Audavis AI
Audavis AI is a German audit platform that reads a client's complete general and subsidiary ledgers instead of a sample, scores the inherent risk of every single booking line and produces standard audit working papers automatically.
What is Audavis AI?
Audavis AI is an audit analytics platform published by Audavis GmbH, a Munich company whose three managing partners combine Big Four audit data analytics, statutory audit practice and cloud security engineering. Its central claim is coverage: rather than drawing a sample, the platform ingests a client's complete general ledger and subsidiary ledgers and examines every posting.
Once an engagement is imported, more than forty rule-based tests and machine-learning models score the inherent risk of each individual booking line, adding a quantitative layer to the auditor's own risk assessment through measures such as item risk and materiality thresholds. The vendor presents this as a complement to professional judgement, not a substitute for it. Standard deliverables are then produced automatically: interactive balance sheet and income statement views, KPIs, journal entry tests, counter-account analyses, analytical procedures and substantive tests on the general ledger, plus receivables, payables, fixed assets and inventory analyses on the subledgers.
A second layer is the engagement copilot. Auditors drag internal-control documentation, process descriptions and contracts into the platform and can then question that context in natural language or ask for working papers to be drafted. The same conversational interface, marketed as AUDAVIS GenAI, runs ad hoc queries and semantic searches in German or English and saves the results as custom dashboards.
The product is sold across four use cases on one codebase: statutory audit, tax audit, internal audit and CFO advisory. On the tax side it consumes German GoBD data exports and adds Benford, distribution, time-series and gap analyses. On the internal audit side it covers Purchase-to-Pay and Order-to-Cash end to end, with three-way matching, cut-off testing and segregation-of-duties checks drawn from ERP systems such as SAP, Microsoft Dynamics and Sage.
Each engagement is provisioned into its own walled-off cloud environment, which the vendor frames as its answer to professional confidentiality and tax secrecy. Setup is claimed to take under an hour with no technical skills required, with results available on the first day of fieldwork. Headline figures on the site claim more than twenty supported ERP and accounting systems, over six hundred completed engagements and 3.6 terabytes of processed data. The ISA references throughout the site describe the procedures the software automates, not accreditations held by the vendor.
What it does
- Analyse a client's complete general ledger and subsidiary ledgers instead of a sample
- Score the inherent risk of every individual booking line using rule-based tests and AI models
- Generate standard audit working papers and dashboards automatically from the imported data
- Run journal entry testing, counter-account analyses, cut-off and expectation-value checks
- Query the accounting data in natural language, in German or English, through the AUDAVIS AI chat
- Test operational processes end to end from ERP data, including three-way matching and segregation of duties
- Comment on working papers and export them as PDF into the engagement documentation
When to use Audavis AI / When not to
A quick filter to help you decide if Audavis AI is the right fit.
When to use Audavis AI
- Statutory audit firms, from boutique practices to large firms, whose engagement teams want full-population testing rather than sampling
- Internal audit departments moving towards continuous auditing and data-driven audit planning
- Tax inspectors and field auditors in tax administrations working from GoBD data exports
- Tax advisers and corporate tax departments preparing for or accompanying a tax audit
- CFOs and finance teams that want to find errors in their own books before the external auditor does
When not to use Audavis AI
- Teams looking for a free or trial-based tool: the cheapest published plan starts at EUR 160 per user per month
- Developers who need a public API: no API documentation exists and none is offered
- Anyone needing mobile or offline access, since the product ships only as a web application
- Buyers who require a published data processing agreement, a named hosting jurisdiction or a stated retention period before signing
- Users outside finance and audit: the platform has no purpose beyond accounting data analysis
How to use Audavis AI
A typical end-to-end flow, from setup to results.
