Aiolum Browse stories
All stories
AI co-pilot for statutory audits: scores the risk of every ledger line and prepares the auditor's working papers · Germany

AUDAVIS: Product, team and how it was built

Small teamFundedB2BAudit software

AUDAVIS (AI co-pilot for statutory audits: scores the risk of every ledger line and prepares the auditor's working papers, Germany): 6 audit firms quoted as users, November 2026. Evidence grade E: no revenue figure was checked. Team: 3 co-founders when the seed round was announced in January 2025 (Munich Startup); later headcount not published. Launched February 2025; still running.Added 7 November 2026

Outcome
6 audit firms quoted as users, November 2026
Evidence
E no revenue figure was checked
Launch
February 2025
Status
Still running
Team
3 co-founders when the seed round was announced in January 2025 (Munich Startup); later headcount not published
Outside funding
Seed round of more than EUR 1.1 million led by High-Tech Gründerfonds, January 2025 (as reported by Munich Startup)
Main channel
Not stated
No revenue figure

What is not verified

No revenue figure was found. We looked for a payment-processor profile, a founder interview with revenue, and a filing in the company register, and found none. Munich Startup reported on 15 January 2025 that the company had "Absichtserklärungen und Vorverträge im sechsstelligen jährlichen Umsatzwert" (letters of intent and pre-contracts with a six-figure annual value) (translation); that is a pipeline before launch, not revenue. The headline, six audit firms quoted as users, rests on the company's own home page. A reader could look for the annual accounts of Audavis GmbH (HRB 286180) in the German company register or for a statement by one of the named firms. This story is grade E: the business is real and its product, price and customers are described from public pages, but no revenue figure was checked. Aiolum shows it as an idea worth studying, not as proof that it makes money.

AUDAVIS is Munich software for the statutory audit, the yearly check by an independent auditor that a company's annual accounts are correct. It reads the client's general ledger, the full list of booked transactions, from more than 20 ERP and accounting systems, gives each booking line an AI risk score, and prepares the working papers, the documents in which an auditor records each test and its result. The English pricing page lists three plans on 7 November 2026, at EUR 160, EUR 220 and EUR 350 per user per month.

Three co-founders started the company in 2023: Philipp Roebruck, an auditor who worked nine years at a Big Four firm, Otrek Wilke and Benjamin Aunkofer. Munich Startup reported on 15 January 2025 that they had raised a seed round of more than EUR 1.1 million led by High-Tech Gründerfonds; in February 2025 the founders said the platform was open only to selected auditors.

No revenue figure was found. The headline is a status read on the home page on 7 November 2026: auditors from six named audit firms, among them Crowe BPG and Moore TK, are quoted as users. An archived copy shows that the paid plans were already listed on 20 January 2026, at lower prices.

AUDAVIS: 6 audit firms quoted as users, November 2026. AI co-pilot for statutory audits: scores the risk of every ledger line and prepares the auditor's working papers, Germany, team of 2 to 10, evidence grade E.
The story in one card. Share it with the address of this page.
01

How much revenue does AUDAVIS make?

Each row shows the evidence grade and the source numbers from the list at the end of this page. The grades are explained in How Aiolum verifies.

