LiveData

The lowest artificial intelligence centrality of any record built in this sweep, and the reasoning is worth stating precisely because the grade is a judgement about the product rather than the company.

Thirty five years of perioperative workflow software sits underneath: scheduling, patient flow, display boards, safety checklists, mobile alerting and analytics. All of it functions exactly as before without any model. The artificial intelligence is a single capability launched on 21 May 2026, a conversational layer allowing a leader to ask a plain language question about block utilisation or turnover and receive an answer from data the analytics product already held.

That is natural language access to existing metrics rather than analysis that did not previously exist. It retrieves and phrases; it does not predict, detect or decide. Remove it and no customer loses a capability, only a convenience.

Graded D rather than the C given to TeleTracking, and the distinction is deliberate. That company's artificial intelligence product forecasts throughput, producing an analytical output the platform could not otherwise generate. This one surfaces numbers already computed. The appointment of a chief product officer to accelerate artificial intelligence work in 2026 indicates the company knows it is at the start of this rather than partway through.

There is essentially no autonomy to bound, and the company has addressed the one risk that matters for the capability it does ship.

The advisor answers questions a human asks and the human reads the answer. Nothing acts, schedules, alerts or writes. Every other module in the suite is a display, a scheduling tool or a checklist operated by staff, and the case milestone capture is explicitly manual, recorded by a clinician pressing a handheld device.

The relevant failure mode for a conversational analytics tool is a confidently stated wrong number, and the company speaks to it directly: answers are stated to be grounded in the health system's own data rather than generic benchmarks. That is a grounding claim addressing hallucination rather than ignoring it, and it is more than several better funded competitors offer.

What is absent is any evidence the grounding works. No accuracy or faithfulness measurement for the generated answers, no description of what the advisor does when a question cannot be answered from available data, and no statement of whether answers cite the underlying query or figures. Ask how often the advisor is wrong and what a user sees when it is uncertain.

The architecture around the model is described more usefully than the model itself, which is an unusual balance and works in the buyer's favour on the questions that determine deployment.

Three specifics are published and each is checkable. The advisor works with any electronic record rather than a supported list. It requires no proprietary data pipeline, meaning no separate extract or warehouse build before value appears. And it is delivered inside the existing analytics cloud with role based access already established, so it inherits controls rather than introducing a new environment. For a hospital evaluating effort and risk, those three answer more than a model card would.

The query domains are enumerated rather than gestured at: block utilisation, case volume, late starts and turnover time.

Below that there is nothing. No model class, foundation model, version or architecture is named, no description of how questions are translated into queries, and no accuracy or faithfulness measurement for the answers. The wider suite's data handling is better described, blending record, anaesthesia and physiological monitoring sources into a single case view. Graded C for strong deployment description with no model specification.

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

The omission is straightforward rather than complicated here. A conversational capability launched in 2026 by a company of roughly seven million dollars in revenue is almost certainly built on an external large language model rather than trained in house, which means a third party processes questions about a hospital's operational performance and, depending on the architecture, the underlying data those questions touch. Which provider, under what terms, and whether any query content leaves the analytics environment are the questions, and none is addressed.

The grounding claim makes the architecture question more rather than less pointed, because grounding answers in a customer's own data implies that data reaches the model at inference time. Whether that happens inside the customer's tenancy or at a provider's endpoint is exactly what a security reviewer would ask.

Ask which model provider is used, whether operational data or query text leaves the analytics cloud, and for the sub processor register.

One credential of a kind not encountered elsewhere in this sweep, and very little else.

OR-Dashboard with Active Time Out has been designated a Leading Practice by the hospital accreditation body. That is a patient safety recognition from the organisation that inspects hospitals, awarded for a surgical safety checklist implementation, and it is categorically different from the analyst rankings and buyer satisfaction scores that constitute third party validation for most records here. It concerns whether the product helps prevent wrong site surgery, not whether customers enjoy using it. Thirty five years of continuous operation is itself a survival signal, and the company was named among 25 Forbes Small Giants in 2019.

Against that, the evidence base is thin. No customer is named anywhere in the material located. No outcome figure, utilisation improvement or turnover reduction is published. No research organisation assessment or peer reviewed study was found. Revenue is estimated by a third party at roughly seven million dollars, which is consistent with a small installed base.

The artificial intelligence capability has no evidence at all, which is unsurprising three months after launch. Graded C on the accreditation credential and tenure. Ask for named reference sites and utilisation outcomes.

