PAI-DOC-15
2026-07 · Sectors
AI in a healthcare practice: documents that answer back, within a framework that holds.
In a healthcare practice, AI can pre-draft letters and reports, make records queryable in natural language, reconcile consumable invoices and equip the reception desk, under systematic human oversight.
Healthcare is our anchored sector: DocAgora, our e-health platform in Portugal, operates in production with health data and a strictly supervised conversational agent.
The framework is as strict as it gets, and it is our standard: sovereign hosting, no data going to a US API, systematic logging.
01
A practice produces documents all day long, and they look alike.
The letters
Reports and referral letters pre-drafted from the record, reviewed and signed by the practitioner before anything is sent.
The queryable record
Find a history item, a treatment or a letter in natural language, instead of reopening every document.
The front desk
Appointment requests sorted and pre-qualified; complex calls stay with the front desk, which decides.
Quotes and financial agreements
Quotes and financial agreements prepared from the treatment plan, checked before they are presented to the patient.
Supplier invoices
Consumables and lab work: invoices reconciled against orders, discrepancies flagged to the manager.
Reminders
Appointment and follow-up reminders drafted and scheduled, approved by the practice before they go out.
The equipped reception
A supervised conversational agent for practical information and appointment booking: this is exactly what DocAgora runs in production.
Coordination
Schedules and handovers prepared from incoming requests, arbitrated by the coordination team.
Compliance
AI-usage register and systematic logging: the health-data framework requires being able to show who accessed what.
Fig. 01 · Three profiles, a set of possibilities: the exact scope is set at the diagnostic.
Referral letters, reports, quotes, financial agreements, consumable invoices, appointment reminders: the administrative day-to-day of a healthcare practice is a flow of documents, and that is exactly the material AI works well on.
The use cases above read by profile, in your own vocabulary. They describe possibilities: what actually applies to you is decided at the diagnostic, based on your volumes and your tools.
02
DocAgora is our proof in healthcare, and it runs in production.
Healthcare is the one sector where we do not need the conditional: DocAgora, our e-health platform in Portugal, the local equivalent of a Doctolib, operates in production with health data.
Its conversational agent is strictly supervised: practical information and appointment booking, never medical advice; human oversight on every sensitive action; systematic logging of every conversation.
It is the standard we apply to every healthcare project: if the architecture holds for a patient-care platform, it holds for a practice's front desk.
03
Health data imposes the strictest framework, and it is a known one.
Health data is sensitive data under the GDPR: processing it requires a solid legal basis, appropriate hosting and complete traceability. In France, the hosting of health data is itself regulated.
Our architecture answers by design: a self-hosted model or one on a sovereign French cloud, no data going to a US API, including during demonstrations, and an audit trail produced by the very way the system works.
The usage policy counts double here: a general-purpose public assistant has no place in a practice; the EU AI Act also requires demonstrable AI-literacy measures from any team that uses these tools.
04
The practice's management follows the same circuits as any SME.
Beyond care, a practice is still a business: consumable and lab invoices, rent, contracts, social-security charges. These documents can follow our document processing circuit: capture, reading, reconciliation, alerts on discrepancies.
The e-invoicing deadline also applies to healthcare structures established in France: from 1 September 2026, every company established in France must be able to receive electronic invoices, under the reform led by the French tax administration (DGFiP). A reconciliation circuit installed before the deadline absorbs it without a jolt.
05
A healthcare project follows the protocol, with reinforced scoping.
A healthcare project follows our five-stage engagement protocol: Diagnostic, Scoping & compliance, Pilot, Deployment, Managed operations. The scoping stage carries more weight here than anywhere else: legal bases, hosting, access rights, logging.
The engine is demonstrated on public data before it touches yours, and every extension is decided on figures measured at your premises. We are paid for the project, never for reselling licences.
06
Continue
07
Commitments
Four commitments, on every engagement
These commitments are set out in our presentation deck. They hold for every engagement, from the diagnostic to managed operations.
- Contact
- Dedicated, end to end
- Progress point
- Weekly
- Environments
- Permanent access
- Reply
- Within 24 working hours
08
Decide
Two minutes to know where you stand.
A straight answer: what applies to you, and whether a project makes sense. If the answer is no, you leave with the answer.
Regulation (EU) 2024/1689 · PAI-DOC-15