Skip to main content
Audact
See the offer

Who this is for

Audact is built for you if you deploy governed AI in regulated Europe.

You run AI for clients under European rules, you need a receipt for every interaction rather than the assurance that one exists, and you sell under your own brand. Below are the five situations where that description does not fit, and the ten things we deliberately do not build.

Five cases

Where Audact is not the right fit.

They are consequences of what Audact is built around, so a shorter sales cycle would not change the answer.

  • Case 01

    If you operate US-only with no EU exposure, and your AI only speaks

    A US-native voice platform will suit you better

    Audact is engineered around EU regulatory primitives: Article 50, the Dutch Telecommunications Act, DORA, the DSA, jurisdiction-aware policy. With no European callers, no European users and no European entity, all of that is overhead you are paying for. A platform built and priced for the US market fits better. One line decides it, and it is not the map: this holds where the call ends and nothing is bound. If your US agents can commit something the business then has to honour — a credit, a discount, a price, a promise — that is a different question, and it may be a Machine Authority Pilot rather than a no. A US-native platform can carry the call underneath while the commitment runs through an authority boundary; it is the infrastructure in that case, not the competitor. Ask us and we will tell you which of the two you are.

  • Case 02

    If you want a fully self-serve, build-it-yourself flow builder today

    A self-serve platform will suit you better

    Audact's intake is sales-led: each agency is onboarded personally, because the first configuration decides what the evidence looks like for every call afterwards. A visual flow builder is on the roadmap. If you need full self-serve today, a self-serve platform fits you better until our builder ships.

  • Case 03

    If you only need free, open-source voice AI

    An open-source voice-agent framework will suit you better

    If you have engineers, time and no compliance budget, the frameworks are freely available and they are good. Build it yourself. Audact is what you move to when a regulator asks what the gate released on a call in March and you need the answer by Wednesday.

  • Case 04

    If you want a generic, channel-agnostic chatbot with no compliance requirement

    A general-purpose chatbot platform will suit you better

    Chat is on our roadmap, with the same disclosure and the same evidence chain as voice. But we are built for regulated interactions in Europe. If you want a marketing chatbot with no compliance burden attached, a general-purpose platform will be cheaper, lighter and easier to live with.

  • Case 05

    If your primary compliance burden is non-EU: US state law, Canadian PIPEDA, APAC frameworks

    A vendor built for that regime will suit you better

    The policy engine is built around European regulatory primitives, so covering another regime is not a translation but a different product. If your buyer's primary need sits outside the EU, we will say so in the first call rather than in month four.

Opinionated by design

Ten things we deliberately do not build.

Every feature we skip is one less thing to defend in front of a regulator, and one less way for the evidence chain to have a hole in it. If you need any of these, pick a platform that builds them. That is a real requirement and we are not going to argue you out of it.

  • Bring your own key for the model

    We control the stack in order to control what the receipt can honestly say. We run only models we have evaluated, because otherwise we cannot stand behind how they behave under the policy.

  • On-premises deployment

    EU-based hosting, not EU-exclusive. That is the environment we host, run and support; in an environment we do not operate we cannot stand behind the same chain.

  • A voice-cloning marketplace

    We will not host a marketplace of synthetic identities: Article 50(2) puts the marking duty on the output, and a marketplace multiplies the number of voices we would have to stand behind.

  • Enterprise contact-centre suites

    Not our segment. We are agency-native, not a replacement for a large-enterprise contact centre, and pretending otherwise would cost you a year.

  • STIR/SHAKEN caller authentication

    A North American caller-ID regime that we do not implement, and that does not apply to EU deployments. Separately: machine-readable marking under Article 50(2) is planned, with C2PA as the intended format; it is not in the product today.

  • Custom models and fine-tuning on your weights

    Leading governed models only, run through our stack. We can record which model ran; what we cannot do is stand behind the behaviour of a model we have not evaluated, so we support evaluated models only.

  • A self-serve credit tier

    White-glove onboarding, or nothing. There is no credit-card hobbyist tier: running an evidence chain per call costs real money and hiding that in usage surprises would be the dishonest version.

  • Every language, counted for the count

    We cover the European languages where we can defend the disclosure wording as well as the transcription. Raw language count is a number that looks good and tells you nothing about whether the disclosure is correct in that language.

  • Squads of agents handing off to each other

    Single-prompt agents. We support one prompt, one policy and one traceable decision per turn today, because that is simpler to check a year later by someone who was not there.

  • A user-tunable denoising mode

    Premium EU-routable synthesis on a standard audio path. No user-tunable denoising: one audio path keeps delivery and support simple.

These are the ten we are asked about most. Ask us about anything else.

Still here? Then you are our buyer.

You deploy AI for regulated clients in Europe, and you need a tamper-evident compliance receipt for every interaction rather than an assurance that one exists. That is exactly what we built.

Apply for a founding seatWhy Audact

If you only need an ungoverned general-purpose voice agent, Audact may be more infrastructure than you need.