- Request access through the contact form on the website; there is no self-service sign-up
- Log in to the web application hosted at app.audavis.io
- Create an engagement, which automatically provisions a dedicated, walled-off cloud environment sized for the job
- Import the client's accounting data: general ledger and subsidiary ledgers, or a GoBD export for a tax audit
- Drag and drop context documents such as internal-control documentation, process descriptions and contracts into the engagement
- Let the platform run: full analysis is claimed to complete in under an hour
- Review the AI risk score to see which booking lines and items deserve attention
- Open the pre-built working papers and dashboards for the general ledger and the subledgers
- Ask ad hoc questions in German or English through the AUDAVIS AI chat and save the answers as custom dashboards
- Comment on each working paper inside the platform, then export it as PDF into your engagement documentation
Pros & Cons
Pros
- Full-population testing rather than sampling, which the platform turns into a shortlist of genuinely audit-relevant items
- Built by practising auditors, and openly aligned with the standard procedures auditors already perform
- Fast to start: an engagement is claimed to be set up in under an hour, with results on day one and no technical skills required
- Transparent public pricing with three named tiers, a per-user monthly amount and stated data volume allowances
- Dedicated, isolated cloud environment per engagement, presented as the answer to professional confidentiality and tax secrecy
- Broad accounting source coverage, with SAP, Microsoft Dynamics, Sage, Diamant Software, Varial, Wolters Kluwer and BMD Business Software shown as supported systems
- Named, checkable testimonials from auditors at identified German audit firms, and an institutional investor on the cap table
Cons
- No free plan and no announced free trial: the entry point is EUR 160 per user per month, reached only through a contact form
- No terms and conditions page exists anywhere on the site, so the contractual framework is not public
- No data processing agreement is published or offered, which is unusual for a product handling third-party accounting records
- The privacy notice covers the website only; there is no public document describing how engagement data is processed
- No hosting country, region or provider is named, no subprocessor list is published and no retention period is stated
- The German and English pricing pages contradict each other on data allowances, quoting 20 GB and 50 GB against 8 GB and 20 GB for the same tiers
- No API, no mobile app and no browser extension: the product is a web application only
Pricing & Plans
There is no free plan and no announced free trial. The cheapest published entry point is the Basic tier at EUR 160 per user per month; Professional is EUR 220 per user per month and Enterprise, marked as recommended, is EUR 350 per user per month. The vendor frames this as roughly the cost of a single billable hour per user per month. Prices are stated exclusive of any mention of VAT, commitment period or annual billing, and every tier button leads to the contact form rather than to online checkout.
- onboarding training
- up to 2 GB of raw data per engagement
- a maximum of 20 users in the organisation
- standard support
- onboarding plus regular refresher training
- up to 20 GB of raw data per engagement according to the German page (the English page states 8 GB)
- a maximum of 100 users
- priority support
- and Professional features such as OCR document matching
- sampling and natural-language analytics
- onboarding plus refresher training
- up to 50 GB of raw data per engagement according to the German page (the English page states 20 GB)
- unlimited users
- individualised support
- all Professional features
- and one additional read-only user for every five paid users
Data, GDPR & hosting
A consolidated view of how Audavis AI handles your data.
GDPR overview
GDPR implementation is explicit but narrow. The privacy notice, valid as of June 2025, states that personal data is processed in accordance with the GDPR and names the controller, AUDAVIS GmbH, with datenschutz@audavis.ai as the contact point. Data subject rights are listed individually: access (Art. 15), rectification (Art. 16), erasure (Art. 17), restriction (Art. 18), portability (Art. 20), withdrawal of consent (Art. 7(3)) and complaint to a supervisory authority (Art. 77). SSL encryption and unspecified technical and organisational measures are mentioned. No data protection officer is named, and as an EU-established company no Article 27 representative is required or designated. Crucially, the notice addresses website visitors only: there is no published statement covering the accounting data processed inside the platform, no hosting location, no subprocessor list and no retention period.
Who owns the data?
The published privacy policy names AUDAVIS GmbH in Munich as the controller, but it covers only the website: visitor log data, the contact form and cookies. No terms and conditions exist anywhere on the site, so there is no public contract clause stating who owns the accounting data an audit team uploads, what the vendor may do with it, or what happens to it when an engagement ends. The product pages promise that each engagement runs in its own isolated environment and that data is processed separately per engagement and per instructing party, invoking professional and tax secrecy. That is an isolation claim, not an ownership clause. Buyers should obtain the contractual terms directly.
Reuse rights
For the website, the policy states that log data is processed on legitimate interest under Article 6(1)(f) GDPR and contact-form data on consent under Article 6(1)(a), and that personal data is not passed to third parties unless the visitor has expressly consented or disclosure is legally permitted. Cookies and analytics tools are used to measure usage. For the platform itself, no equivalent document is published: the site describes an automatically provisioned, walled-off cloud environment per engagement and states that client data does not leave that environment, but it never says whether the customer may reuse outputs freely, nor under what licence. Notably, the site is completely silent on whether customer data is used to train AI models, in either direction.