Two controls are named and the stewardship questions are unaddressed.

Role based access is stated for the analytics cloud, and the public display product is described as designed for privacy compliance. Both are specific rather than generic.

Nothing else is published. No retention schedule, encryption statement, data ownership or deletion position, data minimisation commitment, or statement on whether customer data contributes to model development.

The data footprint is narrower than several records built this session, which works in the vendor's favour and is worth stating for balance. This product observes operational events, schedules and case timings rather than clinical narrative, and it captures no audio or video. Compared with a vendor recording surgical procedures or ambient room conversation, the sensitivity of what is held is materially lower, even though it does ingest anaesthesia and physiological monitoring data during the case.

The new advisor introduces the one open question that matters: whether operational data reaches an external model at inference time. Ask what the advisor sends outside the analytics environment, and what is retained of the questions asked.

No formal posture is published, and one product specific compliance statement is more meaningful than the usual badge.

The Family Waiting Board is described as a compliant display for public areas, showing families the progress of a patient through surgery. That is a genuinely difficult privacy design problem rather than a boilerplate claim: a screen in a public waiting room showing patient status is exactly the scenario health privacy rules constrain, and the company addressing it explicitly indicates the constraint was designed for rather than discovered later. The analytics environment is separately stated to carry role based access.

Beyond those two statements there is nothing. No control enumeration, no de identification position, no retention statement, and no business associate agreement posture, template or execution requirement was located.

Graded C conservatively on the basis applied across this lane. A vendor operating for thirty five years inside operating rooms, integrating anaesthesia and physiological monitoring data, has satisfied privacy review at every customer and executes agreements as routine. None of it is published. Ask for the agreement position and what governs the public display implementation in practice.

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 strongest statement located is that the analytics cloud carries enterprise security and role based access, which names one control and one adjective.

The conclusion is more measured here than for the larger vendors graded D on this axis. A company of roughly seven million dollars in revenue faces a genuine cost barrier to formal certification, which routinely runs into six figures, and thirty five years of hospital deployment without a published credential suggests customers have accepted assessment through their own vendor risk processes instead. That is an explanation rather than a defence, and it does not change what a buyer can verify before contacting sales.

Graded D consistently with the other records in this lane. Note that the accreditation body designation this company holds is a patient safety credential, not a security one, and should not be read across. Ask what security assessments exist and whether any report can be shared.

No device pathway is claimed and none appears necessary. Scheduling, display boards, workflow coordination and analytics are administrative, and the advisor answers questions about operational metrics rather than about patients.

One component sits in a more regulated neighbourhood than the rest and is handled well. Active Time Out supports the surgical safety pause, the mandated verification of correct patient, procedure and site immediately before incision, which exists to prevent wrong site surgery. Software supporting a mandated safety procedure carries quality implications beyond convenience, and the company's position here is stronger than a regulatory claim would be: the accreditation body that mandates the procedure has designated this implementation a Leading Practice.

OR-Dashboard also ingests physiological monitoring data during the case, which is device output. Displaying it for situational awareness rather than analysing it for clinical determination keeps the product clear of device definitions, and nothing published states that position explicitly.

Graded C because the classification is correct and undocumented, with the safety designation as the mitigating evidence. Ask for the stated device status determination and the basis for it.

Nothing is published. No bias or fairness testing, no model validation methodology, no monitoring output, no accuracy measurement and no external audit was located for the advisor.

The grounding commitment is the one adjacent statement and it is a design claim rather than a governance one. Stating that answers come from the health system's own data rather than generic benchmarks addresses where information originates; it does not address whether the model translates a question into the right query, aggregates correctly, or handles an ambiguous question by guessing.

The governance risk in a conversational analytics tool is specific and easy to underrate. An executive asking about block utilisation and receiving a confident, well phrased, wrong number will act on it, and unlike a dashboard there is no visible query to inspect. Decisions about block allocation and staffing follow from those numbers and affect which surgeons and which service lines get theatre time.

Nothing published describes whether answers show their working, whether users can see the underlying query, or how errors are detected. Graded D because a capability three months old ships with no evaluation disclosed. Ask whether answers are auditable back to the query.

No performance figure is published for the advisor, so there is no stated level against which a shortfall could be measured. No accuracy or faithfulness measurement, no confidence expression and no error rate appears for the generated answers.

