dnl
dnl is a German disclosure management platform for external auditors reviewing the notes to financial statements and sustainability reports. Its ai/checklist and ai/numbers products draft checklist answers, tie figures back to evidence and flag inconsistencies.
What is dnl?
dnl is an AI platform for disclosure management, built for the external auditors who review the notes to financial statements and, increasingly, sustainability reports. The Berlin vendor's argument is that traditional audit forces a trade-off between thoroughness and efficiency, and that AI can remove it.
The platform ships as two live products and one announced. ai/checklist turns disclosure checklists into dynamic digital workflows: embedded audit agents draft the answers from the source documents, and each checklist adapts to entity size, materiality and implementation phase rather than staying a static form. The vendor claims 90% of validations are completed in real time. ai/numbers handles the figures — evidence agents perform tick-and-tie in seconds, verify mathematical accuracy on totals, ratios and calculations, and test internal consistency between related line items as well as against supporting evidence and the prior year. ai/accounts, marked coming soon, will map charts of accounts to the balance sheet and income statement.
Framework coverage spans IFRS, CSRD and ESRS, local GAAP, AASB for Australia, IFRS S1 and S2 from the ISSB, and GRI. Firms that already have their own methodology can import their proprietary checklists instead of adopting the vendor's, and the system is said to learn a firm's patterns from one engagement to the next. Results leave the platform as PDF, Excel or Word files, or through an API into existing audit management systems. Two deployment models are offered: plug-and-play SaaS, or a white-label rollout under the firm's own brand.
The limits are stated by the vendor itself. AI proposals rest on pattern recognition and historical audit data, they are advisory only, and every AI-generated element is highlighted in yellow so human validation stays explicit; final responsibility remains with the auditor. A self-declared accuracy of 97.4% on key figure identification comes from the vendor's own AI transparency statement, not from an independent test. The first product shipped in early 2022 under the earlier name Notes Auditor, and product documentation sits behind authentication.
What it does
- Auto-fills disclosure checklist answers straight from the source documents
- Verifies totals, ratios and calculations across the financial statements for mathematical accuracy
- Tests internal consistency between related line items and against the prior year
- Ties reported figures back to supporting evidence such as invoices, bank statements and third-party reports
- Monitors new versions of source documents and reopens only the affected steps
- Records every action in a complete audit trail
- Highlights all AI-generated content in yellow so reviewers can see what the machine wrote
When to use dnl / When not to
A quick filter to help you decide if dnl is the right fit.
When to use dnl
- External audit firms running disclosure and notes review on annual financial statements
- ESG assurance teams working to CSRD, ESRS, IFRS S1/S2 and GRI reporting frameworks
- Audit practices that want to keep their own methodology and import proprietary checklists
- Firms looking to deploy audit AI under their own brand through a white-label rollout
- Engagement teams facing heavy tick-and-tie work across large volumes of figures
When not to use dnl
- Buyers who need a published price list, since entry runs through a demo and a commercial pilot
- Anyone expecting a free plan or a free trial, neither of which the vendor announces
- Teams that want a native mobile app, as no App Store or Google Play listing exists
- Individuals and very small businesses working outside a structured audit practice
- Practitioners hoping to delegate judgement, since outputs are advisory and human validation stays mandatory
How to use dnl
A typical end-to-end flow, from setup to results.
