olduvia
Olduvia / How we work

A clear first step.

From diagnosis to daily use: scope, testing, control and continuity to make informed progress.

A clear first engagement

Before automating,
know where to start.

A focused diagnostic turns a recurring task into a decision: what could improve, what it would require and whether a pilot makes sense.

What we review

One process. Your actual work.

We review representative examples, volume, processing time, corrections and available tools. We agree on the necessary access and data before starting.

  • A conversation with the person who owns the process.
  • Examples of how work arrives and how it gets completed.
  • Integration options and important constraints.
What you receive

A decision you can act on.

  1. A process map and its main bottlenecks.
  2. A measurable baseline, with data gaps identified.
  3. Prioritised opportunities and transparent assumptions.
  4. A proposed pilot with deliverables, cost, schedule and acceptance criteria.

If your existing software already solves the problem, or a pilot is not justified, the conclusion will say so.

From problem to process

Start small.
Progress with evidence.

A pilot must answer one question: does this process work better this way? The next step depends on the answer.

01

Understand

We review a real case, volume, bottlenecks and available software.

02

Test

We define a pilot, the required data and acceptance criteria.

03

Implement

We connect your tools and practise with real team tasks: using the tool, reviewing its suggestions and resolving exceptions. We start with supervised work.

04

Improve

We measure usage, handle exceptions and expand what works.

Founder’s experience · development for Avanss

From a spoken instruction
to a controlled operation.

An example of how Lucas applies AI to technical operations: simplifying execution while keeping controls proportionate to risk.

01

Understand the instruction

A simple spoken request prepares complex network maintenance tasks.

02

Approve according to risk

The operation includes a prior backup. Moderate changes require mobile approval; severe changes require approval from a PC and a subsequent audit.

03

Check the result

Checks run after changes. If they fail, the system restores the previous state and sends a notification.

Developed at Avanss as part of the founder’s prior experience. This is not presented as a commissioned Olduvia project or a savings guarantee.

A starting point, with room to grow

One useful process.
Then the next level.

Start with a defined problem. Once the results support it, connect more work without losing control of scope, cost or responsibilities.

01 / A PROCESS

Prove the value.

One channel, a main integration and an owner. Compare the outcome with the starting point.

02 / DEPARTMENTS

Connect the handovers.

Sales, delivery and administration sharing records, with clear permissions and approvals.

03 / LOCATIONS

Coordinate the operation.

Explore shared reporting, work queues and service indicators across locations or contracts.

Each phase requires a separate feasibility assessment, scope and quote. Expansion is not included by default and the first project does not commit you to later phases.

A clear agreement

Visibility throughout.
Continuity afterwards.

From the outset, the proposal explains what is delivered, who controls access and how your work can continue if the service ends.

Access and documentation

Agreed access, operating instructions, dependencies and a record of the delivered configuration. Ownership and licences are specified in the contract.

A planned handover

Available exports, formats, any migration costs and support for returning to a manual process or moving provider are defined within the capabilities of the tools involved.

A pilot with an exit

Acceptance criteria and a fallback procedure are agreed before live use. If the pilot falls short, the next step is a documented decision, not an automatic expansion.

Before we begin

Good questions.
Clear answers.

Where does it make sense to start?

With a frequent, measurable process and a clear owner. In our first conversation we review what repeats, where work gets stuck and whether your existing tools can already solve it.

Do we need to replace our management software?

We aim to work with what you already use. Before committing to an integration we review access, formats, APIs and provider terms. If a connection is not viable, we explain this before quoting.

Will AI decide or send things on its own?

The scope is agreed for each process. Quotes, technical decisions and important messages can remain subject to human review. The aim is to prepare work better and automate only what has been authorised.

What happens to my company's data?

Before the pilot we agree on necessary data, permissions, providers and retention. The design must separate each client's information and allow system actions to be reviewed.

How much does it cost and how long does it take?

It depends on the process, volume and integration. The proposal separates pilot, implementation and maintenance, with defined usage and support. We do not give a fixed figure before reviewing the actual work.

Can we work together from another country?

Yes. Delivery is primarily remote for Spain and international markets. We agree on working language, time zones, data requirements and integrations. Commercial or implementation visits are quoted when needed.

Who looks after my project?

Lucas González Díaz leads Olduvia and is your point of contact. A compact team of specialists works under his direction, using AI to expand its capacity. Scope, responsibilities and support are agreed before starting.

Can you help us prepare for AI Act compliance?

Yes. We can review use cases and tools, train your team and prepare transparency measures and procedures. We define the technical and organisational support and matters requiring specialist legal review. Explore the scope ↗

Your next step

Let’s talk about your business.

Start a conversation ↗