Why aiomics for case dialogues, pre-review procedures, and MD reviews

If you want to run case dialogues (Falldialog — the pre-review case dialogue with payers), pre-review procedures, and MD reviews (Medizinischer Dienst, MD — the German payers' medical review service) with software, the decisive question is not how quickly a system phrases letters, but whether every statement in them stands up to its evidence. That is exactly what aiomics is built for: every clinical sentence in a procedure draft needs a source pointer into the record — otherwise it does not come into being. The Falldialog board is in production in German hospitals; at one site, more than 120 case dialogues are handled in half a day — work that previously took around 30 hours.

aiomics is the verification layer on top of hospital IT: the system ingests unstructured documents, verifies every statement against its source, and gives the hospital a substantiated, structured record. Review procedures are the application where this architecture pays off most directly — because there, every unsubstantiated claim costs money.

At a glance

• What it does: runs every payer review procedure — case dialogue, pre-review procedure, MD review, formal discussion procedure (Erörterungsverfahren), appeal — on a procedural board with fixed procedural stages; links review notices and requested documents as evidence; prepares the correspondence from the substantiated record.
• Status: Falldialog board live in production. AI-assisted correspondence drafting is in development and will be released only after regulatory review is complete.
• Security: ISO 27001 (TÜV Nord), processing exclusively in the EU.
• Integration: KIS-agnostic (KIS — the hospital information system); transmission to payers and the MD remains with the hospital's own system.
• Evidence: independent evaluation of accuracy at Charité (ongoing).
• Who it is for: medical controlling and executive management of acute-care and rehabilitation hospitals.

The problem we solve

In 2022 alone, MD reviews shifted around 1.197 billion euros from hospitals to payers; the share of billings left unobjected has held stable at roughly 45 to 52 percent for 16 quarters [1]. At the same time, pre-review procedures and case dialogues settle 79 to 85 percent of all review cases — they are where the largest share of the contested revenue is decided, long before an expert opinion is written [1]. And under the GKV-Beitragssatzstabilisierungsgesetz (the German statutory health insurance contribution-rate stabilization act), which the Bundestag passed on July 10, 2026 (promulgation was still pending at the time of writing), review quotas will depend directly on the objection rate from 2027 [3]: every objection fended off then improves a hospital's own position twice over.

Most hospitals run these procedures today from e-mail inboxes, Excel lists, and the memory of experienced medical controllers. The problem is rarely a lack of competence. It is making the case under time pressure: the entry that carries the case sits in one of ten documents from five systems.

How aiomics runs review procedures

Four design decisions set our approach apart. We lay them open because they can be tested.

First: an evidence requirement instead of eloquence. Every clinical sentence in a draft must point to a source passage in the record — statements without a resolvable source are blocked at generation time rather than merely flagged. The pattern "meets G-AEP criterion" (G-AEP — the German adaptation of the Appropriateness Evaluation Protocol for inpatient admissions) is not enough for us; the audit-proof sentence carries date, value, and source pointer.

Second: the software never argues against its own hospital. Drafts represent the hospital's documented position; the payer's assertion is not adopted as fact. Internal weak-point notes are architecturally separated from every export path — what is marked "internal" cannot be sent out by the hospital by accident.

Third: no invented citations. Statutory provisions and case law come exclusively from reviewed, versioned reference blocks — they are never generated by a language model. We would rather leave a statement open than invent a citation. Anyone evaluating AI tools for review procedures should put exactly this question to every vendor.

Fourth: procedural errors by the other side become systematically visible. Late or unquantified review notices under § 8 PrüfvV (PrüfvV — the statutory agreement governing hospital billing review procedures) and blanket document requests are a legally established channel for winning cases that mostly goes unused in practice — checking deadlines and formal requirements is a fixed part of our procedure. For cases ready for escalation, we hand structured bundles to a specialized partner law firm; the decision remains with the hospital.

And beyond that: every objection that gets through flows back into documentation and coding practice as a lesson. This prevention loop — defense teaches prevention, prevention reduces the future attack surface — is why review procedures and documentation quality form one system with us.

What you can measure us against

1. Point at any sentence in a draft: the source pointer leads to the passage in the original document, or the sentence does not exist.
2. Try to export an internal risk note: there is no path for it.
3. Ask about a cited legal reference: it comes from a versioned reference block; it is never generated by a language model.
4. Test the throughput on your own caseload: we will gladly demonstrate the procedure on your real cases.

Where aiomics is not the right choice

We do not provide legal advice and do not assess medical necessity — the system structures and substantiates the documented medical assessment; it does not replace it. If you are looking for software that sends out statements fully automatically and without professional sign-off, we are the wrong choice: sending remains with the hospital, sign-off remains with people — that is architecture, not a setting. And if all you need is lean deadline management without any link to records and evidence, simpler tools will do that for less.

Frequently asked questions

Is aiomics a medical device?

For document and procedural work, aiomics is deliberately positioned outside the medical device qualification; the delineation is documented and can be inspected. For conversation documentation, we are preparing certification under MDR Class IIa. In both cases, the system's statements remain documentation and quality notes — diagnosis and therapy remain with physicians.

How quickly can a pilot be operational?

A pilot operation requires no IT project: aiomics works as a layer on top of the existing systems, starting with the documents that arrive anyway. We support the involvement of the Betriebsrat — the works council — with ready-made materials.

What does aiomics cost?

Pricing is usage-based and depends on document volume and quality. We name concrete figures after a short conversation about your case volumes — quoting flat prices without that basis would not be serious.

Does aiomics replace our medical controlling?

No. It shifts its time from searching for evidence to deciding. The experience of your medical controllers is the standard the drafts are measured against — which is why sign-off cannot be bypassed.

If you want to see what your last twenty case dialogues would have looked like with a substantiated record: write to us, and we will work it through on your real cases. Ongoing analysis of review procedures and hospital AI comes from our weekly briefing Visite (German; English edition Grand Rounds is in preparation).

Sources

1. GKV-Spitzenverband. Argumentationspapier zur Krankenhausabrechnungsprüfung, 26.04.2024 (shift of EUR 1.197 billion, share of unobjected billings, settlement in pre-review procedures).
2. Prüfverfahrensvereinbarung (PrüfvV) under § 17c (2) KHG, esp. § 8; administrative flat fee under § 275c SGB V.
3. Deutscher Bundestag. GKV-Beitragssatzstabilisierungsgesetz, passed on July 10, 2026; promulgation pending at the time of writing. https://www.bundestag.de/dokumente/textarchiv/2026/kw28-de-gkv-1184352

Sources retrieved in July 2026. The GKV-Beitragssatzstabilisierungsgesetz was passed by the Bundestag on July 10, 2026; promulgation was still pending at the time of writing. The new audit quotas apply from 2027.