Skip to content
Sanfeliu Consultancy

Cases

What we’ve done, in two lines.

Real work for real companies. No names, out of discretion toward our clients; no adjectives, out of habit.

  1. 01

    Warehouse management software company

    03 AI in your ERP
    What was there
    Every new client arrived with their own ERP and their data in their own format, and connecting each one was engineering work, one at a time.
    What we did
    An integration engine that brings data from any ERP into the warehouse system along the same path, and new-client setup prepared by AI.
    Today Orders by emailand PDF A person readsand types them in ERP A morning, every day. Copying errors. No one asks the ERP anything. With AI connected Orders by emailand PDF AI reads, checksand validates ERP(the same one) Someone reviewsand approves Five minutes, not a morning. And management asks the ERP in plain English.
  2. 02

    The same logistics software company

    01 Business backends and BI
    What was there
    Their clients tracked shipments in an old application and asked for metrics by email.
    What we did
    The tracking application rebuilt from scratch, multi-company, with built-in reports, and a dashboard where each client defines their metrics by writing what they want to see.
    Today ERP, store,warehouse, bank Someone copiesinto spreadsheets Mondayreport Four spreadsheets that don’t add up. The metrics show up whenever they show up. With the central backend ERP, store,warehouse, bank Centralbackend Dashboard “Margin by family,this quarter” On Monday, management opens their phone. A new metric takes one sentence to request.
  3. 03

    Industrial supplies distributor

    03 AI in the CRM and customer service
    What was there
    A sales team that had never used AI, with knowledge spread across the CRM, internal documents and a catalog of hundreds of items.
    What we did
    AI connected to the CRM, the documents and the live catalog; training for the whole team and a prompt handbook for salespeople. Meeting briefs and proposals in minutes.
    Today Support andsales emails Someone reads,sorts and replies CRM and inbox Urgent emails get lost among the rest. Every call means rereading the history. With AI connected Support andsales emails AI sorts, sums upand drafts CRM and inbox(the same ones) The team reviewsand sends Each ticket to the right person, prioritized, with a draft. Alerts when a deal goes cold.
  4. 04

    Digital health company

    03 AI on your custom software
    What was there
    Their data in their backend, and clinical staff who had to open the interface and click through it to look anything up.
    What we did
    A connector on top of their backend with credential-based access: staff ask and get the answer without opening anything.
    Today Yoursoftware Open three screensand click through The answer Every question means a trip through the interface. Nothing is automated. With AI connected Yoursoftware MCP connectorwith permissions Question inplain language An answer, orthe task done The AI inherits the user’s permissions: it can’t see or touch anything the user can’t.
  5. 05

    Vacation rental agency

    01 New software with AI
    What was there
    Nearly two thousand verified reviews scattered across about a hundred listings on a portal, invisible to Google and to AI assistants.
    What we did
    A review site of their own in three languages that collects, verifies and publishes them so they can be found and cited.
    Review site with 1,940 verified reviews and the average rating
    A rental agency’s own review site: 1,940 verified reviews, recovered from 104 listings on a portal and published in three languages so they can be found and cited.
  6. 06

    Real estate agency with three websites

    01 New software with AI
    What was there
    Three websites on the same listings, cannibalizing each other, and a search that couldn’t find what was there.
    What we did
    An audit scoring each website and a new version of the main one with the same filters, to decide with data before rebuilding.
    Today Threespreadsheets Emailsand calls A process no oneoversees What the company does isn’t written down anywhere: it lives in two people’s heads. With the new app Incomingdocuments AI sorts them andextracts the data Your ownapp A personapproves Your code, your accounts. Working software every week.
  7. 07

    Group of three hotels in Cerdanya and Porto

    02 Websites
    What was there
    Three different websites, with three booking engines and no language in common.
    What we did
    A single website in four languages for the three properties, each connected to its own booking engine.
    Today Old website PDF menu, pricesfrom two years ago Google can’tfind it Anyone who wants to book or order can only call. With the new website Fast onmobile Reservations andcatalog connected Fourlanguages Found by Googleand AI assistants Changes included: you ask, we make them. Domain and content are yours.
  8. 08

    Well-known restaurant

    02 Websites
    What was there
    A website that never got updated and a PDF menu nobody could find.
    What we did
    A new website in four languages, with the menu and allergens generated from their printed menus, and the same red and logo as always.
    Today Old website PDF menu, pricesfrom two years ago Google can’tfind it Anyone who wants to book or order can only call. With the new website Fast onmobile Reservations andcatalog connected Fourlanguages Found by Googleand AI assistants Changes included: you ask, we make them. Domain and content are yours.
  9. 09

    taula.ai

    Our own product 01 New software with AI
    What was there
    Restaurants that handle reservations by phone and WhatsApp, without knowing how many people they have each seating.
    What we did
    A reservation system for restaurants, in production, which we use as a lab for what we later build for clients.
    taula.ai reservation screen on a phone: party size and date
    taula.ai, our own product: guests pick party size, day, and time on their phone, and the reservation reaches the restaurant on its own.
  10. 10

    Guestproof

    Our own product 01 New software with AI
    What was there
    Vacation rental agencies with their verified reviews locked inside Airbnb, Booking and Vrbo, far from their own website.
    What we did
    A SaaS in production that collects the reviews for each property and, on its own, publishes a review website under the agency’s brand, ready for search engines and AI assistants.
    Today Threespreadsheets Emailsand calls A process no oneoversees What the company does isn’t written down anywhere: it lives in two people’s heads. With the new app Incomingdocuments AI sorts them andextracts the data Your ownapp A personapproves Your code, your accounts. Working software every week.

The names, figures and details of each project are shared on the fit call, where appropriate, and always with the client’s permission.

How we work

You see it working before you decide anything.

Let’s start by seeing if we’re a fit.

Eight questions and a call. Free, and we tell you what we think even if the answer is no.

It’s a call with the engineer who would lead the project, not with a salesperson.

Take the questionnaire Or write to us