aiomics

The platform is the product and there is nothing else underneath it. Multimodal capture across speech, handwriting and documents, structuring, completeness and consistency checking, coding and document generation are all model driven, and the company was founded by a physician specifically to build this rather than adding AI to an existing system. Remove the models and no product remains.

This vendor builds the thing the category is supposed to worry about. Rather than generating a summary and hoping it is complete, the product actively surfaces contradictions, gaps and missing detail in the underlying record, checks completeness and consistency through a human in the loop process, and presents everything for physician review and approval with records reviewed and confirmed on the spot. Interactive checklists close the loop on admission packets, discharge letters and insurer forms so incomplete documents are caught before they leave rather than after they are rejected. That is omission handling as a product feature, and no other vendor reviewed in this lane offers it. Held at B because none of it is quantified: no detection rate for gaps, no false positive rate, no threshold and no abstention behaviour is published, so a buyer cannot tell how much the completeness layer actually catches.

The capability description is more specific than most, naming a quality and safety layer, completeness and consistency checking, structured and coded output and FHIR based handoff, which tells a buyer what the system does even though it does not say how. What is absent is everything at model level: no model or model family named, no accuracy figure for extraction, coding or gap detection, no evaluation methodology published, no model card. The independent evaluation underway at the Charité Institute for Medical Informatics may close much of this gap, and if the results are published with methodology this grade should be revisited immediately.

Nothing is published yet, and the reason this is still worth reading carefully is what has been commissioned. The effectiveness of the product is under INDEPENDENT SCIENTIFIC EVALUATION AT THE CHARITÉ INSTITUTE FOR MEDICAL INFORMATICS, with the company stating results will be published on completion. No other vendor in this category has commissioned an independent academic evaluation of its own product. Until those results exist this remains C: deployment across European rehabilitation and acute care hospitals is stated without naming one, no outcome data is published, and the funding round and investor names are commercial validation rather than clinical evidence. Credit where it is due on the way the claims are framed: the company caveats its own efficiency figures by noting actual time savings depend on individual documentation, which is the opposite of the unqualified percentages that fill this category. THIS RECORD SHOULD BE REFRESHED THE MOMENT THE CHARITÉ RESULTS PUBLISH.

A strong structural position resting on jurisdiction and audit rather than on promises. All processing remains within the European Union, the infrastructure carries a C5 attestation, and the company states that information security and data protection are externally audited. Operating natively under GDPR means purpose limitation, data minimisation and data subject rights are legal obligations rather than product features, which is a materially different baseline from a United States vendor asserting good practice. Held at B rather than higher because no explicit retention period, no statement on whether customer data is used to train or improve models, and no de identification posture was located, and because handwriting and speech capture create a broader ingest surface than document processing alone.

Not Rated because the framework does not apply rather than because the vendor is silent. This is a European product serving European hospitals with all processing inside the EU, so HIPAA is not the governing regime and no business associate agreement exists. The equivalent instrument is a data processing agreement under Article 28 of the GDPR, and the relevant assurances are the GDPR posture, the C5 attestation and the external audit recorded on the safety and security axes. A buyer outside Europe evaluating this product should establish whether a United States deployment is even offered before treating this axis as a gap.

The strongest security credential in this category and it is strong on the specifics this index checks. ISO/IEC 27001:2022 CERTIFICATION, with both the version year and the certifying body named, issued by TÜV NORD. That is a genuine certification against an audited standard rather than an attestation, and naming the certifying body makes it verifiable, which puts it above every SOC 2 claim in this lane including those that specify Type II. Alongside it sits a C5 attestation on the infrastructure, the German federal cloud computing criteria catalogue, and a statement that information security and data protection are externally audited. Held short of a perfect record only by the absence of a trust centre or public certificate registry link, so the artefacts must be requested rather than downloaded.

No device certification is held, and the company engages the applicable frameworks more directly than almost anything else in this category. It references the EU AI Act explicitly and states it is building toward ISO 13485, the medical device quality management standard, and Medical Device Regulation alignment. Naming MDR at all is unusual, since most documentation vendors avoid the question entirely by asserting they are administrative tools. Graded C on the standing principle that building toward is not holding: an intention to align is not a certification, and a buyer should establish the timeline, the intended device class if any, and what the company believes its current regulatory status to be while that work is in progress.

The highest governance grade in this category, and it is earned on behaviour rather than on a policy page. Four things stand out. The company has commissioned an INDEPENDENT ACADEMIC EVALUATION of its own product at the Charité Institute for Medical Informatics and committed to publishing the results, which no competitor here has done. It states plainly that it believes in evidence over claims, for its own tools as much as for a hospital's records. It caveats its own efficiency figures rather than presenting them as universal. And it engages the EU AI Act and medical device frameworks directly instead of asserting itself out of scope. Inclusive design is also stated as an objective, with multilingual and multimodal capability framed as removing communication barriers in workplaces where staff and patients do not share a first language. Held at B rather than A because no fairness, subgroup or demographic performance data has actually been published yet, and because the independent evaluation has not reported. Publication of subgroup results would make this the first A on this axis in the category.

The right shape with the usual naming gap. Output is returned to existing systems as structured, coded and audit secure data with FHIR based handoff, which is genuine bidirectional intent rather than a read only summary the clinician retypes, and ingestion covers speech, handwriting, scanned letters, printed reports and PDFs, which matters disproportionately in European hospitals where paper persists. The company also positions its architecture as adapting granularly to individual hospitals, departments and specialties rather than requiring workflow change. Held at B because no hospital information system or EHR vendor is named anywhere, and the European market has a different and more fragmented set of incumbents than the United States, so an organisation cannot infer support from a customer list.

A real residency commitment, externally audited, which almost nothing else in this category offers. All processing remains in the European Union and the infrastructure carries a C5 attestation, so a buyer knows where clinical content is handled and that an independent party has checked. That is materially stronger than the silence or the unnamed cloud region that characterises most records here. Held at B rather than A because the commitment is to a jurisdiction rather than to customer choice: no on premise, private cloud or customer environment option was located, so an organisation that needs processing inside its own infrastructure has no described path. Ask whether one exists.

No price, tier or pricing mechanism was located. Not Rated is the house convention for absent pricing rather than a low grade. The company does publish an economic impact framing around protecting revenue, increasing occupancy and relieving staff, which describes claimed value rather than cost. Note that reimbursement mechanics differ substantially across European health systems, so any economic case presented should be checked against the buyer's own national funding model rather than accepted as generic.

Focused on a setting nothing else in this index addresses. REHABILITATION alongside acute care is the core market, and rehabilitation carries its own documentation instruments, its own funding applications and its own multidisciplinary team structure, none of which a product built for United States ambulatory care handles. Role coverage extends beyond physicians to nurses, therapists and social services, and the document types named reflect that: admission notes, discharge summaries, therapy briefings, insurer packets, rehabilitation applications and payer correspondence. A second product line addresses practices and telemedicine providers and is described as in pilot. Graded B rather than A because coverage is described at document and role level rather than through demonstrated instrument level behaviour, and because the second product is not yet generally available.