Alcidion

The platform is the product and the algorithms run on top of it. Management's own description is the clearest evidence: Miya Precision standardises data and helps deploy algorithms safely in regulated healthcare settings. That is a statement about being the substrate rather than the intelligence.

Strip out the learned components and a working business remains. Data standardisation across legacy systems, patient flow and bed management, emergency department workflow, care coordination and, at some sites, a full electronic patient record are software and interoperability rather than modelling, and they run across more than 400 hospitals today.

The reason this stays inside the membership boundary rather than falling to the infrastructure filter that rejected Redox, Segmed and Vim is that the company builds its own intelligence as well as hosting others'. Real time clinical decision support is native to the platform, and the generative suite covering clinical documentation, summarisation and problem detection is the company's own capability rather than a third party model passed through.

Graded C on the same reasoning applied to TeleTracking: healthcare only software, neither horizontal nor services, with genuine but non central intelligence. Watch the trajectory. If the product becomes predominantly a deployment substrate for external algorithms, the infrastructure filter would need revisiting rather than the grade adjusting.

Everything here is advisory and the company does not pretend otherwise. Clinical decision support surfaces alerts and recommendations, the generative capability drafts documentation and summarises records, and problem detection flags issues for clinicians. A human acts in every case. No autonomous action on a patient, a bed or a record is claimed anywhere.

That clarity is appropriate for the domain and it is the whole of what is published.

No accuracy, sensitivity, specificity or alert precision figure appears for any decision support component. No confidence expression, no false alert rate, and no description of what a clinician sees when the system is uncertain. In deterioration and problem detection the operative failure mode is alert fatigue, which is well documented in this category and is what determines whether a decision support deployment succeeds or is quietly ignored, and nothing published addresses it.

The generative components raise a second question with no published answer: whether drafted documentation is marked as machine generated in the record, and whether a clinician must affirmatively accept it. Ask for alert precision in production and for the attribution behaviour of generated text.

Capability areas are named and mechanism is not. The company describes a growing suite of generative artificial intelligence tools within the platform covering clinical documentation, summarisation and problem detection, stated as reducing clinician administrative burden while improving decision making and data quality. Naming three distinct applications is more specific than a general claim to be artificial intelligence enabled.

Underneath there is nothing. No model class, architecture, foundation model, version or training data is described for either the decision support layer or the generative layer. No accuracy figure of any kind is published for any component.

The platform layer is better described than the model layer. Data standardisation across heterogeneous legacy systems is the stated technical foundation and is credibly evidenced by deployment across three national health systems with entirely different underlying estates, which is a harder engineering problem than it sounds and is the company's real differentiator.

Graded C for named applications with no technical specification behind them. A published evaluation of the problem detection capability would move this more than any amount of further product description.

No party in the chain is named. No foundation model provider, model class or version, no cloud platform, no sub processor list, and no statement on whether customer data contributes to model development.

The omission is sharper here than for most records because of what the company says it does. Describing the platform as one that helps deploy algorithms safely in regulated healthcare settings means third party models are expected to run on it, which makes the identity and governance of those models a first order question rather than a technical footnote. Nothing published says whose algorithms, under what assessment, with what accountability when one performs badly.

The generative suite raises the same question from the other direction. Generative capability for clinical documentation and summarisation is very unlikely to be built from scratch by a company of this revenue scale, which implies an external model provider that is not identified.

Jurisdiction compounds it. Clinical data from National Health Service trusts, Australian state health services and New Zealand regions moving through an unnamed model provider is a procurement question in all three systems. Ask which foundation models are used, which third party algorithms are hosted, and for the sub processor register by jurisdiction.

The deployment record is the strongest in this lane and the outcome measurement is absent, which is the familiar split.

Scale is documented rather than asserted, because a listed company has to substantiate what it says. More than 400 hospitals, more than 50,000 beds and 130,000 active users across three countries. More than 40 National Health Service trusts, providers in every Australian state, and all four regions of New Zealand Health. Named customers include University Hospitals Sussex, Harrogate, Northumbria, Dartford and Gravesham, Hywel Dda and Western Health.

Retention is the most persuasive evidence on the record and is rarely available elsewhere in this index. Western Health renewed for a fifth time across twenty years, several trusts renewed multi year, and a customer adopting an additional module during renewal discussions is disclosed. Twenty years of repeat purchase is a harder signal than any satisfaction score.

What is missing is measurement of the intelligence. No published clinical outcome, no evaluation of the decision support components, no peer reviewed study located, and no third party research assessment. For a platform whose decision support touches deterioration and problem detection, the absence of published clinical evaluation is the gap worth pressing. Ask for outcome data from any long standing site.

