AI at work in the daily running of your company.
We build agents and automations on the systems your company already uses, for admin work, customer enquiries, sales and finding information inside the company. We train the team that will use them and look after them once they are live.
A quote request, from email to sent
Step 1 of 5The request arrives09:31
Ana Ribeiro asks for a quote for 400 cardboard boxes, delivered to Braga by the 20th.
The email lands in the company’s orders inbox, as it always has.
OutlookEmail received
Good morning. Could you quote for 400 corrugated cardboard boxes, delivered to Braga by the 20th?
Tools we build on
- Microsoft 365
- Google Workspace
- Claude
- ChatGPT
- Hugging Face
- Llama
- Mistral
- Qwen
- Ollama
- vLLM
- n8n
- Make
Four kinds of work we start with
Whatever the sector, we start from a task your team repeats every week and only then choose the tools.
- Admin and documents
- AI drafts proposals and reports from your own templates and enters invoices and forms into the systems you already use. A person checks everything before it goes out.
- Customer replies
- Enquiries that arrive by email, WhatsApp or the website are sorted and passed to whoever should handle them, with a first reply already drafted in Portuguese, English or Spanish.
- Sales and marketing
- Follow-ups go out on time and the client list stays in order. For marketing, AI prepares first drafts of posts and newsletters in the company's own voice.
- Company knowledge
- Anyone on the team can ask about a procedure, a price or a reply already given to a client, and the answer names the document it came from.
From the first call to tools in daily use
The price depends on the size of the team and the work, and we cover it in the 30-minute call.
Services in detailThe technology behind it
We choose the models and tools for what each task needs, and explain each part before we install it in your company.
- Agents
- They take on tasks with several steps: reading the request, checking systems, drafting, then handing the result to a person.
- Connections to your systems
- We connect email, documents and the ERP through MCP, the open standard the main vendors have adopted.
- Reasoning models
- We use them where a task needs analysis or calculation, such as building a proposal or finding anomalies in data.
- Open models on your servers
- We run Llama, Mistral, Qwen or Gemma through Ollama or vLLM on a company server, and the data never leaves it.
- Models tuned to your company
- When one task runs at high volume, we adjust a small model to the company's documents and tone.
- Document reading
- Invoices, forms and scans become structured data in the right system.
- Search across company knowledge
- The team asks a question and gets an answer drawn from procedures, price lists and past emails, with the source always shown.
- Voice agents
- They answer calls in European Portuguese to take bookings, give the status of an order or pass the caller to the right person.
- Screen and browser use
- For older systems with no direct connection, the agent works on screen as a person would.
- Testing and monitoring
- Automated evaluations and logs that show, month by month, whether everything still works.
- Image, video and audio
- Marketing material generated with AI and reviewed by your team before it is used.
Rules we keep on every project
- Built on what you already pay for
- We work on Microsoft 365 or Google Workspace, with mainstream AI and automation tools. We only propose a new platform when a task justifies it, and we explain why before going ahead.
- Your data stays in your accounts
- The tools run in your company's own accounts, and we use your data only for your project. Before we start building, we sign a GDPR data-processing agreement with you.
- Independent
- No software vendor pays us commission, so we choose tools for what the task needs.
- Work we turn down
- We do not build tools that screen job applicants or assess the creditworthiness of individuals, two uses the AI Act classes as high-risk.
- Legal support
- When a project needs legal support (GDPR, the AI Act, contracts), we work with lawyers who provide it.
Work delivered
Ontia Labs, formerly Arqmetrica.
At a family-owned Portuguese manufacturer of about 150 people, the board stopped two vendor pilots that had no clear success criteria and, within six months, put one use case into production.
Read the caseCommon questions about our AI consultancy
Do we need to buy new software?
Usually not. We build on the Microsoft 365 or Google Workspace you already pay for, and only suggest a new platform when there is a clear reason.
Where does our data stay?
In your company’s own accounts, or on your own server when we use open models. We sign a GDPR data-processing agreement before any build.
How long until something is working?
The Start programme takes 30 days. In that time the team takes the workshop and gets two working tools, with our support while it gets used to them.
What happens after launch?
We keep the tools running with monthly support, in three levels, depending on how many tools you have and how much help the team needs.
Start with a 30-minute call
Bring one task that takes your team too long, ideally one that comes round every week.
Book a 30-minute callBy the end of the call you will know where we would start, and whether this is work we do.