- Book a demo through the website form — there is no self-service sign-up
- Pick the regional demo track that matches the firm: EU, UK, US, Australia or global
- Agree the deployment model with the vendor: plug-and-play SaaS, or white-label under the firm's own brand
- Scope a focused pilot — a small team, a defined engagement scope, a single audit cycle
- Work with the implementation and customer success teams, who accompany the pilot end to end
- Import the firm's own disclosure checklists and methodology to keep the existing approach
- Upload the source documents — Excel and PDF files — into the shared cloud workspace, with access granted by role
- Let the audit agents draft the checklist answers and run the tick-and-tie and consistency checks
- Review every AI-generated item, recognisable by its yellow highlight, and validate it yourself
- Export the results as PDF, Excel or Word, or push them through the API into the audit management system; by the second engagement the AI has learned the firm's checklist patterns and by the third the workflow is standardised
Pros & Cons
Pros
- Narrow specialisation on disclosure and notes review, which is rare in this market
- Complete, verifiable legal identity — HRB 206004 B, VAT number DE323717681 valid in VIES, named managing directors — alongside a public, dated funding history that includes European co-funding from Investitionsbank Berlin and the ERDF
- Audit trail and systematic highlighting of AI output, matching the traceability the profession requires
- Named client references and a signed testimonial, with Forvis Mazars presented as a development partner
- Stated compliance posture: ISO 27001 claimed by the vendor, GDPR, and alignment with the EU AI Act's limited-risk tier
- Choice of data residency between the EU, the United States and Australia
- Firms can keep and import their own methodology rather than adopt the vendor's
Cons
- No public price: neither a pricing page nor a single figure appears across the 107 URLs of the sitemap
- No free trial and no free plan is announced
- No terms and conditions are published, and product documentation sits behind authentication, so both the contractual and the technical detail stay private until contact
- ISO 27001 is presented as a certification without a certificate number or a certifying body
- No retention period is ever quantified in the privacy policy
- The vendor never states whether customer data is used to train its models
- The AI transparency statement is frozen at version 1.0 of 12 February 2025, is still written under the former product name Notes Auditor, and still carries an uncorrected template placeholder pointing to yourcompany.com/ai-info
Pricing & Plans
No pricing is published. dnl operates on a contact-sales basis: there is no pricing page, no amount appears anywhere on the site, and a direct check of the /pricing path returns a genuine 404. No free plan and no free trial are announced. Access begins with a booked demonstration, followed by a scoped commercial pilot on a single audit cycle; commercial terms — whether plug-and-play SaaS or a white-label deployment under the firm's own brand — are then agreed directly with the vendor. Prospective buyers should expect to contact the sales team to obtain any figure at all.
- dnl lists no tiers and no amount appears anywhere on the site
- Plug-and-play SaaS deployment — offered without a published price
- White-label deployment under the audit firm's own brand — offered without a published price
- Commercial terms are agreed after a demo and a scoped pilot
- the FAQ states the platform suits audit firms of all sizes
Data, GDPR & hosting
A consolidated view of how dnl handles your data.
GDPR overview
GDPR implementation is documented rather than merely asserted. The FAQ states that dnl is GDPR compliant for data handling, and a full Data Protection policy sets out the controller, the legal bases under Art. 6(1), data subject rights and the conditions for transfers outside the EEA. The policy is organised service by service, naming each third party used on the website. An external data protection officer has been appointed under Art. 37 — heyData GmbH, reachable at datenschutz@heydata.eu — while no Art. 27 representative is named or needed, the company being established in Germany. Two gaps remain: no retention period is ever quantified, the policy relying on generic wording, and no minimum age is stated anywhere on the site.
Who owns the data?
No terms of service are published, so ownership of client content is not documented contractually anywhere on the site. The only formal statement is the privacy policy, which names DNL Deep Neuron Lab GmbH, Methfesselstraße 43, 10965 Berlin, as the controller within the meaning of Art. 4(7) GDPR, reachable at contact@dnl.ai. The company is represented by Andreas Schindler, Iason Georgakopoulos and Livia Jansen-Winkeln, and has appointed heyData GmbH in Berlin as its external data protection officer. Audit content stays in the firm's chosen hosting region — the EU, the United States or Australia — but no published clause states who owns it or what the vendor may do with it.
Reuse rights
With no terms of service published, nothing on the site grants or restricts the customer's right to reuse the platform's outputs, and no licence terms are stated. What the vendor does document is the processing of personal data on its marketing website: the legal bases invoked are Art. 6(1)(a) consent, (b) contract, (c) legal obligation and (f) legitimate interest, and transfers outside the EEA rest on adequacy decisions or, failing that, on standard contractual clauses under Art. 46(2)(b) GDPR. The named providers — Vercel, JOIN Solutions AG, HubSpot, Usercentrics, Microsoft Clarity and UET, Google Analytics and Tag Manager, LinkedIn Ads — all relate to that showcase website, not to the audit platform, whose own processors are never listed. The HubSpot contact form loads only after functional cookies are accepted, and the site's robots.txt explicitly allows GPTBot, ClaudeBot and other AI search agents.
