Installation and start

DAOS is installable standard software, not a rebuild project. This page shows what happens between your decision and live operation – and what you need for it.

The path to operation

First call

Potential analysis

Installation

Policies

Handover

Illustrative path – the detailed walkthrough is below.

1
First call
30 minutes
No obligation and no preparation needed. We work out where DAOS creates value in your company right away. Afterwards you know which next steps are needed.
2
Potential analysis
(Optional) 5 days
Two days on site for current processes and a technical inventory, three days of analysis, then a closing presentation of 90 to 120 minutes to the executive team. You receive a strategy paper with a prioritised roadmap, a business case, and the technical prerequisites for operation.
3
Installation
We install on a pilot machine, connect your model providers, and verify together that monitoring, resource discovery, and the dashboard work cleanly. Only then does it go wide.
4
Setting policies and budgets
Before the rollout, it is decided which models are approved, which security tier applies to which data, and where the budget limits sit. We set the system level so that nobody can water it down later.
5
Enablement and handover
In hands-on workshops your employees learn to build and steer agent workflows themselves. Once that works, our involvement ends – DAOS stays, operated by your people.

What the installation technically does

The installation itself is unspectacular, and that is deliberate. In one pass, the following happens:

The operating system and prerequisites are checked.
DAOS checks whether the Podman container runtime and – if local models are wanted – Ollama are present, and assists with the initial setup if they are not.
Program files, configuration, and the directory structure are created.
DAOS is registered as a system service and will start automatically from then on.
You set an administrator password (at least twelve characters).
Credentials and keys are stored encrypted in the operating system's keychain – not in files.

After the first start, DAOS discovers for itself what is available: processor, memory, graphics card, container runtime, locally installed models, connected interfaces, and databases. In the dashboard you immediately see what agents could access – before the first agent even runs.

Complex AI infrastructure solved with a single command: download the package, run daos install, done. What follows is configuration – not development.

What you prepare

Machine or server

One device matching the figures in the technical specification. For pilot operation, a well-equipped workstation is enough.

Container runtime

Podman from version 4.0. If local models are to run, Ollama as well – and for that a graphics card pays off.

Network and access

Outbound internet access for the installation and for cloud models. Nothing inbound is needed: dashboard and interfaces are reachable only on the local machine by default.

Model provider access

Your own API keys for the cloud providers you want – or the decision to start purely local for now.

Decisions on rules and budgets

Which models are allowed, who may start agents, where the cost limits sit. These are management decisions, not technical ones – and they belong before the rollout.

Points of contact

One person each from IT and from compliance. The contracts – terms and data processing agreement – are handled digitally through this portal.

How fast this goes

What we commit to depends on who is asking:

Department leads: a fully operational, policy-compliant infrastructure within 48 hours.
IT leadership: a production-ready installation on your infrastructure in under seven days, meeting your security requirements.
Compliance: measurably aligned with the requirements of the EU AI Act within one week.
Executives: complete transparency over all AI costs and access from day one.

These commitments refer to the local installation, that is, to what is production-ready today. Timelines for company-wide operation via the Enterprise Hub are agreed separately.

What you are left with

A team that configures and operates AI workflows independently.
Running, productive agents.
DAOS under licence, under your control, with your data in-house.
No permanent dependence on us as a service provider, as soon as you are ready for that yourself.

Book a slot now

The next step is a meeting, not a quote. Get in touch or book a slot directly – we look at your environment and tell you what is possible in 48 hours.