Volundra

Volundra builds AI that takes over entire services, completely. A system that does the work itself, inside the limits the client sets, and stands behind every decision it makes.

Our AI plans, decides, executes and documents. It asks the client only what the client has to decide, and answers everything else itself. Today it runs advertising budgets; the technology underneath is built for more.

Volundra GmbH Operated in the EU Audit trail on every decision
The Volundra AI

Our own decision system.

Our own rulebooks, our own safety architecture, our own memory, orchestrated on leading models and operated in the EU. We do not train a foundation model of our own, and we do not claim that we do.

Rulebook

Verified knowledge instead of gut feeling

Under every decision sits a verified reference work of the platform rules, every statement with its source and the date it was retrieved, anything unsourced marked as unsourced. It grows with the rule changes of the platforms, not with the fashions of the market.

Safety architecture

Limits the client sets

Budget, risk and product range are set by the client. The system checks itself against those limits before it acts, stops instead of crossing them, and can be halted from the outside at any time.

Memory

Every run leaves a trace

The numbers, the decision and the reasoning in plain language are written down, the doubts included. From that record the system learns what carries an account and what does not, so the next run does not start from zero.

Orchestration

Models are replaceable, we are not

For language and judgment we call leading models, chosen separately for each task. They can be swapped out without the system losing what it knows. Our rules, our limits and our memory stay.

Case study: Adtronaut

Adtronaut is the first service we took over completely: paid advertising, run end to end. Every number below is measured there, in live accounts. It is what coverage looks like when it does not shrink as accounts grow.

100%

of an account's search terms are reviewed by Adtronaut, every day. Whether an account produces five hundred terms a day or eleven thousand, each one is rated and either keeps receiving spend or gets blocked. Adtronaut does not sample and it does not skim.

Measured in production, not estimated
25
automated checks run before Adtronaut moves any budget, on four-figure accounts and six-figure accounts alike. If one fails, nothing happens.
9,500
lines of verified platform rulebook sit underneath those checks, each statement with its source and retrieval date. It grows with every platform change, so the thousandth account starts smarter than the first.
7
days a week, around the clock. Adtronaut works on the platforms' schedule, not on office hours, and it does not take August off.
Principles

Four statements you can hold us to.

They are here not because they sound good, but because the way we build enforces them. If we break one of them, that is a defect, not a compromise.

  1. I

    Measured, not promised

    We say beforehand what will be measured, and afterwards what came out of it, with the raw data next to it. We make no commitments about the result. A report you cannot recalculate is, to us, not a report.

  2. II

    The rule first, then the model

    Whatever a fixed rule can decide is decided by a fixed rule. The language model comes in where language and judgment are needed. That keeps decisions verifiable and operations affordable, at the thousandth account as much as at the first.

  3. III

    Data in the EU, described honestly

    Processing and storage run in European regions, on services from providers based in the United States, covered by standard contractual clauses and a transfer impact assessment. For the AI models, content leaves the EU. We write that down instead of writing around it.

  4. IV

    Every application separable, with clear responsibility

    Its own technology, its own data, its own contracts, from day one. Who answers for what is in the contract, not in the fine print. Every application should stand on its own: possible to value, possible to audit and one day transferable on its own.

Applications

Where the Volundra AI works today.

An application is the visible side of the system: one service, one brand, one client. It counts only once it runs for a real client.

Application 001 live in production

Adtronaut

A complete advertising operation in one system: it creates the ads themselves, copy, images and video, runs the campaigns across channels, replaces creative before it fatigues, steers the budgets and puts every result in the open. The work of an agency, carried out by the Volundra AI.

See Adtronaut

Its own brand, its own technology, its own data, its own contracts. Transferable on its own.

In development unnamed until it earns a name

Further applications

The next one is already being built, on the same AI and with the same memory. It appears here when it runs for a real client. Not a day earlier.

Company

Volundra GmbH

One company behind the AI and behind every application built on it. The technology, the data and the contracts sit with it.

Company

Volundra GmbH

Application

adtronaut.com