No service level agreement, warranty, indemnity or remediation commitment was located for any product, and no pilot or trial terms were found.

The recourse exposure is modest by the standards of this index and should be described accurately rather than inflated. The advisor answers management questions rather than clinical ones, and a wrong figure produces a poor block allocation decision rather than a patient safety event. The wider suite carries no autonomous action at all, since milestones are recorded by staff and displays show what the source systems report.

The one place where the stakes rise is the safety checklist. Active Time Out supports the mandated verification before incision, and any failure there is a patient safety matter, though the failure mode would be availability rather than model error. Nothing published addresses uptime commitments for a component used in every case.

Only a published accuracy measurement for the advisor, or an availability commitment for the theatre facing modules, will move this grade.

The strongest axis on this record, and the disclosure is specific in ways that matter operationally.

Three claims stand out. The advisor is stated to work with any electronic record rather than a supported list, and to require no proprietary data pipeline, which together describe a low integration burden a buyer can test. The suite shares scheduling data bidirectionally with the record system, positioning itself as an improved version of the record vendor's own surgical calendar rather than a parallel system. And OR-Dashboard blends three distinct source types into one live view: the record, anaesthesia systems and physiological monitoring devices. Integrating device output alongside record data is a materially harder class of work than record integration alone and few vendors in this index attempt it.

Ecosystem presence is evidenced rather than asserted. The company markets a solution specifically for customers of one major record vendor and exhibits at that vendor's user conference, which identifies at least one platform concretely.

Held at B rather than A because no interface standard is named, no certification or marketplace listing was located, and any electronic record is a claim rather than a list. Ask which standards the integrations use and which record systems are in production.

The delivery model is partially described and location is not.

What is published is a hybrid shape that follows from the products. LiveData Insights is a cloud environment and the advisor runs inside it, inheriting its access controls. The workflow modules are different in kind: display boards in perioperative corridors and waiting areas, handheld devices in theatres, and real time integration with anaesthesia and monitoring equipment all imply on premise components tied to hospital infrastructure. A buyer therefore knows roughly which parts sit where, which is more than most records here disclose.

What is absent is every specific. No cloud provider, region, tenancy model, residency commitment or customer controlled option was located, and nothing states whether case data leaves the hospital for the analytics environment or is aggregated locally first.

The advisor sharpens the question, since a conversational layer grounded in customer data implies that data is queryable from the cloud environment. Ask where the analytics environment runs, what leaves the hospital network, and whether the on premise components can operate if the cloud connection drops during a case.

Nothing about cost is published. A dedicated pass located no pricing page, no unit of charge, no range, no implementation or onboarding fee position, no minimum commitment, no pilot terms, no return calculator and no percentage saving or utilisation improvement figure.

There is not even a directional efficiency claim of the sort that lifts several records in this index above D. The product marketing describes capabilities and benefits qualitatively throughout, with no number attached to any of them.

The modular structure makes the omission more consequential than for a single product vendor. The suite comprises at least nine named modules plus an analytics cloud and now the advisor, and nothing indicates whether they are licensed individually, as a suite, or in tiers, which is the first thing a hospital buying only scheduling or only display boards would need to know. Whether the artificial intelligence advisor is included for existing analytics customers or carries incremental cost is also unstated, and matters because it launched into an installed base.

Ask for the licensing basis, whether modules are separately priced, what a single module deployment costs, and whether the advisor is an upgrade or a separate line.

Narrow, coherent and undocumented internally.

The setting is perioperative: hospital operating rooms and surgery centres, extended by Procedure Suite Manager into procedural areas outside the main theatre complex, which is a sensible adjacency since procedure suites face the same scheduling and turnover problems with less software attention. Coverage runs the full perioperative journey from clinic scheduling through check in, the case itself, and family communication in the waiting area, which is broader within the setting than most competitors attempt.

What is not published is any breakdown inside it. No surgical specialty enumeration, no statement of facility size range, no distinction between academic and community deployment, and no named customer from which a reader could infer any of it. Nothing addresses markets outside the United States.

The family waiting board is worth noting as genuine coverage of a constituency almost nothing else in this index addresses. It is a compliant public display keeping families informed of a patient's progress through surgery, which is a real operational problem and an unusual thing for a vendor to build. Graded C for a coherent narrow setting with no internal detail. Ask for facility size range and specialty coverage.