A funded three-founder company in Munich with public prices and named audit-firm users, and no revenue figure published
FactValueGradeHow we knowSources
Audit firms quoted as users, November 20266 (Dr. Dienst & Partner, Moore TK, Gebr. Scheurer, NorCon Revision, Crowe BPG, smartaudit & advisory)CHome page, read 7 November 2026: seven testimonials, each with an auditor's name; six give the firm. One says: "AUDAVIS wertet bei uns den gesamten Buchungsstoff vollständig aus und nicht nur stichprobenartig" (at our firm AUDAVIS evaluates the entire booking data completely, not only by sampling) (translation). No firm statement outside the company's site was found.[1]
Founded2023, MunichCMunich Startup profile: "Gründung 2023" (founded 2023) (translation). The VCNews list of German AI startups also gives 2023 and the stage "Seed".[6, 7]
FoundersPhilipp Roebruck (CEO), Otrek Wilke (CTO), Benjamin Aunkofer (Chief AI Officer)CMunich Startup, 15 January 2025, gives the roles. The interview of 7 February 2025 says Roebruck worked "neun Jahre als Wirtschaftsprüfer" (nine years as an auditor) (translation) at a Big Four firm. The imprint lists all three as managing directors of Audavis GmbH, Amtsgericht München HRB 286180.[4, 5, 8]
Team3 co-founders; later headcount not publishedCThe home page and Munich Startup show the three founders. The founders told Munich Startup in February 2025 that they prefer "einen kleinen, aber sehr schlagkräftigen Kern" (a small but very effective core team) (translation) and build all software in-house. No source gives a later headcount.[1, 4, 5]
Fundingseed round of more than EUR 1.1 million, January 2025CMunich Startup, 15 January 2025: led by High-Tech Gründerfonds, with business angel Martin Ostermayer and private investors from the Baystartup investor network.[4]
Prices, November 2026EUR 160, EUR 220 and EUR 350 per user per monthCCompany's English pricing page, read 7 November 2026.[2]
Prices, January 2026EUR 120, EUR 154 and EUR 210 per user per monthCInternet Archive copy of the English pricing page, 20 January 2026, with the same plan names (Basic, Professional, Enterprise).[3]
Price change, January to November 2026+33.3% (Basic), +42.9% (Professional), +66.7% (Enterprise)calcCalculated: 160/120, 220/154 and 350/210, each minus 1.[3, 2]
What the AI doesa risk score for every booking line, from more than 40 rule-based tests and AI modelsCHome page: ">40 regelbasierte Tests und KI Modelle analysieren Ihren gesamten Buchungsstoff" (more than 40 rule-based tests and AI models analyse your entire booking data) (translation). The features page says every transaction line is analysed for risk at the start of the audit.[1, 9]
Company counters, November 2026600+ engagements, 3.6+ terabytes of data, 20+ ERP and accounting systemsCHome page section "In Zahlen": "600+ Erfolgreiche Mandate" (600+ successful engagements) (translation). The section says the numbers show the technology and "die praktische Erfahrung, die wir in datengetriebene Prüfungen einbringen" (the practical experience we bring to data-driven audits) (translation), so they may include work from before AUDAVIS.[1]
Time from the selective launch to the November 2026 status21 monthscalcCalculated: February 2025 (platform open to selected auditors) to November 2026.[5, 1]

How to read this table: grade C is the company's own page or an article that repeats the company's statements; grade X is calculated on this page. No value is grade A or B, and no row is a revenue figure.

AUDAVIS prices read on the company's English pricing page on 7 November 2026, in euros per user per month, not converted to US dollars (grade C)
PlanPer month, billed yearly (US$)Per month, billed monthly (US$)Limit
Basicnot stated160 per userup to 2 GB of raw data per client; up to 20 users; onboarding training; standard support
Professionalnot stated220 per userup to 8 GB per client; up to 100 users; OCR document matching, sampling, analyses in natural language; priority support
Enterprisenot stated350 per userup to 20 GB per client; unlimited users; one extra read-only user for every 5 users; individual support

All values in this table are euros. The page shows one price per plan and does not say whether it is billed monthly or yearly, so the yearly column says "not stated". The page says the cost per user per month equals "just one billable hour". On 20 January 2026 the same three plans cost EUR 120, EUR 154 and EUR 210 per user per month (Internet Archive copy).

02

When did AUDAVIS launch, and how did it grow?

From a 2023 company and an archived product page in 2024 to a seed round, a selective launch and public prices. Each entry carries its grade.

  1. 2023CAudavis GmbH is founded in Munich by Philipp Roebruck, Otrek Wilke and Benjamin Aunkofer. [6, 7, 8]
  2. 1 March 2024CThe earliest archived home page describes a data-analysis platform that prepares data-driven working papers and extends risk assessment to the whole ledger with AI. It shows no prices. [11]
  3. 15 January 2025CMunich Startup reports a seed round of more than EUR 1.1 million led by High-Tech Gründerfonds, and third place in the Munich business plan competition 2024. [4]
  4. 7 February 2025CThe founders tell Munich Startup that the core features are open to selected auditors and that they plan to open the platform to the whole market at the end of the second quarter. [5]
  5. 20 January 2026CAn archived pricing page lists three paid plans at EUR 120, EUR 154 and EUR 210 per user per month. [3]
  6. 7 November 2026CThe pricing page lists EUR 160, EUR 220 and EUR 350 per user per month; the home page quotes auditors from six named firms. [2, 1]
03

Why did AUDAVIS work?

