It could be designed to help, but Work-specific pre-consultation preparation is not yet established. The strongest supported hypothesis is assistance with selecting relevant patient context and outstanding work before an appointment. Whether that uses previous Heidi sessions, the wider medical record or additional sources remains an important unanswered question.
This is pre-launch analysis dated 27 September 2026. Heidi Work is the anticipated name for the forthcoming product, not a confirmed specification. The Melbourne event material places chart preparation among the problems surrounding consultations. Importantly for GPs, the current waitlist excludes the UK and Europe; this is not a guide to a Work feature already available to a UK practice.
What would a genuinely useful brief contain?
The purpose of preparation is not to summarise everything known about a patient. It is to identify the information that helps the clinician begin the particular appointment, while making important gaps visible. That is an editorial and workflow distinction, not a claim that an AI system can reliably infer relevance without evaluation.
An ideal brief might identify the recorded reason for attendance, relevant recent events, outstanding actions and questions that remain unresolved. Each factual statement should connect to its source and date. Where the booking reason is vague, the brief should preserve that uncertainty rather than manufacture a confident narrative.
A useful design would distinguish patient-reported information, a previous clinician's assessment and an administrative status. Those are different kinds of context. A referral requested in an earlier note is not necessarily a referral submitted, and a document received is not necessarily a document reviewed.
The desired result is a short orientation with a route back to the record, not a replacement record that asks the clinician to trust an unexplained selection.
Preparation already exists in Heidi's broader positioning
Heidi's primary-care page, checked on 27 September 2026, describes adding patient information and reviewing patient lists and key context before interactions. This supports a broader preparation proposition. It does not establish that the anticipated Work product autonomously reads every source or prepares every appointment without configuration.
The distinction matters for existing users. A new interface might make preparation easier without introducing a wholly new activity. A more substantial expansion would need to show what additional information becomes accessible, how relevant context is selected and what manual work is removed.
The event's emphasis on preparing charts makes this a reasonable launch question. It does not settle the answer. A demonstration should identify the starting information rather than allow viewers to assume a complete longitudinal record was available.
A fictional morning clinic shows why one summary format is insufficient
Imagine a family physician preparing for a mixed morning clinic. The first appointment concerns a planned review after a previous discussion. Useful preparation would prioritise what was agreed, which actions were expected before today and whether the relevant information has arrived. Repeating the entire historical problem list could obscure that purpose.
The next appointment follows a recent encounter elsewhere. Here, the clinician may need to know which external documents are available, what remains missing and whether the reason for today's visit relates to that event. The brief should not treat the absence of a document in Heidi as evidence that no external encounter occurred.
A third appointment is for a patient the clinician has not previously met. The booking description is nonspecific. A useful brief could identify the limits of preparation and show essential source context without pretending to know the patient's agenda. The opening conversation remains necessary.
A fourth appointment concerns an administrative request. The relevant context might be an earlier agreed action and its status, rather than a broad clinical narrative. A preparation tool should not turn this into a treatment recommendation merely because it can generate one.
These examples are fictional and non-prescriptive. They illustrate why evaluation should include different appointment types, not only a carefully selected case with a clear history and a predictable purpose.
Previous sessions and the wider record are different starting points
A system using earlier Heidi sessions could offer useful continuity within that information. It might still lack events recorded elsewhere, recent correspondence or actions completed through another route. The correct output should reflect that scope.
Access to a wider clinical record would change the task, but it would not remove the need for source selection and permissions. The product would need to show which information it read, how recently it was refreshed and what it could not access. A connection to a clinical system is not evidence that every relevant data type is included.
The distinction becomes especially important when preparation happens before the working day. A brief assembled earlier can become outdated before the appointment. A useful design would make the preparation time and any subsequent refresh visible, with a way to inspect new information that arrived in between.
None of these data-access arrangements is confirmed for Work in the sources reviewed on 27 September 2026. They are questions that would materially alter the value of a pre-consultation feature.
A shorter record can still be a worse brief
Compression can remove information indiscriminately. Selection should depend on the intended task. A brief that is concise but misses the unresolved reason for follow-up may save reading time while creating another search later.
Conflicting entries also need deliberate handling. If an older note and newer correspondence describe different plans, a summary should not silently blend them into a single invented agreement. The clinician needs to see the disagreement, the source dates and which instruction appears current, without the interface pretending that chronology alone resolves every conflict.
A useful hypothetical design would separate established facts, outstanding work and unresolved context. It could then offer deeper detail only when requested. That is a design proposal, not evidence that Work already uses those categories or that every clinician would prefer the same layout.
How a practice could evaluate preparation
The comparison should begin with the existing preparation process. Record the time spent finding relevant information, the sources opened and the important facts needed for the appointment. Use an appropriate authorised test environment or carefully designed synthetic cases before considering a live clinical pilot.
The same appointment set can then be reviewed using the proposed tool. Evaluate preparation time, important omissions, irrelevant material, unsupported statements and repeat searching. Record whether the clinician could locate the source quickly when a claim needed checking.
| Evaluation question | Why it matters | What a useful record would contain |
|---|---|---|
| Did the brief support this appointment's purpose? | Relevance differs by visit type | The stated appointment task and required context |
| Was important information omitted? | Brevity alone is not sufficient | The missing item and its source |
| Did the clinician repeat the search? | A short brief may shift rather than remove work | The additional search and its reason |
| Was the information current? | Preparation can precede later record changes | Brief generation time and source refresh information |
| Were uncertainties visible? | Missing context should not become false certainty | How the gap or conflict was represented |
This is a proposed evaluation worksheet, not a published Work performance result. A small successful demonstration would support feasibility for that scenario, not general clinical reliability.
Patient information and clinical reference answer different questions
A pre-consultation brief answers what is known about this encounter. A clinical-reference search asks what relevant guidance or evidence says. Combining the two without showing the boundary can make a generated suggestion look like an existing patient-specific plan.
Per iatroX product information, September 2026, Ask-iatroX is free and uses linked NICE, CKS, SIGN and SmPC information from emc. Its published methodology describes retrieval and checking processes, not automatic access to a practice's patient records or integration with Heidi.
A clinician might therefore use preparation software for encounter context and a separate source-linked reference tool for a focused knowledge question. That is a complementary workflow, not a claim that either product substitutes for the other.
What would make the launch important for this use case?
A meaningful demonstration would begin with an identifiable appointment list, show what information is available, produce a relevant brief and expose uncertainty without making the clinician reconstruct the whole process. It would also explain what remains manual and which deployments can use it.
The strongest result would not be the shortest summary. It would be a preparation process that leaves the clinician better oriented, with less repeated searching and no hidden claim of completeness. Until those functions are demonstrated and documented, pre-charting remains a plausible Work direction rather than a verified specification.
Frequently asked questions
What is AI pre-charting?
It is the use of software to help assemble or organise information before a clinical encounter. Its usefulness depends on the available sources, the appointment's purpose and how clearly it shows omissions and uncertainty.
Could Heidi Work prepare a patient summary before an appointment?
Preparation is a plausible focus because Heidi's public event material highlights work before consultations. A Work-specific preparation workflow was not established by the public sources reviewed on 27 September 2026.
Would it need access to the medical record?
A limited brief could use information already supplied to Heidi, while a broader brief would need access to the additional relevant sources. Those starting points should not be treated as equivalent or assumed from an integration label.