One relevant statement exists and no detail sits behind it. Management describes the platform as helping deploy algorithms safely in regulated healthcare settings, which is a claim about governed model deployment and is the closest thing to a safety position on the record.

Nothing operational accompanies it. No retention schedule, encryption statement, access control description, data ownership or deletion position, data minimisation commitment, or statement on whether customer data contributes to model development or to the generative capability.

The stewardship question that matters most here is jurisdictional segregation. The company holds clinical data from three national health systems whose laws differ on where data may sit and who may access it, and it recently acquired a customer base from another vendor, which involves data migration. Whether United Kingdom, Australian and New Zealand data are segregated, and what happened to the acquired customers' data on transfer, are the questions a buyer in any of those systems would ask first, and neither is addressed.

Graded C conservatively for the same reason as the privacy axis: national health system supply implies controls exist and none is published.

This axis assumes a United States vendor and does not fit this record, so it is graded against the equivalent obligation rather than the literal one. The reasoning is recorded here so the grade is not misread.

The federal privacy framework this axis names is largely inapplicable to a business that is 63 percent United Kingdom and 37 percent Australia and New Zealand by revenue. The obligations that actually govern it are the United Kingdom data protection regime, the Australian Privacy Act and its state health records legislation, and the New Zealand Privacy Act with its health information code. Those are the frameworks a buyer in its markets would ask about.

Nothing published addresses any of them. No privacy posture statement, control enumeration, de identification position, data processing agreement template or equivalent of a business associate arrangement was located for any jurisdiction.

Graded C conservatively rather than punitively, on the basis applied to Optum and TeleTracking. Supplying more than 40 National Health Service trusts requires satisfying that system's data security requirements at every one of them, so a posture demonstrably exists in contract form. None of it is published where a prospective buyer would look. Ask for the data protection position by jurisdiction and how the three regimes are kept separate.

No credential was located. A targeted pass found no controls report, no information security certification, no health specific security framework certification and no cloud authorisation. There is no trust center, no security page, no report availability process, no penetration testing disclosure and no vulnerability disclosure policy in the material retrieved.

The absence is almost certainly publication rather than substance, and more confidently so here than for most records. Supplying more than 40 National Health Service trusts requires meeting that system's mandatory data security standards, and Australian and New Zealand public health procurement carries comparable requirements. Certifications of some form therefore exist as a condition of the business the company has already won.

That does not change the grade. This index measures what a buyer can verify before contacting sales, and a vendor whose entire customer base is public health systems with published security requirements has both the artefacts and an unusually strong reason to publish them.

Graded D on the same basis as TeleTracking: a focused company with a single corporate property where a targeted search returning nothing is meaningful evidence of non publication. Ask for the security certifications held in each market and whether reports are available on request.

The regulator this axis names has no jurisdiction over a business operating in the United Kingdom, Australia and New Zealand, so the grade is reasoned against the equivalent regimes and the reasoning is recorded here.

The question is not moot, which distinguishes this from every other record in this lane. Patient flow and bed management are administrative and unregulated as devices. Clinical decision support is not automatically so. Software that detects clinical deterioration or flags a clinical problem can meet the medical device definition under United Kingdom medical device regulation, and equivalent thresholds apply under the Australian therapeutic goods framework, which expressly captures certain clinical decision support software. The exemptions generally turn on whether a clinician can independently review the basis of the recommendation, which is precisely the kind of design detail this record could not establish.

Nothing published addresses device status in any jurisdiction: no marking, no registration, no inclusion in the Australian register, no regulator correspondence and no statement that the products fall outside the definitions.

Graded C because the position may well be correct and is entirely undocumented. This is the first record in the index where the device question is live rather than formal. Ask which components are regulated as devices in which markets, and on what basis the others are exempt.

A governance posture is asserted at the right level of abstraction and nothing beneath it is published. Management's framing of the platform as standardising data and enabling algorithms to be deployed safely in regulated environments is a governance claim, and it is a more thoughtful one than most vendors offer, because it locates the risk in deployment discipline rather than in model quality alone.

No content follows. No bias or fairness testing, no model validation methodology, no monitoring output, no distribution reporting, no drift detection and no external audit was located for the decision support or generative components.

Two governance questions are specific to this record. If the platform hosts third party algorithms as management describes, the vendor is making assurance decisions on behalf of customers about models it did not build, and nothing published describes the assessment applied before an algorithm is allowed to run. And deterioration and problem detection models trained on data from one national health system may perform differently in another, given different populations, coding practices and documentation norms, with nothing published on cross jurisdiction validation. Ask what assurance a hosted algorithm must pass, and whether models are revalidated per country.

No performance figure is published for any intelligent component, so there is no stated level against which a shortfall could be measured and nothing to hold the vendor to. No accuracy, sensitivity, specificity, alert precision or error rate appears for decision support, problem detection or the generative capability.

