Decide on data and ask your documents
We turn your sales, your website, your CRM and your spreadsheets into clear dashboards, reports that write themselves and AI predictions. And we let your team ask natural-language questions about your own documents, with the source always cited.
Projects from €799
Your company already knows more than it uses
The information is there, but scattered across places that never talk to each other: the till, the sales rep’s spreadsheet, web analytics, the CRM, the accounts. And next to it, another pile of knowledge locked inside PDFs, manuals, contracts and shared folders that nobody can find when it matters. The problem is not a lack of data, it is that nobody has time to join it up or to go looking. We connect the sources into a single panel anyone on your team can read, automate the reports you now build by hand, add models that anticipate demand and spot customers about to leave, and stand up an assistant you question like a colleague, which answers citing the exact document and page. If something is not in your sources, it says so.
What you get
The numbers that move your business and the knowledge you already hold, both finally within reach.
One screen, not ten tabs
All your sources connected in a custom panel with your real metrics, on desktop and mobile, readable by anyone on the team with no training needed.
Anticipate demand
Models tuned to your business that predict sales, stock turnover and footfall, so you plan purchasing and shifts on evidence rather than instinct.
Segment your customers
Segmentation ready to use in marketing and sales, plus early detection of the customers about to walk, while there is still something to be done.
Ask in natural language
Manuals, contracts, regulation, procedures and tickets turned into an assistant your team questions like a colleague, answering in seconds.
Always with the source cited
Every answer links to the exact document, page and passage the fact comes from. If it is not in your sources, it says so instead of making it up.
Alerts before it hurts
Scheduled reports that arrive on their own and warnings when a metric goes out of range, so you find out in time rather than at year end.
What you get
How we build it
From audit to production in under a month.
Data and knowledge audit
We review which sources you have, what state they are in and what documentation your team handles. We define the metrics that genuinely matter and the questions that must be answerable.
Connection, cleanup and ingestion
We connect the sources, resolve duplicates and inconsistencies, and load and index the documents together with their permissions.
Dashboards and reports
We build the panel with your metrics and schedule the automatic reports that replace the weekly spreadsheet.
Models and tuning
We train the prediction and segmentation models, and tune the search assistant against real questions from your team.
Integration and handover
We take it to where you already work (intranet, Slack, Teams, WhatsApp or the CRM), train the team and leave you the panel to manage documents without depending on us.
Two ways to use what you already have
Over your numbers
Dashboards, automatic reports and prediction. We connect the till, CRM, web analytics, accounting and spreadsheets, clean up whatever needs it and build the panel and the models.
Over your documents
AI search across your own knowledge (technically, RAG). We load your PDFs, Word, Excel, regulation and tickets, with role-based access control so each person only ever queries what is theirs.
Where it fits
Frequently asked questions
Do I need my data tidy before we start?
No. That is precisely part of the job. We always find data scattered, duplicated and spreadsheets built on different criteria, and the connection and cleanup phase exists for that. If something is unusable we will tell you, but you do not need to do homework before calling us.
Is this not just for large companies with a data team?
No. We build it exactly for companies with no data team: the panel is meant to be read by anyone and the reports arrive on their own. No endless projects and no expensive licences.
Are my documents private? Do they train any model?
They are private and they train nothing. We work in European territory under GDPR, your documents are indexed to answer you and are never used to train third-party public models. On top of that, role-based access control ensures each person only queries what is theirs.
How do you stop the AI making answers up?
The assistant answers only from your documents and always cites the origin: document, page and passage. When the answer is not in your sources, it says so instead of filling the gap. Being able to check every fact against its source is what makes the system usable for serious decisions.
Can it really predict my sales reliably?
Within an honest margin and with enough history, yes. A prediction is not a crystal ball: it is a range with a confidence level, which improves as more data comes in. We tell you upfront what reliability to expect in your case, and if your history does not support it, we say that too.
Which document formats does it take?
PDF, Word, Excel, PowerPoint, plain text, emails, intranet pages and ticket exports, among others. If you have scanned documentation, we run it through text recognition before indexing it.
How long does it take to go live?
From audit to production, under a month in most cases. The first panel or the first assistant is usually running within weeks, and from there it grows with more sources and more questions.