Hi. Meet Meridian Bank. The name is fictional, but the challenge is real. Migrating out of their Slough data center to AWS. One application per wave across five Saturday night cutovers, all driven from a single plan. Three waves are down, but here's how they ran them. A static spreadsheet copied wave to wave with critical lessons buried in yellow cells that nobody actually reads at 10PM on a Saturday night. Cutover AI turns that static record into a continuous loop. Three features, one core idea. A runbook that learns from its own history. Let's break it down. First, AI Create. You take the legacy documents you already have spreadsheets, Word docs, PDFs and write a single paragraph describing your intent: streams, sequence, durations, and notes. Attach the file, hit create, and half a minute later, your runbook is ready. 12 tasks across four streams, proper dependencies, a go, no go milestone, and it even catches that yellow cell, planning DNS propagation for thirty minutes instead of ten. Second, the AI assistant. Just ask in plain language. You tell it to turn this runbook into a reusable wave template, swapping names for roles, stripping bridge numbers, and fixing approvers. The assistant drafts the edits, but nothing changes until you approve. One click applies all edits in a single batch. Without AI, human teams repeated the exact same two mistakes across three waves, skipping rollback rehearsals and underestimating DNS time. The data was there but no one acted on it. That brings us to the main building block, agentic tasks. These sit right inside your runbook plan. When execution reaches them, instead of waiting on a human operator, the task hands control to an agent. That can be your own organization's already approved custom agent. Pulling live data into the run, a monitoring check, a change record, a reconciliation result or Cutover's built in AI agent. The same tools as the AI assistant, proactively reading history, completing steps, and refining templates as the run unfolds. The difference is nobody has to ask, it acts automatically when the run reaches it, under the same approvals. Here's wave five, Treasury. The change manager creates the runbook from the template, and the live Saturday night cutover begins. The initial tasks are handled by people, confirming backups and change windows, and the rollback rehearsal that waves one to three skipped. Then we hit task 13, a pre cutover readiness check powered by an agentic task. Without anyone clicking a button, it triggers automatically, opens the conversation, and analyzes all four prior waves. Once you grant session approval for low risk actions, it runs unattended. Its job is to review everything that went before and give the change manager its own view on whether to go, posting a comprehensive briefing right before the gono go gate. Which tasks overran, which were skipped and why and what to do differently tonight, citing each wave by name. The final decision remains with a human and it's a go. The team completes the cutover, stopping the app, syncing data, repointing DNS, and running smoke tests. Once business sign off is complete, the plan hits the second agentic task. Fully automated, it pulls data across all five waves, compares planned versus actual durations, calculates task medians, and determines what needs updating in the master template. Then it proposes the changes. Anything that changes the plan waits for a person. It batches template updates into a single approval. Writes the change list back to the runbook. Runbook complete. Instead of someone spending hours building a post mortem slide deck, a clean wave report lands instantly on the runbook homepage. Even better, the template itself has learned. Future rollback rehearsals and closeout steps are now automatically scheduled with real world durations. Wave six starts off smarter. That is the Cutover AI loop. AI Create builds the runbook, the AI assistant shapes the templates, and agentic tasks execute, capture, and feed operational wisdom straight back into the stream. Every wave makes the next one better. And best of all, every screen you just saw was captured from a live un-staged runbook in Cutover. Thanks for watching.