Four causes, in the order they mattered. Each one is a reusable lesson.

01

Test every line, not a sample

Auditors normally draw a statistical sample of bookings. AUDAVIS scores every booking line, and one quoted auditor writes that it narrows more than one million lines to a few hundred that matter for the audit. The auditor still decides; the score is an extra input to the risk assessment.

02

Automate the data engineer

The founders told Munich Startup that other audit-analytics vendors need data analysts to prepare each client's data. AUDAVIS automates that step: the company says setting up a client takes less than an hour and the working papers are ready on the first day of the audit.

03

Founders from the buyer's profession

The CEO is a qualified auditor from a Big Four firm, and the Chief AI Officer has years of data analysis for audits at large audit firms. The working papers follow the audit standards that auditors cite (ISA 315, ISA 240, ISA 520), which the how-it-works page names.

How it was built

These details come from the company's own pages and from what the founders told Munich Startup. They are grade C. [1, 4, 5, 9, 10]

  • A separate, isolated cloud environment is created for each client engagement, because auditors must keep client data confidential.
  • Data import from more than 20 ERP and accounting systems; the founders name SAP, Sage and DATEV.
  • More than 40 rule-based tests and AI models; the home page counts 3 dedicated AI systems, 48 data-driven audit procedures and 37 journal entry tests.
  • The Chief AI Officer told Munich Startup that the AI system has several layers and runs as distributed computing with federated deep learning.
  • All software is built in-house; the public website is built with Next.js, as its page source shows.

Tools named in the sources

Each one opens its page in the Tools tab, with every story that uses it.

04

Who tried this before AUDAVIS, and why did they fail?

Three companies in our database of failed startups promised to automate accounting work or to give non-technical people data analysis without data scientists. The values below are from that database and are not independently verified.

Died 2014 to 2019Burned US$100M

ScaleFactor

The promise

Automated bookkeeping and financial analytics for small and mid-sized businesses.

Why it died

The database says it relied too much on automation that did not work well enough and promised more than the product did.

Product/Tech Failure
Died 2006 to 2011Burned US$33M

NetBooks

The promise

Online accounting for small businesses.

Why it died

It could not stand out against QuickBooks and its costs were high.

Competition
Died 2012 to 2015Burned US$2M

Zillabyte

The promise

Turn raw data into decisions for people without data-science skills.

Why it died

Better-funded analytics platforms won, and it had no narrow market focus.

Competition

The failed startups come from our database of 1,749 dead startups, retold here in our own words. Its entries are AI-assisted summaries of public sources, so these cards carry no evidence grade.

What changed

ScaleFactor sold automation that, in the database's account, still needed people behind it, and Zillabyte sold general data analysis to everyone. AUDAVIS chose one profession with fixed procedures: the audit standards define which tests an auditor runs, so the platform can prepare the same working papers for each client automatically.

Two things are newer. AI models can now give a risk score to every booking line at low cost, which makes testing the whole ledger practical instead of a sample. And the shortage of auditors in Germany, which the founders cite from a 2022 survey, makes firms pay for software that saves audit hours: the pricing page compares the monthly price per user with one billable hour.

05

Can one person copy AUDAVIS now?

Not as a solo project. The idea of automating one profession's standard procedures, yes.

AUDAVIS has a qualified auditor, a cloud-security engineer and an AI specialist as founders, seed funding, and a buyer that must trust it with confidential client data. One person cannot easily repeat that. The narrower pattern is open: pick a profession whose work follows written standards, automate the data preparation and the standard tests, and price it against the hourly rate of that profession.

Where it is still open: Similar tools for tax auditors of small firms, for internal audit in mid-sized companies, or for audits in countries with other accounting standards.

  1. 1Choose a profession with written standards that define the routine tests, such as auditing, tax review or quality control.
  2. 2Automate the data import from the systems that its clients use, so no data engineer is needed.
  3. 3Produce the documents the professional must keep, in the format they already use.
  4. 4State the price in the buyer's own unit, such as one billable hour per user per month.

