BEAMON AI alternative: what the price includes

BEAMON AI is the legal AI workspace built by BRYTER GmbH in Frankfurt, with list prices from EUR 99 per user and month on an annual plan. Several of the legal databases it advertises require a second subscription. CASUS covers the same tasks, works inside the Word add-in on the open document, and includes DACH research with no separate publisher contract.

What BEAMON AI is and where it comes from

BEAMON AI is a young product on an older foundation. BRYTER GmbH was founded in 2018 as a platform for rule-based no-code automation of legal processes; BEAMON AI arrived in May 2025 as an AI layer on top, which is why the trade press called it "Bryter 2.0". That explains the shape of it: the strength lies in process automation, and legal document work is the later addition.

In concrete terms the offering has three building blocks: Assist for drafting, review and research, Extract for structured data out of large document sets, plus a Word plugin and the rule-based workflows. iManage and SharePoint are connected.

The review module deserves to be taken seriously. BEAMON describes it as "high-speed document review with clause comparison, risk highlighting and policy checks aligned to your playbooks" (beamon.ai, as of August 2026). Describing this provider as a pure research tool would be describing it wrongly.

Three things are not stated, and they are what a review turns on in practice: analysis from the perspective of a named contracting party, a graded severity per finding, and a playbook comparison that returns a percentage.

CASUS took the opposite route and was built for work on the legal document from the start. Seven modules each carry their own job: AI Chat with Agent Mode, Risk Review, Benchmark, Dossiers, Legal Research, AI Data Room and Proofread. The comparison runs along six axes, and on five of them the two providers differ.

What the list price covers and what needs a second subscription

The price list names three tiers: Single, Team and Pro, at EUR 99, 149 and 199 per user and month on an annual plan (as of August 2026). Workflows and single sign-on are available only in the Pro tier. The row below is more interesting: under "integrations and publishers" the pricing page lists six positions, and for four of them it states explicitly that they cost extra:

Research across "more than 80 legal databases" is advertised. Which of them belong to which tier is not stated; the two named explicitly, Otto Schmidt and Haufe Lexware, both carry a second subscription. For a firm doing the maths, the licence fee is where the calculation starts, not where it ends.

CASUS costs EUR 155 per user and month, or EUR 130 billed annually, which is EUR 1,560 per year (as of August 2026). Research covering Austrian, German and Swiss law is part of that price, with no publisher contract alongside it. The pricing overview sets out both payment options.

The work happens in the Word add-in

Both providers offer a Word integration, and neither means the same thing by it. At CASUS roughly 80 percent of all usage takes place in the Word add-in rather than in the web app. That is an operating figure, not positioning, and it explains why the modules are fully available there.

In practice: Risk Review, Benchmark, Proofread and the AI Chat including Agent Mode all work on the open document. Agent Mode inserts clauses, rewrites and extends them, respecting structure, numbering and formatting, without copy and paste. Add-in actions come on top: Translate carries a document across without altering formatting or legal meaning, Anonymize identifies personal data before a document leaves the firm.

A concrete case: a framework supply agreement arrives from the other side, and the review is promised by the afternoon. Risk Review runs over the open document, findings appear alongside the text, the proposed liability wording moves into place with one click, and Proofread checks the cross-references before it goes out. No export, no second window, no file leaving the firm.

For daily work that is the difference between a tool that has to be opened and one that is already where the contract sits.

Review with party perspective, severity and playbook match

A contract review is only useful once it is clear whose side it was written from. Risk Review identifies the contracting parties and analyses from the perspective of one of them. Every finding carries an attribution, a relevance and a severity from low to high, along with a concrete wording option that can be taken straight into the document. From live operation: a review takes one to three minutes and returns around twenty findings on average, figures that come from usage rather than a datasheet.

Next to it sits Benchmark, which holds a document against a reference standard, either an in-house playbook or best practice for SPAs, NDAs and DPAs. It names missing areas such as liability without a cap or confidentiality without a deletion duty, and expresses the match as a percentage. A playbook comparison that produces a number can be enforced as a team standard; an assessment without one invites debate.

What makes a review checkable is attribution: every finding traces back to the passage it came from. An assessment a partner has to sign off needs that route back, or it stays an opinion.

Four further modules carry their own job. Dossiers keeps context open across up to 100 documents without re-uploading files. The AI Data Room analyses hundreds of documents in parallel and returns a clause matrix across liability, IP, SLA and termination, prioritised by risk. Proofread checks cross-references, definitions, annexes, numbering and open placeholders. And the AI Chat with Agent Mode answers questions about the document and jumps to the passage in question.

Research without a second subscription

Legal Research draws on more than 1.8 million court decisions and statutory provisions from Switzerland, Germany and Austria. For Austria that means the entire RIS, plus the Supreme Court, the Administrative and Constitutional Courts, Findok, the financial market authority and the competition authority. For scale: the federal open data interface reports roughly 895,000 documents of case law alone in the RIS (data.bka.gv.at, queried August 2026).

Research runs on statutes and case law rather than general web sources, and results carry citations that can be reused. The output is more than a hit list: a legal position yields a risk analysis with its drivers, arguments for and against, and a recommendation, meaning hold the position, offer a fallback or propose an additional clause. Because research sits in the same add-in as the review, the result lands in the document rather than in a note beside it.

One gap belongs in the picture: EU law is not covered, neither the case law of the Court of Justice and the General Court nor EUR-Lex. Anyone working with Union law daily needs a separate source for it.

Security: where the computation happens and who sees plaintext

On security the two providers do not differ on whether it is taken seriously, but on what they disclose.

