Real processes that stopped depending on someone’s memory.

The breakdown that gets lost along the way never gets fixed

In a hotel, housekeeping and maintenance work in parallel, but they almost never talk to each other. The housekeeping team handles cleaning requests, linen changes and the occasional clash over who was supposed to do what every day. When something breaks, the report passes from hand to hand: a guest or a housekeeper tells reception, reception tells maintenance. If the chain breaks at any point, the issue can be lost for good. And a breakdown that gets lost doesn't just stay there: it ends up as a complaint at reception or a bad review.

We built a solution where every room has its own QR code. The guest scans it, says what they need in their own language —without calling or registering— and reception sees it live, translated with one click.

Each employee logs in with their own account and role: maintenance receives exactly what is theirs, the head housekeeper assigns tasks to the housekeepers, and reception and administration oversee the whole. Every request carries its priority and its history: what was done, what's missing and who it was assigned to.

The notice is no longer written by hand

A transfer agency schedules pickups from train stations and airports to the hotel, coordinating each city with its own transport provider. Before, every request was written by hand to send to the provider: origin, destination, arrival and departure times, contact details… With the volume of orders the agency handles, that led to constant errors.

Our tool pulls that data straight from the agency's internal system, reads the train ticket or the flight note, and calculates the pickup time according to the mode of transport (train, plane or bus). If two booking references share a car and schedule, it merges them into a single transfer, optimizing cost instead of sending two car requests. The system warns about schedule inconsistencies, but some decisions remain human: grouping travelers is as simple as dragging and dropping into boxes. The notice comes out already written: on one hand, it's sent to the provider to request the transfer (by email, downloaded as a PDF or shared via WhatsApp); on the other, it generates the note with the driver's time, date and contact details, ready for the agency to pass on to the client through its own system.

Everything pending, in one place

An inbound tour operator organizes sightseeing tours for groups, and each tour scatters its tasks across different places: train tickets to issue, notes inside each booking about allergies or special requests for the hotels, vouchers and rooming lists that must be sent to the guides. Before, you had to go and look for each thing separately.

The agency already had its own system, but everything was scattered and it wasn't practical to check it that way. Our tool connects to that system and gathers all those tasks in a single place. Tickets pending issue show up automatically, and a reminder calendar keeps what's left to resolve in view.

Each guide's paperwork is no longer downloaded file by file: it's requested once, the system downloads and organizes it into folders, and with one button the emails are generated to send each guide theirs.

There’s also a button that does something nobody used to have enough time to do, given the huge volume of bookings: it reads the remarks on all the bookings in a period and flags any note about allergies, intolerances or a request for the hotel that nobody marked so it would reach the corresponding rooming list — without that mark, it stays as an internal note and, if someone misses it, it doesn’t arrive in time.

In addition, the phone is connected to the same system: incoming calls to the office appear on screen, and from the browser extension, just by being inside a booking and pressing a button, you call that agency directly without dialing any number.

From a stray message to a campaign in one click

A company that connects brands with influencers manages the entire relationship from the platform we built: sign-up, contract and collaborations. Before, finding out who wanted to take part in each campaign meant writing to them one by one and waiting to see who replied, and then cross-checking that data by hand against what each brand was looking for.

Each influencer signs up with their details and signs the service contract —clauses and commissions included— inside the app. From there they become part of the roster the team manages.

When a campaign comes up, it’s published as an open request —anyone can sign up— or a closed one, aimed at a specific profile. On open ones, each interested person signs up with one click and the team sees at a glance who wants to take part. That list also helps put brands in front of the profiles that truly fit what they’re looking for.

Once it's decided who takes part, tracking begins: an encrypted internal chat, availability marked by each influencer, and the collaboration moving forward through visible phases, from acceptance until the brand gives the green light.

The platform also reminds you to update each influencer's details, or asks for specific information before a deadline if something needs reviewing. The documents for each collaboration are stored there too, and the brand can follow the project's progress through a link, without having to ask.

The schedule that fits the law, not the other way around

In Porto’s historic center there is a traffic restriction on weekdays, from 18:00 to 20:00, and the loading and unloading of each service has to fit within that regulation. Before, fitting the schedule was manual trial and error, and if it didn’t fit you had to improvise with the group already at the destination.

Now the tour data is entered just once and the fit is calculated automatically: if it doesn’t fit before the cutoff, it tries trimming combinations —the walking stretch down to 5 minutes, the panoramic drive down to 15— and keeps the one that needs the least adjustment, noting the trim in the document itself. If there’s truly no way to make it fit, it proposes alternatives instead of leaving it blank.

With one button the PDF report is generated and, right after, the route map, ready to send through the municipality’s portal. Before, even with the template already built, it took about 20 minutes, and over an hour if something changed, because every small change meant redoing everything. Now it takes about 4 minutes, and each document comes out different depending on the data and the order it’s given.

A bank isn't designed for downloading a hundred receipts in a row

Filing bank transactions on paper means pulling the receipt for every deposit and payment from the bank. Done by hand it's slow and repetitive: log in, filter, open, download, and again. The tool generates a script that runs in the browser itself and downloads the unread receipts matching the description you ask for, marking each one as read so it isn't downloaded again next time. If the work piles up over several months and a large volume of transactions arrives, it used to mean losing hours at a stretch; now each receipt downloads in seconds and a full session is done in about 30 minutes.

With the same tool they are then sorted by type, date, payee or sender, and even deposits by invoice number —since they are sequential and written in the description, the tool reads them and sorts directly by that number. Everything is combined into a single document ready to print.

It finds out about the change before the client calls

A train’s schedule can change after the ticket has been issued. If nobody finds out before the client does, the problem blows up at the worst moment —already at the station, with the client having to sort it out on the spot and the agency taking the heat. Checking each ticket by hand, one by one against the official timetable, isn’t feasible with the volume an agency handles.

The tool downloads the issued tickets, reads the train, date and time of each one —from several companies, including those that don’t publish official timetables— and compares them against the reference source. Anything that doesn’t match is grouped by booking, with a note already written, ready to paste and notify. This makes something previously unthinkable possible: reviewing the entire ticket portfolio periodically, every few days, without blocking the team for half a day. By hand, reviewing a batch like that —without interruptions— would take almost 6 hours straight. Repeating it every week wasn’t an option. Now it is. Even on transport strike days, it cross-checks the official minimum-services resolution: it marks which trains are guaranteed to run and which aren’t, and looks for real alternatives covering that same route that day, so the client’s trip is guaranteed.

From booking to issued ticket, with no retyping

Issuing a train ticket goes through several repeated screens: you have to move the client's data from the internal system to the train ticket provider's website, and once the ticket was issued, the trip details had to be recorded by hand inside the internal system in a template with the travel information. With the volume a wholesale agency handles, that repeated typing is where the most time is lost and where it's easiest for something to be mistyped.

Our tool takes the client's data and fills in the form by itself across the different booking screens. Once the ticket is issued, it reads it from the inside and fills in the internal trip template with that data, which used to be done by hand.

And if, while processing new bookings, a client who has already traveled with the agency is recognized, the tool searches the history for who else they traveled with on previous occasions, to speed up the booking and move on to the next call faster — you can act on it with a single click.