What could go wrong for AUDAVIS?

The risks are the mirror image of the causes above.

What is not verified

No revenue figure was found. We looked for a payment-processor profile, a founder interview with revenue, and a filing in the company register, and found none. Munich Startup reported on 15 January 2025 that the company had "Absichtserklärungen und Vorverträge im sechsstelligen jährlichen Umsatzwert" (letters of intent and pre-contracts with a six-figure annual value) (translation); that is a pipeline before launch, not revenue. The headline, six audit firms quoted as users, rests on the company's own home page. A reader could look for the annual accounts of Audavis GmbH (HRB 286180) in the German company register or for a statement by one of the named firms.

Trust signals on the home page

The row "Partnerschaften & Auszeichnungen" (partnerships and awards) shows five logos. High-Tech Gründerfonds and the Munich business plan competition match the Munich Startup article, and the NVIDIA logo image reads "Inception Program". A plain Microsoft logo and the logo of the KI Bundesverband (German AI association) have no text that says what the relation is, and no source we read explains them. The counter "600+ Erfolgreiche Mandate" (600+ successful engagements) is introduced as showing the team's practical experience, so it is not clear whether it counts only engagements on AUDAVIS since the 2025 launch.

Competition

The founders told Munich Startup that the Big Four audit firms build their own tools and that other audit-analytics vendors need manual data work. On this site, Punchcard (AI agents that test audit evidence and prepare workpapers, grade E) sells a similar product to audit firms.

Dependence on model providers

No source we read names an outside AI model provider. The home page describes 3 dedicated AI systems and more than 40 rule-based tests and models, and the Chief AI Officer told Munich Startup in January 2025 that the system uses federated deep learning. The Professional plan's analyses in natural language may use a large language model; the pricing page does not say which.

Team size

The company was started by three co-founders, who were the team shown when the seed round was announced in January 2025. No source gives a later headcount, and no source suggests more than ten people.

Sources

The grade next to each source is the best grade that the source can support. "Read in full" means that the text of the page was downloaded and the numbers were checked in it.

  1. C AUDAVIS home page, read 7 November 2026
    Read in full (German). Source for the product, the AI risk score, the counters, the testimonials with firm names, the founders and the partner logos (the logo images were also opened).
  2. C AUDAVIS pricing page (English), read 7 November 2026
    Read in full. Source for the three plans, their limits and the billable-hour comparison.
  3. C Internet Archive copy of the AUDAVIS English pricing page, 20 January 2026
    Read in full. The earliest archived pricing page; source for the January 2026 prices.
  4. C Munich Startup, 15 January 2025: Audavis sichert sich Seed-Finanzierung
    Read in full (German). Source for the seed round, the investors, the founders' roles, the letters of intent and the business plan competition.
  5. C Munich Startup, 7 February 2025: Audavis: KI-Co-Pilot für die Wirtschaftsprüfung
    Read in full (German). Interview; source for the founding story, the selective launch, the market plan, the competition and in-house development.
  6. C Munich Startup startup profile: Audavis GmbH
    Read in full (German). Source for the founding year 2023 and the B2B model.
  7. C VCNews: Startup-Liste, sector KI (German AI startups)
    Read the AUDAVIS entry in the copy downloaded by the pipeline on 7 November 2026. Source for the founding year, the stage and the description.
  8. C AUDAVIS Impressum (imprint), read 7 November 2026
    Read in full (German). Source for the company name, the managing directors and the register number.
  9. C AUDAVIS features page (English), read 7 November 2026
    Read in full. Source for the per-line risk analysis, the isolated client environments and the subledger procedures.
  10. C AUDAVIS how it works (English), read 7 November 2026
    Read in full. Source for the first-day working papers and the audit standards named.
  11. C Internet Archive copy of audavis.ai, 1 March 2024
    Read in full (German). The earliest archived home page.
  12. C Database of failed startups, local export
    Local export. AI-written summaries, not independently verified. Source for the failed predecessors.

Run this company? Send corrections or a verified figure to aiolum@livyden.com

Back to all stories