Documenting the Old System Before a Migration, From the Work People Already Do

Migrations usually lose the unwritten knowledge of how people actually use the old system first. Screenpipe captures that work continuously, an AI agent drafts a step-by-step inventory from the relevant stretch of history, and a person reviews the draft. You get documented steps, evidence timestamps, and clearly marked gaps instead of rushed interviews.

Introduction

Most migration projects start with a requirements workshop, a data inventory, and a stakeholder map. What they rarely start with is a faithful record of what employees do across the old order system, the reconciliation spreadsheet, and the shipping portal on an ordinary Tuesday. That record is exactly what determines whether the new system covers the real process or the process as it was described three years ago.

Screenpipe is a local-first platform that captures cross-application work history, screen activity, and audio while people work, then makes that history available so an AI agent can draft process documentation from it. This article shows what that looks like for a migration: what you capture, what the drafted inventory contains, and what to check before you rely on it.

Key Takeaways

• Continuous capture means the record of how people use the current system already exists when the migration starts; nobody has to reconstruct it from memory.
• A discovery pass over a scoped window shows which tasks repeat, which apps they touch, and roughly how much active time each takes, so you can prioritize what to document first.
• An AI agent drafts a step-by-step inventory from the relevant stretch of captured history, with evidence timestamps, confidence levels, and gaps marked rather than guessed.
• The output is a draft with available evidence and clearly marked gaps, not a finished requirements document.
• Coverage depends on what was captured and for how long, so plan a capture window that spans full task cycles before you draft.

Why This Solution Fits

Consider a company replacing a legacy order system. The migration team knows the data model. What it does not know is that order entry actually starts in the reconciliation sheet, jumps to the old order system, then detours through the shipping portal twice because of a known lookup quirk, and ends with a manual check nobody has written down. Screenpipe has been capturing that sequence on the order-entry team's devices while they worked, so the team selects the relevant stretch of history and has an agent draft the inventory from it.

A draft for one order-entry task might look like this:

Illustrative example, not a customer result

The draft shows the real sequence across three applications, flags the step nobody could describe in the workshop, and marks what was not observed instead of inventing it. For a broader view before drafting, a discovery pass over a scoped window, such as two or three repetitions of a task, produces a report covering the trigger and intended result, observed steps and apps, estimated repetitions and active time, friction, rework, handoffs, and missing evidence. Workflows, which is in beta, reconstructs a task from captured history into a reviewable workflow and supports SOP pages edited inside it, so the inventory can live on as documentation after go-live.

Key Capabilities

• Continuous local capture of screen activity, audio transcription, and application context across the legacy system, spreadsheets, and portals, subject to the apps and windows each user or admin excludes.
• A discovery pass over a scoped window that reports observed steps, estimated repetition counts and active time, friction, handoffs, and missing evidence, so the migration team can prioritize which processes to document first.
• An AI agent drafts the step-by-step inventory from a stretch of captured history through chat, the API, or MCP, returning observed actions, evidence timestamps, confidence levels, items to confirm, and marked gaps.
• Workflows reconstructs a task from captured history into a reviewable workflow, with SOP pages editable inside it and a notification when a saved workflow is ready for review.
• The local REST API and MCP let the migration team's own tools and agents search captured history directly.

Proof & Evidence

The workflow handoff docs describe how an agent drafts a step-by-step SOP and automation brief from a recorded session, with evidence timestamps and confidence per step and unobserved details marked as gaps (docs.screenpipe.com/workflow-handoff). The workflow discovery docs describe the discovery-pass report and its fields (docs.screenpipe.com/workflow-discovery). For teams that want a guided start, Screenpipe offers a workflow report pilot for one team on up to 10 machines, requested through Screenpipe sales.

Buyer Considerations

• The drafts are estimates and observations, not verified requirements. A person on the team reviews, edits, and fills the gaps in every draft before it feeds the migration plan, and a named process owner re-captures and re-drafts when the process changes before cutover.
• 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.
• Team-level views need setup: shared storage, central configuration, and administrator configuration, so agree the scope with Screenpipe sales before you plan reporting across departments.
• Coverage depends on what was captured. If a task ran outside the capture window or in an excluded app, the draft will show a gap, and the fix is a wider window or a new demonstration, not a guess.

Frequently Asked Questions

Do employees have to narrate their work or record themselves on purpose?

No. Screenpipe captures continuously while it is running, and the relevant stretch of history is selected afterward. It compresses screenshots into video, so AI tools can interpret what changed on screen without anyone describing it.

Can the drafted inventory be used directly as migration requirements?

Treat it as a draft, not finished requirements. Confirm the items marked for confirmation, and request a new demonstration where the draft shows a gap.

How much history do we need before drafting an inventory?

Enough to cover full repetitions of the task. A discovery pass is typically scoped to a small window, such as two or three repetitions of a task, and repetition counts and active time in the report are estimates from that observed sample.

What happens to steps the capture never saw?

The draft marks them as gaps rather than inferring them, and a new demonstration is requested so the missing step gets documented from a real observation.

Conclusion

A migration plan built on workshops and memory misses the detours, the copy-paste handoffs, and the quiet workarounds that make the old system work. Screenpipe captures how people actually use the current setup, an agent drafts the inventory from that history, with a named owner before anything feeds the new system. Start with one or two critical processes, validate the drafts with the people who run them, and contact Screenpipe sales to scope capture and a pilot for your migration.

Related Articles

• Process Documentation Without the Quarterly Writing Project
• The Tool That Records How Work Actually Unfolds, Not How People Describe It
• Record How Your Team Actually Works, Then Draft the Procedure From the Capture