Business lens demos/biz-05-raid-board.html Fictional data ← all artifacts
RAID log · Project Lighthouse

Project Lighthouse RAID log board

working session · Thu 11 Jun 2026 · facilitated by P. Mehta

ABC Corp is a fictional company. Every name, number and date is invented. This is a reference artifact generated with an LLM coding agent; the brief that produces it is at the bottom of this page.

This board does not save. It's a working surface for the meeting: export before you close the tab.

All items

IDTypeItemOwnerSeverityStatus
R-01RiskDuplicate customer records in wave-2 source dataUnresolved duplicates could corrupt billing history when batch-billing replays invoices after migration.S. de WitHIGHopen
R-02RiskVendor API rate limits during peak migration windowsThroughput caps on the pricing-service vendor API may stretch the wave-2 schedule past 26 Jun.R. OkaforMEDmonitoring
R-03RiskKey SMEs on leave through JulyDecommission review capacity is thin; deputies are named but have not yet shadowed a full review cycle.A. KowalczykMEDmonitoring
R-04RiskRollback procedure untested at wave-2 volumeThe staging dry-run on 18 Jun is the first test of order-gateway rollback above 50k records.M. TanakaHIGHopen
R-05RiskTraining material drift after tooling changeScreenshots in the onboarding deck predate the new case tooling; low impact, easily refreshed.J. LindqvistLOWopen
A-01AssumptionWave-2 data volumes match wave-1 profileCapacity planning assumes ±10% of wave-1 record counts; a profiling run will confirm by 16 Jun.R. OkaforMEDmonitoring
A-02AssumptionChange board approves decommission plan first passSchedule has no slack for a resubmission cycle; legal pre-review was done to de-risk this.P. MehtaMEDopen
A-03AssumptionNo regulatory reporting changes before go-liveThe compliance calendar shows nothing landing before 28 Aug; confirmed with the policy desk in May.A. KowalczykLOWclosed
I-01IssueEnvironment access delays blocked wave-2 startMigration started two days late because firewall changes for batch-billing missed the change window.M. TanakaHIGHopen
I-02IssueReconciliation report mislabels currency columnEUR amounts show under a column headed “local”; cosmetic but it confused the wave-1 sign-off review.S. de WitLOWclosed
I-03IssueDuplicate alerts from old and new monitoringBoth stacks page the on-call for the same order-gateway events; suppression rule is drafted.J. LindqvistMEDmonitoring
D-01DependencyVendor night-window batching agreementWave-2 throughput plan depends on the vendor honouring the agreed 22:00–05:00 batch window.R. OkaforHIGHmonitoring
D-02DependencyLegal review of decommission planSign-off moved to 17 Jul to fit the legal review cycle; everything after it shifts if this slips.P. MehtaMEDopen
D-03DependencyFinance freeze calendar for AugustGo-live cutover needs a two-day posting freeze; dates requested from finance, answer due 20 Jun.A. KowalczykMEDopen

Owner workload at a glance

Snapshot of the 14 seeded items by owner, so nobody leaves the session without knowing what they hold. High-severity counts are what to watch.

OwnerItemsHigh severityOpen / monitoring
M. Tanaka222 open
R. Okafor313 monitoring
S. de Wit211 open · 1 closed
A. Kowalczyk32 open · 1 closed
P. Mehta22 open
J. Lindqvist21 open · 1 monitoring

Add an item (during the meeting)

New items get the next free ID for their type, status open, and exist only in this tab until you export.

Export

Exports the current filtered view (all items, current order).
How this was made: the brief, how to reproduce it, and an honesty note

The brief

Build a single-file HTML RAID log seeded with these items [paste].
I want to filter by Risks/Assumptions/Issues/Dependencies, add items
during the meeting, sort by severity, and copy the whole log out as
a Markdown table afterwards. Make it obvious that the page doesn't
save anything.

How to reproduce

Paste the brief into any capable LLM: GPT, Claude, Gemini, Grok, DeepSeek, or the assistant your company provides. Iterate a few rounds on layout and content until it reads well. Save the final answer as a .html file and open it in any browser. Expect similar output, not identical: every model has its own taste, and that is fine.

Honesty note

This reference artifact was built with Claude Code, an LLM coding agent, over several iterations. Treat it as the bar to aim for, not as a guaranteed first answer. All data on this page is fictional.

Next artifact Town-hall announcement →
vishalshah.app