Twenty-five questions, answered in full.
What does LookAit do?
LookAit is a business intelligence consultancy based in Córdoba, Argentina, working remotely with companies across the Americas, the UK and Europe. We rebuild the data architecture of businesses that already hold large volumes of data, and build end-to-end systems on Google Cloud — from extraction to dashboard — that turn that data into business decisions. On top of that we apply machine learning and AI to project where the data is heading, so the decision gets taken before the change arrives rather than after it.
Who do you work with?
Companies, startups and agencies whose data is already digitised and already substantial — several databases, hundreds of thousands of records. Our proven ground is high-volume fintech. The filter is not the size of the company but whether the data exists in a usable form: a small business with clean, structured data is a fit; a large one still working on paper is not.
Can I offer data services to my clients without having a team of my own?
Yes, and it is how a good part of our work runs. You keep your own contracts and your own client relationships; ours with you is separate, and neither touches the other. The capability sits on our side, the relationship stays on yours. A cybersecurity company has worked this way with us for three years, across its own operation and the service it delivers to its clients.
Do you appear in front of my client, or work behind the scenes?
Whichever you need. Some partners introduce us as part of the team; others work with us entirely in the background, which is usually the case in sensitive markets. It is your call, not ours, and it can change from one project to the next.
What happens to my clients’ confidentiality?
It is respected exactly as it stands. We work in an industry where that is not negotiable, and we know how to handle data under those conditions. The work we build belongs to whoever commissioned it, and we can deploy it inside your own cloud accounts, so the data, the systems and the control never leave your side.
Do you work with several consultancies at once?
Yes, including within the same sector. Each relationship is its own contract and its own workspace, and nothing crosses between them.
Should I hire a data scientist or bring in an outside team?
Hiring is a reasonable thing to want. What is worth weighing is what the first stretch actually needs: an analyst, a data engineer and a data scientist are usually three different people, and a single hire covers one of them. With us that team is in place from the first month, on a fixed budget, without recruitment, without profile searches, without the risk of the wrong fit. And if the area grows to the point where an internal team is justified, what we built is the foundation it stands on — current technology, documented, and handed over with our teams alongside yours until the responsibility is fully theirs.
I already have a technical team. Do you replace it?
No — we work alongside it. Calls, meetings, shared work: we are firm believers in co-working, because a project like this runs on feedback. The rules of your business live with your people, and there is no way to learn them from the outside.
How long does a project take?
A full cycle runs three to eight months depending on complexity. The initial diagnosis — the first deliverable — takes around two weeks.
What does it cost?
Every system is bespoke, so the budget is fixed after the diagnosis, which defines scope, stages and a closed price. A share is paid up front and the rest on delivery. After that we keep watching what we built — our relationship continues.
What do I take away if the diagnosis does not convince me?
The diagnosis itself. It is yours, whether you continue or not: the findings, the data-model map, the metrics we validated together. In practice it rarely ends there, because the diagnosis is a two-way stage — we cross what we find against what you already sense about your own business, and that exchange tends to open questions nobody had thought to ask.
What if you do not find anything in my data?
Then the product is smaller, and that is a legitimate outcome. It might be a forecast, or a report that monitors how one specific tool is behaving and which faults keep repeating. The work grows to the size the data justifies, no further.
What access do you need to my systems?
Only what the work actually requires, and always with credentials you create and authorise. You decide what to share and what to keep back. The trade-off is worth knowing: the more varied the raw material, the more we find, because most discoveries come from crossing sources that had never been put side by side.
How do I know whether my company has opportunities hidden in its data?
You usually do not, which is the point. A real example: crossing customer numbers against credit capacity on a map, one location showed up as a very small, very healthy dot surrounded by much larger struggling ones. Nobody was looking there. That reading came from an industrial-engineering habit — asking which part of the business process is worth opening before looking at anything — rather than from the data alone. The result is in our cases.
What technologies do you use?
We work on Google Cloud Platform — BigQuery, Vertex AI, Cloud Storage and Looker — with Python and SQL for processing, machine learning and AI. We also build AI-agent systems.
Do I need my data to be tidy before we start?
No — tidying it is part of the work. What you do need is data: digitised, and in enough volume to be worth mining. The cleaning, the architecture and the analysis are ours.
Do you build AI and agents?
Yes. Beyond analysis and dashboards, we apply machine learning and AI to your data and build AI-agent systems integrated into your processes.
Can you bring runaway cloud costs under control?
Yes. Poor data architecture is the most common cause of a runaway cloud bill. Rebuilding how data is stored, processed and queried can cut that cost sharply without losing accuracy.
What does it cost to build a data department from scratch?
Three salaries, recruitment, the months it takes to find the right people, and the risk that one of them is not. Against that sits a closed budget and a team already working. Which one makes sense depends on how much data work you have ahead of you — that is the question worth answering first.
Do you work with companies in other countries?
Yes. The work is remote and the only fixed points are the calls: one or two hours at a time, planned in advance, mostly during validation stages. Time zones are agreed between both teams, and we invoice from Argentina.
What happens when the contract ends?
The contract ends; the relationship does not. Every engagement is quoted on its own, so there is nothing tying you to the next one. We keep monitoring what we built, and when you want to extend it, that is new work with its own scope and its own price.
What happens after delivery?
We hand the project over fully documented and keep monitoring how it performs as your data grows. You pay again only when you decide to extend it.
Do you build custom dashboards and data portals?
We build the full stack. A company sitting on transaction records, sensor feeds or CRM data can end up with a portal that ingests it, models it and serves it back as a live decision surface — not a static report someone exports on Fridays. The demos on this site are software we built and run ourselves, on the same stack.
What is the difference between a BI consultancy and building data-driven software?
A BI consultancy connects to data that already exists and helps you read it. We do that too. The difference is when the data does not come in a usable form, or when the output needs to be a product rather than a report: a client-facing portal, a monitoring tool, a scoring engine. That is software, and we build it end to end, from the data model to the interface.
Do you build the software or only the strategy?
Both, in that order. Strategy without implementation is a deck. We diagnose, design and build — and we stay involved after delivery, watching what we built perform. The cases on this site are systems running in production, not recommendations left in a document.