No service level agreement, warranty, indemnity or remediation commitment was located, and no pilot, benchmark or validation offer of the kind several competitors publish was found.

The contract disclosure creates a specific and answerable question that the absence of performance terms makes pointed. Individual agreements are published with durations of seven to ten years and values in the tens of millions. Multi year contracts of that length in a domain where the underlying models will be replaced several times over the term ought to carry performance and change provisions, and nothing published indicates what they are.

One pre emptive note for future passes: further financial or scale disclosure cannot move this grade. Revenue, recurring revenue, contract value, bed counts and user numbers measure commercial performance and adoption, not the accuracy of the intelligence or the buyer's recourse when it is wrong. Only a published evaluation of a decision support component, or a contractual commitment on performance, will change it.

Interoperability is not a feature of this product, it is the product thesis, and deployment scale evidences it more convincingly than a vendor list would.

The company positions itself as modernising legacy health systems by standardising data across heterogeneous estates, and it does so simultaneously across three national health systems with entirely different underlying technology. Patient administration system customers are referenced as a distinct base, several of whom renewed multi year, which indicates the platform sits above and integrates with incumbent systems rather than replacing them. It also operates through a systems integrator partnership in Australia and New Zealand.

At the other end of the range it can be the record system itself. The May 2026 University Hospitals Sussex agreement is a full electronic patient record win rather than a modular sale, which is a materially different capability claim from integrating with someone else's record.

Held at B rather than A because no interface standard is named anywhere in the material retrieved, no specific record system is identified as supported, and no certification or marketplace listing was located. The capability is demonstrated by deployment and undocumented in specification. Ask which standards the data standardisation layer consumes and which record systems are supported in each market.

Nothing was located, and this is a record where the question is unusually consequential.

No hosting model, cloud provider, region, tenancy model, residency commitment or customer controlled option appears in the material retrieved.

The markets served have stricter and more explicit residency expectations than the United States norm this index usually encounters. National Health Service data handling carries specific requirements on where patient data may be processed and on offshore access. Australian government health procurement applies data sovereignty conditions, with several state health services requiring onshore hosting. New Zealand applies its own health information rules. A single platform serving all three either operates separate regional instances or moves data across borders, and which of those is true is the first question any of those buyers would ask.

The recent acquisition of another vendor's product line and customer base adds a migration dimension with the same question attached.

Graded C consistent with how silence has been treated across this lane, with the note that the bar here should be higher than for a domestic vendor. Ask for hosting regions by country, tenancy model, and whether any support access occurs from outside the country of deployment.

The most financially transparent record in this index, for a structural reason rather than a philosophical one: public listing compels disclosure that private competitors withhold.

A buyer can see the shape of the business. Revenue of 51.6 million dollars for the year to 30 June 2026, up 27 percent, split 63 percent United Kingdom and 37 percent Australia and New Zealand. Annual recurring revenue of 38.3 million dollars, up 34 percent. A record 78.5 million dollars of new and renewal total contract value. Cash of 15.1 million dollars with no debt. Contracted revenue described as extending into the mid 2030s.

Individual contract economics are published, which no private vendor in this index offers. The University Hospitals Sussex electronic patient record agreement is stated at roughly 35 million dollars over seven years, extendable to roughly 49 million over ten. The acquired Kyra product line was disclosed at roughly 3.7 million dollars of revenue, more than 90 percent recurring, with underlying earnings of roughly 1.1 million. A buyer can derive an order of magnitude for a comparable deployment from those figures.

Held at B rather than A because none of it is a price. There is no rate card, no unit of charge, no published basis for scaling cost with beds or modules, and statutory disclosure is not the same as commercial transparency toward a buyer. Ask for the licensing basis and how modules price into an existing deployment.

The broadest jurisdictional coverage of any record in this index, and the only one operating at national scale outside the United States.

Three national health systems are served concurrently: more than 40 National Health Service trusts in the United Kingdom, healthcare providers in every Australian state, and all four regions of New Zealand Health. Each has a different funding model, different clinical coding and data standards, different privacy law and a different procurement regime, and the platform runs in all three. Stated expansion interest covers Canada and the Middle East.

Functional coverage is correspondingly wide rather than deep in one workflow: patient flow and bed management, command centre operations, emergency department workflow, care coordination, virtual care, clinical decision support and full electronic patient record deployment. The Sussex contract is a full record deployment rather than a modular sale, which demonstrates the platform can be the system of record rather than only a layer above one.

Graded A because the coverage claim is evidenced by deployment at national scale in three countries rather than asserted, and because operating across that many regimes is a substantive capability rather than a marketing footprint. Every other record in this index carries a note that nothing outside the United States is addressed. This is the exception.