Data retention & training
Hosting summary
Client data is hosted in the audit firm's own region — the European Union, the United States or Australia — with the corresponding data-residency controls, according to the vendor's FAQ. Beyond that regional choice nothing is published: no data centre, no cloud provider and no infrastructure subprocessor is named for the audit platform itself. The measures described around it are enterprise-grade encryption, role-based access control, single sign-on, regular penetration testing and continuous monitoring. A distinction matters here. The providers named in the privacy policy — Vercel, HubSpot, Google, Microsoft, LinkedIn, Usercentrics, JOIN Solutions — concern the marketing website only. Vercel Inc., based in Covina, California and certified under the EU-U.S. Data Privacy Framework, hosts that showcase site; it is not presented as the host of the audit platform. Website data transferred outside the EEA relies on adequacy decisions or on standard contractual clauses under Art. 46(2)(b) GDPR. In practice a firm can obtain EU residency for its audit content, but should ask the vendor to name the underlying infrastructure in writing, since the site never does.
Things to keep in mind
Risks and trade-offs to weigh before adopting dnl.
- Over-reliance is the central risk. The vendor states its AI proposals are advisory only and rest on pattern recognition and historical audit data; an auditor who accepts the yellow-highlighted suggestions without testing them hands judgement to a machine while keeping the legal responsibility.
- The 97.4% accuracy claimed on key figure identification is self-declared, taken from the vendor's own AI transparency statement of February 2025 rather than from an independent assessment.
- ISO 27001 is claimed on several first-party pages but no certificate number and no certifying body is given; the information security page is an ISMS policy, not an attestation. Ask for the certificate before treating it as established.
- The published list of processors covers the marketing website (Vercel, HubSpot, Google, Microsoft, LinkedIn, Usercentrics, JOIN Solutions) and not the audit platform. Vercel hosts the showcase site; platform hosting is described only by region — EU, US or Australia — with no data centre or infrastructure provider named.
- No terms and conditions are published, no retention period is quantified, and the vendor never states whether customer data is used to train its models. For audit working papers that is a material gap to close contractually before any rollout.
- The AI transparency statement is frozen at version 1.0 of 12 February 2025, is still written under the former product name Notes Auditor, and still contains an unedited template placeholder pointing to the fictitious yourcompany.com/ai-info — a maintenance signal on the very page meant to document AI governance.
- The client logos on display (KPMG, BDO, RSM, Forvis Mazars, Baker Tilly, HLB, Nexia, KBHT, LTS) are commercial references; only one named testimonial is signed, by Forvis Mazars, and the '10,000+ auditors' figure is a vendor claim that cannot be verified independently.
Setup & Integrations
Technical difficulty
Low for the audit team, but not self-service. The FAQ states that no technical skills are required, and the vendor's implementation and customer success teams accompany the rollout end to end. There is no sign-up: onboarding runs through the vendor, starting with a focused pilot on a single audit cycle with a small team and a defined engagement scope. The configuration work is functional rather than technical, mapping the firm's own checklists and procedures into the platform. The vendor says the workflow is standardised by the third engagement, once the system has learned the firm's patterns.
Deployment
Behind dnl
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Who is dnl built for?
What products does dnl offer?
Which reporting frameworks are covered?
How much does dnl cost?
Is there a free trial?
Does the AI replace the auditor?
Where is client data hosted?
Is there an API?
Is dnl certified?
Can a firm keep its own audit methodology?
Should you pick dnl?
dnl is a narrow tool for a narrow job, and that is its main strength. Disclosure and notes review is a well-defined, heavily regulated task, and the platform is built around it rather than around audit in general: ai/checklist drafts the checklist answers, ai/numbers ties the figures back to evidence, and both leave a full audit trail with every AI-generated element highlighted in yellow. For firms working to IFRS, CSRD, ESRS or the ISSB standards, that alignment with the actual working method is worth more than breadth.
The company behind it is unusually well documented. The legal identity is complete and verifiable — a Berlin GmbH with named managing directors, a commercial register number and a VAT number that checks out — and the funding history is public and dated, from a EUR 1 million seed in 2021 through European public co-funding to a further EUR 2 million in 2022, with a former KPMG EMEA chairman on the advisory board.
What cannot be assessed is the commercial offer. No price appears anywhere on the site, there is no free plan and no announced trial, and no terms and conditions are published at all, so a buyer cannot weigh cost, contract or exit before speaking to sales. The compliance claims deserve the same caution: ISO 27001 is asserted on several pages without a certificate number or a certifying body, the privacy policy quantifies no retention period, and the vendor never says whether customer data feeds its models. The AI transparency statement, still frozen in a February 2025 version under the product's former name and still carrying an unedited template placeholder, does not help.
Worth a demo for an audit practice with a disclosure review problem, and worth a firm list of questions before signing anything.
- Choosing a selection results in a full page refresh.
- Opens in a new window.