The AI assistant in CaRD Plan: with your own AI from the plan to the lessons learned
Whether OpenAI, Claude or Llama in your own data centre: you connect your AI service per tenant and switch the functions on one by one. The AI prepares, summarizes and suggests – people still decide and release.
Request a demoPreparation
Plan from documents
Upload an Excel schedule, Word runbook, PDF work instruction or text: the AI proposes the tasks with order, duration, group and owners. You select and create catalogue entries and a template from them – the fastest way from an existing Excel sheet to a rehearsed plan.
Content review of the plan
Complements the consistency check with what a rule cannot see: missing data backup or fallback, missing second check before a release, questionable order, unclear instructions.
Improve instructions
Translate descriptions into the five interface languages, turn them into a checklist or make them clearer – right in the catalogue.
Impact of a change
For management of change: describe the change, the AI names affected areas and matching activities from the catalogue – which become a template.
Execution
Status report at the push of a button
Progress, delays, blockers and the next tasks in a few sentences – for management or straight to the Microsoft Teams channel.
Report back by photo or voice
Read values from a photo of a test report or type plate, turn a voice note on the phone into the comment – marked, and only kept with “Save”.
Did we have this before?
When a problem comes up, the AI shows what happened with the same task in test runs and earlier live runs – and what helped back then.
Follow-up
Lessons learned and a better template
Planned and actual times, delays, force reasons and comments become a draft of the lessons-learned report – plus suggestions for the template, such as more realistic durations, applied with one click.
Ask your data
Ask questions in your own words – such as which tasks had the most delay in recent runs or what the last live run cost.
Your AI – your choice
OpenAI
Directly or via Azure OpenAI in your Azure region – also other services with an OpenAI-compatible interface.
Claude
Anthropic’s models via the Anthropic API.
Llama
Open models via Ollama in your own data centre – the data does not leave the company.
Control and data protection
- Off by default: without setup, no data goes to an AI service
- Switchable per tenant and per function, daily limit against unexpected costs
- The key is stored encrypted; your AI service bills you directly
- AI texts are marked as drafts; the AI never sets a status and never releases anything
- Every request is logged with time, person and function – without content
- No evaluations of individual persons
Frequently asked questions
Do I need AI to use CaRD Plan?
No. The AI assistant is an add-on you can switch on per tenant. CaRD Plan works fully without it – switched off, the application looks as always.
Which AI services are supported?
OpenAI (including Azure OpenAI and other services with an OpenAI-compatible interface), Claude by Anthropic and Llama via Ollama in your own data centre. You enter address, key and model yourself.
Does my data go to the AI?
Only if you switch the assistant on – and then, for each request, only the data needed for it, to the service you connected yourself. If data must not leave the company, use Llama in your own data centre.
Who decides – the AI or people?
People. The AI suggests, summarizes and reads; nothing is applied without your confirmation. It never sets a status, never closes a task and never releases anything.
What does the AI assistant cost?
Your AI service bills the usage directly. In CaRD Plan you set a daily limit for requests and see the usage of the last 30 days.
See your own procedure in CaRD Plan
Tell us briefly what it is about – in 30 minutes we show you what the plan looks like.
CaRD Gesellschaft für EDV-Beratung mbH
Gutenbergring 2 · 69168 Wiesloch, Germany
Phone +49 6222 9256-0 · E-mail info@card.de