Five sittings between a named app and a readable week

Intake · 60–90 minutes

Name the app

We agree which production app counts, which environment is in bounds, and which days form the window. You name a counterpart. If logs cannot leave the building, we book sitting days in SS15 or at your table instead of pretending a screenshot is a window.

This is also where we refuse work: no observable app, no counterpart, or a request to rewrite prompts under the assessment fee.

Days 1–3

Take the call inventory

Every invocation site gets a human name, a mark for waiting or not, and columns for retries and fallbacks. Route names sit beside the human names; they do not replace them. Unlocated calls stay on the list as a finding.

If your own list is missing a nightly job or a save-hook, that absence is usually the first useful sentence in the report.

Days 3–10

Read the window

We reconstruct the agreed days from traces, request logs, a supervised export, or a sitting session. Provider invoices are a cross-check, not the inventory. Gaps are written down. Timestamp-only reconstructions are labelled as inferred.

Wei Liang stops the read if parent and child calls cannot be told apart and the team expected them to be. That stop is cheaper than a confident paragraph.

Days 10–13

Write the picture

Page one is the inventory in waiting order. Then the narrative of the window. Then the gap list and the disputed items. Token totals, if they appear, are a footnote. Aisha writes in sentences that can be read aloud in a mixed room.

Day 14 or 15

Walk through it

Ninety minutes. The document is already in your hands. We read it together — in person at SS15, at your office in the Klang Valley, or on a call if the room is split. Priya attends when product, engineering, and operations will share the table. Slides do not replace the report.