenue today, there's nothing to quote for.
- ×We don't replace tools that already work. If your business software does the job, it stays. We connect things; we don't replace them just to bill for a migration.
- ×We don't let AI talk to your customers unchecked. Every message that goes out in your name passes through a rule you've approved, or through a person.
- ×We don't deliver slide decks. Every phase produces something that works or a measured number. A document isn't a result.
- ×We don't hold your system hostage. At the end the client owns what's been built, credentials included, and is free to walk away.
- ×We don't take on work we don't believe in. If the problem can be solved with half an hour of reorganising, we'll tell you, and we won't send a quote.
The most common questions.
How much does it cost?
It depends on the piece, and we can't know until we've looked. What we can tell you up front is the shape of the cost: you pay for one piece at a time, each with a written scope and a fixed price agreed before work starts. No subscription that kicks in before the system works, no day rates.
The first piece is deliberately small: it has to cost less than what the business is already losing on that process, otherwise there's no point proposing it. If the first call shows there's no return to be had, we'll tell you and won't send a quote.
Do we have to change our business software or throw out what we use?
No, and that's the whole point: we don't replace the tools you have, we make them pay off. Business software, email, WhatsApp, the Excel sheets the team knows by heart: they all stay where they are. The work is getting them to talk to each other and cutting out the manual steps in between.
We only suggest replacing a tool when the analysis phase shows it's the bottleneck, never because we'd prefer a different one.
Our team isn't technical. Will they manage?
The design rule is that the system mustn't need any new skills to work. In most cases the work happens behind the scenes, and things arrive already sorted in the places the team already looks: the inbox, the spreadsheet they already use, a message. If a piece means someone has to learn a new tool, that piece has been badly designed.
Where does our data go, and is the AI trained on it?
No. Our contracts with the AI model providers rule out using the data for training, and the orchestration runs on servers in the European Union under our direct control, not on third-party platforms that resell access.
How we protect access, data and backups is set out line by line, with no gloss, in AI transparency.
We've already tried AI and not much changed.
That's the norm, not the exception: in Italy 81% of SMEs already use AI tools, but only 25% have really built them into their processes. A tool in one person's hands helps with that one half hour; it doesn't touch the process that person repeats eighty times a month.
The difference isn't the tool: it's whether someone has mapped the process, decided where human judgement is needed and where it isn't, and connected the result to the tools the business already uses. That's exactly the work we do before writing a single line of the system.
How much of our time will it take?
The analysis phase is the most demanding, and even then it only comes to a few hours in total: we need to see how you really work, not a questionnaire filled in in a hurry. From the build onwards the work is on us, with a check-in whenever there's something concrete to look at.
If a project needs weekly meetings for months, it isn't a project: it's a consultancy subscription.
A method is best judged on a real case.
The phases above describe how we work. What they would do for your business only becomes clear from your numbers. Itria builds custom digital systems for Italian SMEs, starting from the outside in. For you, that means more enquiries, fewer things falling through the cracks and less work done twice. Drop us a line about what's slowing you down. We'll make the first move: we'll look at what a customer sees when they search for you, and tell you what we found. Even if we never end up working together.