Article · AI and automation · Dan Tardif
Screenpipe: from morning brief to meeting follow-through
I connect my work history to email, my calendar and project tools. The goal: have AI prepare my priorities, turn a meeting into a brief and carry out the tasks I assign.
I want to start my day knowing which projects need attention: what I promised yesterday, what is blocked today and what I need before my next call. Those answers are spread across meetings, messages and project tools.
I use Screenpipe to pick up the thread of that work. Combining it with other connectors is what makes AI especially useful: it can recover a decision, check the current project status, then prepare or carry out the next steps.
What Screenpipe adds
Screenpipe keeps the screen activity and audio I allow it to record. My AI assistant can search that history for a conversation, something I read or the reasons behind a decision. It is a memory of my work I can query: “Why did we change the launch date?”
A good fit for my workday
I am an ideal candidate for this kind of tool: I work 100% remotely, and all my professional interactions go through my computer. Teams calls, email, research, media plans and campaigns all happen there. Screenpipe can retain that context, including what happens between meetings. I can retrieve it without having to explain everything to AI again.
Memory becomes useful when it connects to other tools
A connector gives AI access to a tool: Gmail or Outlook for email, a calendar for appointments, Jira for tasks and advertising platforms for results. Some connectors provide read access; others also let the assistant create tasks, edit documents or prepare drafts inside an application.
Screenpipe provides the history. The original tools provide the current project status. The assistant brings both together and can work in connected applications. That is the combination I am building around: a meeting decision can become a brief, assigned tasks and documents ready for review.
- ScreenpipeMeeting decisions and commitments
- Email and calendarRecent requests and deadlines
- Business toolsTask status and current results
The assistant prepares the brief and tasks
Connectors carry out authorized actions
Three business scenarios
These are fictional examples of requests and expected results. They illustrate the approach I am putting in place; available actions depend on the connectors and permissions provided.
The strategic morning brief
At 9 a.m., my brief should identify projects that need a decision. Screenpipe recovers yesterday’s commitments; the calendar identifies meetings to prepare for; email and project tools help check what has changed.
Prepare my three priorities for today. Cross-check meeting commitments against my calendar, recent messages and task status. For each priority, give me the issue, deadline and next action.
The result should help me decide what to do first. For example:
- Before 10 a.m.: get approval for the creative blocking Friday’s launch. Prepare the follow-up message.
- Before the 2 p.m. call: gather campaign results and compare them with the objective agreed on in the last meeting.
- Today: check whether yesterday’s promised deliverable was sent. If no confirmation is found, keep the follow-up open.
The next step can also be assigned to AI: prepare the figures, draft the follow-up and update the relevant tasks.
Leave a meeting with a usable brief
A marketing team is preparing a Friday launch. During the call, it agrees on the offer, two ad versions to produce and a need for approval before noon on Thursday. I want those instructions in a form the team can use.
Use this meeting to write the production brief and action list. Keep the owners and dates actually agreed on. Flag missing information.
The assistant can structure the result into three parts:
- Brief: the objective, audience, offer, messages and requested formats.
- Deliverables: two ad variants and a landing page to check.
- Actions: who produces, who approves, by when and what depends on that approval.
It can also compare the brief with recent emails to catch an instruction changed after the meeting. A transcript summary alone would miss that update.
Have AI carry out the actions
Once I have reviewed the brief, I can assign the execution. The assistant then uses connectors with write access to the tools. Screenpipe helped recover the request; the other connections handle execution.
Create Jira tasks from the approved brief, with the confirmed owners and deadlines. Write the two ad variants and add them to the working document. Prepare the follow-up email as a draft.
I can then check concrete results:
- Links to the created tasks, with their owners and dates.
- The document containing the copy to review.
- The email draft ready to send.
AI handles part of the coordination and production work. I review the brief and deliverables; sending, publishing and campaign changes follow the agreed permissions.
- The meetingScreenpipe recovers the instructions
- The briefAI extracts deliverables and actions
- The reviewI confirm the request and priorities
- The executionTasks, documents and drafts created
How is this different from Copilot Premium?
Copilot Premium works with the email, documents and meetings it can access, as well as connected tools. Screenpipe adds another source: a memory of what I saw and heard on my computer, even if I never wrote it down.
For example, I compare two offers in supplier portals, then close the tabs. The next day, I want to recover the terms of the second offer.
If I did not record those terms and the portal is not accessible to Copilot, it does not have that information. Screenpipe can recover it from the captured screen history.
What interests me is recovering information I never thought to file away. I can then ask my assistant to use it with other connectors to prepare a comparison, update the project or create a task.
Here, Copilot Premium means Microsoft 365 Copilot for work accounts. Screenpipe can only recover material that was actually recorded.
What I check
A transcript can be incomplete, and a task discussed on a call may have changed since. I check important points in the original tools. I also choose what may be recorded or sent to AI: raw data stays local by default, but cloud AI or synchronization can send data outside the computer.
Try it on a simple project
Start with a recent meeting: recover the decisions, have the brief prepared, then choose an action to assign to AI in a connected tool. You will quickly see where that memory helps your work.
Explore Screenpipe through my affiliate link ↗Affiliate link: I may receive a commission if you make an eligible purchase through this link.
Sources and notes
Comparison checked against official documentation on October 22, 2026. Scenarios are fictional; they do not describe a fully deployed execution workflow or a client result.