Data retention & training
Hosting summary
Hosting is described functionally, never geographically. The site says that when an engagement is created a dedicated, protected cloud environment is provisioned automatically, sized to the requirements of the job, and that data is processed separately per engagement and per instructing party so that professional confidentiality and, for tax work, tax secrecy are preserved. One solutions page states plainly that client data does not leave that environment. What is missing is everything a compliance reviewer would need: no country, no region, no cloud provider and no data centre is ever named, no subprocessor list is published, and the privacy notice covers only the website rather than the platform. The website itself resolves to an IP address geolocated in the United States behind a content delivery network, but that describes the marketing site and says nothing about where engagement data is stored. Buyers with residency requirements should ask the vendor directly and obtain the answer in writing.
Things to keep in mind
Risks and trade-offs to weigh before adopting Audavis AI.
- No terms and conditions exist on the site and no data processing agreement is published, so the contractual position on third-party accounting data must be obtained privately before any upload
- No hosting country, region or provider is named and no subprocessor list is published, which matters for engagements bound by professional secrecy
- The site never states whether customer data is used to train AI models, neither confirming nor excluding it
- No retention period is published, and with no terms page there is no public rule on returning or deleting engagement data at the end of a contract
- Automated risk scoring can encourage auditors to follow the machine's shortlist and quietly narrow their own professional scepticism; the vendor itself positions the score as a complement to judgement, not a replacement
- The German and English pages contradict each other on data allowances, and the how-it-works page still carries an unremoved internal editorial note, both signs that published detail should be verified rather than trusted
- Standards such as ISA 240, 315, 330 and 520 and the GoBD export format are what the software helps apply; they are not certifications held by the vendor, and no ISO, SOC or comparable certification is claimed anywhere on the site
Setup & Integrations
Technical difficulty
Low on paper. The vendor claims an engagement is configured in a few clicks with no prior technical knowledge, the cloud environment is provisioned automatically, context documents are added by drag and drop, and onboarding training is included in all three tiers. The real effort sits upstream, outside the tool: obtaining a clean, complete accounting export from the client, whether a general ledger extract, subledger data or a GoBD file. Firms that already handle such exports routinely should find setup straightforward; those that do not will spend their time there rather than in the software.
Deployment
Integrations
Supported languages
Behind Audavis AI
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Audavis AI actually do?
Who is it for?
How long does it take to set up an engagement?
How much does it cost, and is there a free version?
Which accounting and ERP systems are supported?
Which languages does the tool work in?
Is there an API or a mobile app?
Where is the data hosted?
Is customer data used to train AI models?
Who publishes the tool?
Should you pick Audavis AI?
Audavis AI is a narrow, well-aimed product rather than a general-purpose assistant. Its proposition, analysing the whole population of postings instead of a sample and ranking every booking line by risk, addresses a real constraint of audit work, and the alignment with the procedures auditors already perform makes it legible to its buyers rather than exotic. The claim that an engagement is up and running within an hour, with no technical skills required, is unusually concrete for this market, and the published per-user pricing spares prospects the usual guessing game. Named testimonials from auditors at identified German firms give the coverage claims a degree of external weight.
The reservations are about documentation rather than capability. For a platform whose whole purpose is to process the accounting records of third parties, the public paperwork is remarkably thin: there is no terms and conditions page at all, no data processing agreement, no named hosting jurisdiction, no subprocessor list and no retention period. The privacy notice, current as of June 2025, covers website visitors only. Whether customer data feeds model training is never addressed either way. The German and English pricing pages also disagree on the data allowances attached to the Professional and Enterprise tiers, which is the kind of inconsistency that should be settled in writing before a purchase order.
None of that is disqualifying, and much of it is normal for a young vendor still building out its public materials. It does mean, however, that the contractual and data-protection questions have to be asked directly of the sales team rather than read off the site. Audit firms evaluating the platform should treat the commercial demonstration and the legal annexes as two separate exercises.
- Choosing a selection results in a full page refresh.
- Opens in a new window.