Workflow Mining When Your Work Lives in Spreadsheets, Email, and Web Portals

When the real process is "download from the portal, paste into the tracker, email the exception," the useful question is which steps repeat, which apps they cross, and roughly how long they take. Screenpipe has been capturing that cross-app work locally, and a discovery pass turns a stretch of that history into a written report.

Introduction

If your team's workflow is spread across a web portal, two spreadsheets, and an inbox, traditional process mining has a hard job reaching it. Process mining reads event logs from systems like ERP and CRM. Task mining captures desktop-level activity to show how individual tasks are done across apps, which is the layer where spreadsheet-and-portal work actually happens.

Screenpipe is built for that layer. It runs on macOS, Windows, and Linux, captures screen OCR, audio transcription, and accessibility metadata continuously while it is running, and stores the history locally. From that history, you get a time summary of where the week went, a discovery report on a repeated task, and an agent-drafted SOP for the process you decide to document. This article walks through how that works for a purchase-order team, and what to check before you buy.

Key Takeaways

• Screenpipe captures work continuously across every app on the desktop, including spreadsheets, email, and web portals, so the process does not have to live in one system to be visible.
• The time summary shows where time went across the desktop, and a discovery pass produces a written report on a repeated task with observed steps, repetition counts, and estimated active time.
• The result is a shortlist of automation candidates to validate in a pilot, not a finished ranking.
• The output is concrete: a report excerpt you can read in a few minutes, with friction, handoffs, and missing evidence marked.
• Team-level reporting depends on shared storage, central configuration, and administrator setup, which you scope with Screenpipe sales.

Why This Solution Fits

Consider an operations team at a distributor that suspects purchase-order entry is its biggest time sink. The work does not happen in one place: an order arrives by email, someone keys it into the ERP portal, and the shipment is logged in a tracking spreadsheet. Nobody can describe the full sequence from memory, and the process document, if one exists, was written a year ago.

Because Screenpipe has been capturing in the background, the team does not need to stage a recording session or interview anyone. The time summary shows where the week went, and in this example it points to the mail client, the ERP portal, and the tracking spreadsheet as the windows where PO work accumulates. The team treats that as a candidate, checks it against a second week of captured history, and then pilots one improvement.

For the deeper look, Workflows, which is in beta, reconstructs how a task was done from captured history into a workflow a person can review. Its discovery pass is scoped by a person to a concrete boundary, such as three repetitions of the PO task in a time window, and produces a report the team can read. Here is what an excerpt looks like:

Illustrative example, not a customer result.

Three repetitions at about 22 minutes each is an estimate from a small observed sample, and the team ranks candidates afterward using frequency, active time, failure cost, and ease of testing.

This fits spreadsheet-and-portal teams because capture follows the person, not the system. The portal, the tracker, and the inbox are all just windows on the desktop, so the handoffs between them, which is where the real process lives, are part of the record.

Key Capabilities

• Continuous local capture of screen OCR, audio transcription, and accessibility metadata across apps while Screenpipe is running, subject to the apps and windows a user or admin excludes.
• The time breakdown pipe summarizes where time went across the desktop, giving the team a factual starting point for the "what should we automate first" conversation.
• The Workflows discovery pass produces a per-task report with observed steps, estimated repetitions and active time, friction, handoffs, and missing evidence, written as a local Markdown report when run as a scheduled task.
• The weekly automation-candidates job described on the use-cases page: a weekly schedule over repeated work that surfaces repeated actions, handoffs, and time sinks as candidates.
• An AI agent can draft an SOP from a stretch of captured history through chat, the API, or MCP, returned as a draft for the team to edit before it is shared.

Proof & Evidence

The workflow discovery docs describe the discovery report and its scoping in detail, including how the user sets the boundary for a pass. For teams that want a guided starting point, Screenpipe offers a workflow report pilot for one team on up to 10 machines, described on the task mining page. The pilot structure is to measure a baseline, test one improvement, then build the business case before expanding.

Buyer Considerations

• Screenpipe does not publish a standalone ranked-by-hours or team-level ranking report. The time breakdown and discovery reports produce estimates, not verified figures, and team-level views depend on shared storage, central configuration, and administrator setup. Scope what you need with Screenpipe sales.
• Tell people what is recorded. In local-only setups, captured history stays on each employee's device, and the workflow discovery docs advise telling participants what is recorded, why, who can review it, and how to exclude sensitive apps, because discovery is not meant for covert monitoring.
• A person reviews and owns the output. The discovery report, the SOP draft, and the pilot decision all involve human review, and when a process changes, a named process owner re-captures and re-drafts.
• Coverage depends on capture. Results reflect what was actually captured after exclusions and how long capture has been running; if no matching data exists for a task, you widen the range or fix capture.

Frequently Asked Questions

Do we need to move our work into one system for this to work?

No. Screenpipe captures at the desktop level, so a process spread across a web portal, spreadsheets, and email is captured the same way a single-system process is. The handoffs between windows are part of the record.

Where do the time figures come from?

The time breakdown pipe and discovery reports derive estimates from observed windows in captured history. Treat them as estimates from a sample, check them against a second week, and validate candidates in a pilot before committing budget.

Can Screenpipe rank tasks across a whole team automatically?

Team-level views are a setup question rather than a self-serve report, so agree the scope with Screenpipe sales before counting on them.

Does someone have to record each task deliberately?

Capture is continuous while Screenpipe is running, so nobody stages a recording session for a task. The relevant stretch of history is selected afterward, when the team decides what to examine.

Conclusion

For teams whose real process lives in the space between a portal, a spreadsheet, and an inbox, Screenpipe starts from what actually happened on the desktop: continuous local capture, a time summary and discovery report built from that history, and a shortlist of automation candidates to validate in a pilot, with a named owner at each step. If that matches the visibility gap you are trying to close, contact Screenpipe sales to scope a pilot for your team.

Related Articles

• Task Mining for Order-to-Cash: See the Steps People Actually Take
• Task Mining on Managed Laptops Without an Integration Project First
• What Tools Uncover the Manual Work Hiding in Browser Tabs and Spreadsheets?