BEAMON undertakes not to pass customer data outside the EU or EEA and rules out training use. The disclosure stops there: which region hosts the data, where AI inference runs, and who can reach client content in plaintext appear on none of the provider's pages.

Those are precisely the points section 43e paragraph 4 BRAO and section 40 paragraph 3 no. 4 RL-BA 2015 turn on: where the work is performed, and who can see the plaintext. CASUS answers by name. Client content sits on Google Cloud in Zurich, encrypted with AES-256 at rest and TLS 1.2+ in transit. AI inference runs solely through Vertex AI in Belgium and Azure OpenAI in Switzerland North and Sweden Central, with none in the United States. An opt-out from abuse monitoring is in place for Azure, so there is no thirty-day buffer and no human review by Microsoft.

Four undertakings are contractual and named individually: no use for AI training, no processing in third countries, no access by cloud provider staff, no mixing of matters. Plaintext access on the CASUS side is limited to technically necessary troubleshooting, and the full list of subprocessors is published in the trust centre. The details are set out there.

For procurement the difference is concrete: undertaking an economic area does not answer the question of where processing happens, while naming the regions does.

Professional secrecy is the real selection test

In Austria and Germany, choosing an AI provider is a professional conduct question, and both jurisdictions frame it through the lawyer rather than the vendor.

In Austria, section 40 paragraph 3 of the RL-BA 2015 permits external data processing only where five conditions are met cumulatively: client interests preserved, careful selection of the provider, a contractual undertaking to give immediate notice of a search of premises, technical and organisational measures reflecting the state of the art, and information to the client about the categories of providers involved.

The third condition rarely appears in contracts. The Austria annex to the CASUS terms contains it verbatim, committing the company to inform the client without delay in the event of a seizure, a search of premises or any other official access, so far as legally permitted. No other provider in the Austrian field publishes such an undertaking. The competitor's page for bar members cites unlimited confidentiality, a data processing agreement under Article 28 GDPR, EU processing and a ban on training use, plus a signed bar association checklist that is not publicly available. The notification duty is not among them.

In Germany the equivalent test is section 43e of the Bundesrechtsanwaltsordnung, whose fourth paragraph permits services performed abroad only where comparable protection of professional secrets exists, and whose fifth paragraph requires prior client consent for matter-specific services. The Germany annex to the CASUS terms classifies the service as firm-wide infrastructure across all matters, so that consent requirement does not arise. Worth noting for paragraph 4: BRYTER GmbH holds subsidiaries in New York and London while the location of inference is not stated.

For the file there is something no marketing text replaces: firms can request written confirmation of auxiliary-person status at any time, and equally a confirmation that infrastructure and AI providers have no human plaintext access to client content.

BEAMON AI and CASUS compared

For a wider view of the Austrian market, see the comparison of legal AI providers in Austria, and a separate article on the other provider promoted by the bar association.

Testing takes less time than evaluating a price list. CASUS can be trialled for fourteen days at no cost and without a minimum term, sensibly on a real contract rather than a sample. Anyone curious about a review from party perspective can create an account and upload the next draft already on the desk.

Which tool for which firm

CASUS fits the ordinary case. A firm working daily on contracts, pleadings and review instructions finds six arguments. The work runs in the Word add-in on the open document, where roughly 80 percent of usage happens. The review is written from a party perspective, graded by severity and traceable back to the passage, which is what makes it fit to be signed off. The playbook comparison returns a percentage, which is what turns a house standard into something enforceable. Research across Austrian, German and Swiss law is included, so the list price is the whole calculation. The place of inference is named and lies outside the United States. And the professional secrecy undertakings sit in the contract rather than in a brochure, with written confirmation available for the file.

BEAMON AI fits three particular situations. A firm that already subscribes to Otto Schmidt or Haufe Lexware carries existing publisher content forward, and the second-subscription point falls away. A house with an iManage estate keeps its precedents where they are. And a team with a real need for rule-based process automation, for recurring assessments or standardised intake paths, finds BRYTER's origin and continuing strength.

All three describe an existing investment or a requirement outside document work. A firm deciding without such a commitment is deciding about daily work on the document, and there the answer is a different one.

Frequently asked questions

What does BEAMON AI cost?

The public price list names three tiers at EUR 118, 176 and 234 per user and month billed monthly, or EUR 99, 149 and 199 billed annually (as of August 2026). Workflows and single sign-on are Pro-only, and several publisher and register integrations require an additional subscription elsewhere.

Is the RIS integration included in every tier?

The published material does not answer this. The RIS integration is described on the page for Austrian bar members, while the pricing table under "integrations and publishers" lists German sources only. Before signing, this is a question for the sales team.

Does CASUS meet the requirements of section 40 paragraph 3 RL-BA 2015?

The five conditions address the lawyer, not the provider; the firm has to satisfy them. For the third it needs a contractual undertaking from the provider, and that undertaking appears verbatim in the Austria annex to the CASUS terms.

Which Austrian legal sources does CASUS cover?

The entire RIS including case law from the Supreme Court and the Administrative and Constitutional Courts, plus Findok, the financial market authority and the competition authority. EU law is not covered.

Can CASUS work inside Microsoft Word?

Yes, and that is where most usage happens. Risk Review, Benchmark, Proofread and the AI Chat with Agent Mode run in the add-in on the open document, applying changes in the correct format.

Is client data used to train AI models?

No, neither by CASUS nor by the model providers. An opt-out from abuse monitoring is in place for Azure, so there is no thirty-day buffer and no human review. Both providers rule out training use; the difference lies in where processing happens.