ChatGPT knows a lot. Inora has your protocol.
People reach for ChatGPT because the answer is not at hand. Inora gives them that answer from your own agreements, on record and safe for client data.
Nursing home · night shift · 03:10
ChatGPT
1 general answer · 2 steps by hand · nothing on record
Inora
Asked once · 3 steps taken off their hands · all on record
Mr Visser has fallen and is lying next to his bed. Can I lift him myself?
TypedAsked once · 3 steps taken off their hands · all on record
- KnowABCDE first, do not lift alone
With pain in the hip, leave him lying and call the on-call colleague; then use the hoist.
Protocol after a fall §2 - DoOn-call colleague called with an SBAR
The SBAR was ready before the nurse called.
- RecordReport ready
Into the record on the care worker’s approval; the incident report waits as a draft.
Report
From your protocol, and rated well. By the people who work with it.
95%
of answers to care questions rated well by care workers
The same question, now with your protocol.
- What was saidWhere it differsChatGPT
In the worker’s own account
In a personal account the chats sit with the worker. The worker can delete them; a deleted chat cannot be restored.
source: help.openai.comInoraOn record at Axiomatic
Every answer is traceable to its source and the version that applied at the time. 15 steps behind every question, every one on record.
- Your own agreementsChatGPT
What the workspace connects
On Business, Enterprise and Edu, the Company Knowledge plugin also searches connected apps, with source links where the answer gives them.
source: help.openai.comInoraYour whole handbook, sourced
Your handbook organised as a system, with owner and version, and the source behind every answer.
- The day preparedChatGPT
Tasks the user schedules
Tasks at a time the user sets, once or recurring, and on events the user picks, where available.
source: help.openai.comInoraThe shift already prepared
Before the shift starts, today’s clients and what changed since yesterday are ready.
- At night, aloneChatGPT
An answer to check yourself
OpenAI itself: ChatGPT can be wrong, even when it sounds confident. Check important information.
source: help.openai.comInoraYour own protocol, at 03:10 too
From your protocol, with the source to check. 95% of answers to care questions rated well by care workers.
- What staysChatGPT
The user’s own chats
On Free, Go, Plus and Pro the user exports their own chat history as a ZIP. On Business, Enterprise and ChatGPT for Healthcare, through the workspace owner.
source: help.openai.comInoraThe handbook and its structure
Your organised knowledge, the files, the structure and the record of every answer. The organisation’s.
Show all 6 points
- Client dataChatGPT
Depends on the plan
On Free, Plus and Pro, OpenAI may use the content for training unless the user turns it off. On Business, Enterprise and Edu, not by default.
source: help.openai.comInoraIn the EU, trains no model
In the EU, and it trains no model, not even the model provider’s. Built to NEN 7510 and ISO 27001.
The same question at 03:10. Now with your organisation’s protocol beside it.
Hello
Draft the SBARPrepare the incident report
Your night shiftFri 6 Nov
- Mr Visser03:10From your fall protocol, §2First ABCDE, do not lift aloneWith hip pain leave him lying and call the on-callThen with the hoist, with two colleagues
- Mrs Smit04:00
- Mr Kok05:30
- Mrs Bos06:15
- SBAR for the on-callDrafted, to call with
- Incident reportDraft, waiting for your approval
ChatGPT knows a lot. How your organisation does it is in your handbook.
- Fall protocol §2The version that applies todayOn-callNumber at hand
Your protocol, with source.
No general answer, but the version that applies today.
- ProtocolAnswerDraftedSBAROn approvalReport
Then the work.
SBAR drafted, incident report as a draft, report for your approval.
- Client dataIn the EU, trains no modelSign-inSSO and SCIM, on your own accountsBuildTo NEN 7510
Your own environment.
Client data in the EU, and your data never trains a model.
Read the full comparison
Your people already use ChatGPT, on their own phones, and it helps them. A ban takes that help away. A better tool takes away only the risk.
Your people looked for help, and found it.
ChatGPT reasons and writes well, instantly, for anyone. Wording a report properly, explaining what a medicine does, starting a difficult email: it helps, and your people notice. Nobody did anything wrong by looking for help.
On their own phones, ChatGPT answers in general.
The protocol after a fall, the agreement with the on-call colleague, the format of the incident report: that is in your handbook. ChatGPT on a personal account answers from what it knows in general, and what it said then sits with the worker.
OpenAI also sells ChatGPT for Healthcare to care organisations. It reads SharePoint, Teams and Outlook so answers reflect the organisation’s approved policies and care pathways; it does not train on the data, and the data can stay in Europe (OpenAI, OpenAI). It finds what is in the folder, with no owner or version that counts today, and the follow-up stays by hand.
The Dutch Data Protection Authority received several reports of data breaches because staff shared personal data with an AI chatbot. In one, an employee of a GP practice had entered patients’ medical data, against the agreements made (Autoriteit Persoonsgegevens).
Client data belongs in your organisation’s own environment.
Between visits, a care worker enters client information. With Inora, your organisation gets its own EU environment for the procedure and the report; your data never trains a model.
At 03:10 your own protocol counts.
Mr Visser has fallen and is lying next to his bed. The nurse types the same question they would otherwise have typed into ChatGPT: can I lift him myself? Inora answers from their own organisation’s protocol, the version that counts today: ABCDE first, with pain in the hip leave him lying, call the on-call colleague. The nurse calls with an SBAR that is ready; the report goes into the record on their approval; the incident report is ready as a draft in Inora.
The easy tool is also the safe one.
Inora is the one tool the organisation provides itself. For every answer Axiomatic records what it rested on: the source and the version. What happens, the professional decides.
“I have worked in care for twenty years, but Inora gives me ideas I had never thought of.” Eight weeks. Your manager signs.
Live from day 15
One team, a fixed price, within your manager’s own budget. You see up front what each part costs.
Stop after the evaluation, no notice period
After eight weeks you decide whether to go on.
Two regions chose a second year
ZorgConnect Zuid-Limburg (16 organisations) and VVT Platform Zuidoost-Brabant (13), after their first year with Inora.
Your people looked for help. Give them the right kind.
Do we ban ChatGPT?
No, offer something better.
Your people went looking for help, and they were right to. Give them one good tool and make clear agreements on AI chatbots, as the Dutch data protection authority advises organisations to do.
Isn’t Inora just ChatGPT with our documents?
No.
Inora organises your agreements as a system, with owner and version; the owner sees what is out of date. After the answer, Inora drafts the follow-up and drafts the report, which goes into the record on approval.
Can our staff enter client data?
Yes.
Client data stays in your organisation’s own environment in the EU and never trains a model. An answer to a protocol question needs no name or citizen service number.
What does the pilot mean for our care workers?
Your pilot team works with Inora from day 15.
Your care workers get answers from your own approved work instruction, day and night, with source and version.
What if we take ChatGPT Business, Enterprise or ChatGPT for Healthcare?
Then you gain privacy, not yet a system.
OpenAI then does not train on your data by default, it can stay in Europe with Enterprise and ChatGPT for Healthcare, and ChatGPT also searches your connected apps, with a source. What it finds there is what is there: no owner and no version that applies today, no follow-up as a draft and no competence file.
Inora works in the apps your team has open.
Also compared: Attendi · Microsoft Copilot · The intranet handbook · Zenya · Tandem Health · Autoscriber · Building it yourself · Nedap Ons · Vilans Protocols. All comparisons