Starting point

Meetings & tasks
Turn meetings into concrete work
Every commitment, decision and deadline from a meeting is recognised automatically and created as a task with an owner directly in the right work system. The minutes remain as a record, but the actual work starts right after the meeting – not once someone reads through the minutes again.
Process architecture
From the spoken word to an assigned task.
The process takes the recording or transcript of a meeting, recognises commitments, decisions and deadlines within it, and creates structured tasks with an owner in the right work system. Ambiguous commitments without a clear owner go to the meeting lead for clarification instead of being assigned automatically to whoever happened to be mentioned.
01Input
Open quotes and their context come together.
The process starts with silence rather than a message: a quote that has had no answer for days. The leading system supplies the quote, its deadline and the contact, together with the correspondence already attached to the case.
2 Steps
02Process
The right message is derived from the state of the case.
The draft comes from the quote itself, not from a template full of placeholders. The rules then check whether a follow-up is allowed at all; anything outside the agreed frame goes to a salesperson.
4 Steps
03Output
The message goes out and the case stays traceable.
The follow-up sits in the same conversation as the quote, sales sees the current state without asking, and every attempt is documented on the case.
3 Steps
Source01Record the meeting
A meeting is captured by recording or transcript – via Microsoft Teams, Google Meet or a custom conferencing system, depending on the environment.Details
Record the meeting
A meeting is captured by recording or transcript – via Microsoft Teams, Google Meet or a custom conferencing system, depending on the environment.
- Microsoft Graph API
- SharePoint API
- Create the transcript
Data02Create the transcript
The audio or video recording is turned into a searchable transcript with speaker attribution.Details
Create the transcript
The audio or video recording is turned into a searchable transcript with speaker attribution.
- Azure AI Document Intelligence
- Mistral OCR
- Dataverse API
- Recognise commitments and decisions
Action03Recognise commitments and decisions
A language model reads the transcript and flags passages that contain a task, a decision or a deadline.Details
Recognise commitments and decisions
A language model reads the transcript and flags passages that contain a task, a decision or a deadline.
- Power Automate
- Azure Functions
- Microsoft Graph API
- Assign the owner
Decision04Assign the owner
For every flagged commitment, the process checks whether a person was clearly named in the conversation; if not, the passage counts as ambiguous.Details
Assign the owner
For every flagged commitment, the process checks whether a person was clearly named in the conversation; if not, the passage counts as ambiguous.
- Azure OpenAI GPT
- Anthropic Claude
- AI Builder
- Draft the taskstandard case
- Clarify an ambiguous commitmentexception
Approval05Clarify an ambiguous commitment
For missing or unclear ownership, the meeting lead receives the affected passage for a brief confirmation before a task is created.Details
Clarify an ambiguous commitment
For missing or unclear ownership, the meeting lead receives the affected passage for a brief confirmation before a task is created.
- Teams Adaptive Cards
- Power Apps
- Draft the taskafter approval
System06Draft the task
The clarified commitment becomes a concise task title with context from the conversation, an owner and, if mentioned, a deadline.Details
Draft the task
The clarified commitment becomes a concise task title with context from the conversation, an owner and, if mentioned, a deadline.
- Dynamics 365 API
- Azure Service Bus
- Create the task in the work system
- Notify those involved
Result07Create the task in the work system
The task is created directly in the work system in use, with a reference back to the passage in the transcript it came from.Details
Create the task in the work system
The task is created directly in the work system in use, with a reference back to the passage in the transcript it came from.
- Power BI API
- Teams Webhook
- File the minutes with the task list
Result08Notify those involved
The owner receives a short message with the new task, without having to read the entire set of minutes.Details
Notify those involved
The owner receives a short message with the new task, without having to read the entire set of minutes.
- Power BI API
- Teams Webhook
Result09File the minutes with the task list
The full transcript and the list of every recognised task are filed together, as a record for later questions.Details
File the minutes with the task list
The full transcript and the list of every recognised task are filed together, as a record for later questions.
- Power BI API
- Teams Webhook
System landscape
The same process in three environments.
- Microsoft
Microsoft Graph API
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
SharePoint API
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Azure AI Document Intelligence
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Mistral OCR
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Dataverse API
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Power Automate
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Azure Functions
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Azure OpenAI GPT
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Anthropic Claude
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
AI Builder
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Teams Adaptive Cards
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Power Apps
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Dynamics 365 API
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Azure Service Bus
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Power BI API
Möglicher Baustein für die Microsoft-Systemvariante.
- Microsoft
Teams Webhook
Möglicher Baustein für die Microsoft-Systemvariante.
- Interfaces in use
Google Workspace APIs
Google environment: reads the Meet recording and creates tasks in Google Tasks.
- Interfaces in use
REST API
Individual environment: connects the custom conferencing system with the existing project or ticketing system.
- Interfaces in use
Azure AI Speech
Creates the searchable transcript with speaker attribution in the Microsoft environment.
- Interfaces in use
Google Cloud Speech-to-Text
Creates the searchable transcript with speaker attribution in the Google environment.
- Interfaces in use
Webhook
Forwards a newly created task to the target system immediately, without waiting for the next scheduled run.
- Possible apps & platforms
Microsoft Teams und Microsoft Planner
Recording of the meeting and storage of tasks in the Microsoft environment.
- Possible apps & platforms
Google Meet und Google Tasks
Recording of the meeting and storage of tasks in the Google environment.
- Possible apps & platforms
Jira oder Asana
Target system for tasks in the individual environment, independent of the conferencing system in use.
- Possible apps & platforms
Slack
Notification of the owner about the new task in the individual environment.
Implementation in context
What the process does and what it deliberately leaves to people.
The process replaces neither facilitation nor a substantive decision – it only makes sure a commitment made in conversation is not lost again but becomes visible as a task.
The environment determines the tools, not the process. In a Microsoft 365 environment tasks arise from a Teams recording directly in Planner, in a Google Workspace environment from a Meet recording in Google Tasks, in a self-hosted environment from a custom conferencing system directly in the existing project or ticketing system. Which variant fits depends on where the company already runs its meetings and its tasks today.
Recognition of commitments is deliberately cautious: only passages with a clearly named owner become a task automatically. If the conversation lacks that assignment, the passage goes to the meeting lead for a brief clarification instead of assigning a task to whoever was merely mentioned in passing.
The full transcript stays alongside the task list and points to the exact passage in the conversation a task came from. That way every task can be traced back to its original wording if needed.
Target picture
Every commitment from the meeting becomes visible as its own task with an owner in the right work system, right after the conversation ends. The full minutes remain as a record, but are no longer the only source the work has to come from.
Expected benefits
- Commitments from the meeting become visible as a task immediately, instead of disappearing into the minutes.
- Every task has an owner and, where mentioned, a deadline – without manual follow-up work.
- Ambiguous commitments are recognised and presented to the meeting lead instead of being assigned incorrectly by default.
- The owner receives the task directly in the work system they already use, without having to read the minutes.
- The full minutes remain as a record and stay findable at any time through the task.
- The process adapts to the existing environment instead of making an additional meeting tool a precondition.
Prerequisites to start
- 01A conferencing system that provides a recording or transcript of the meeting.
- 02A work system for tasks that the process is allowed to write to.
- 03A named meeting lead per meeting type who can quickly clarify ambiguous commitments.
- 04Clarity on which recurring meetings are in scope for automatic task recognition.
- 05A decision on the environment: Microsoft 365, Google Workspace or self-hosted – and the matching access rights.
Your next step
Which process costs your team unnecessary time every day?
Give us a short outline. We will tell you honestly where automation is useful—and where it is not.






