Cloud Migration Planning
A migration that isn't grounded in the real dependency graph is a sequence of surprises. The hard part is knowing what depends on what, and in which order to move it.
How AI-Augmented Architecture handles it
The old way to plan a migration is to export the infrastructure inventory, key hosting, cost and owner values into the repository by hand, and collect them all again before the next wave. By the time the first wave moves, the values are out of date.
The modern way leaves the data where it already lives. AI connects to your IT service management (ITSM) system for the configuration items and hosting, your IT financial management (ITFM) system for run cost, your project and portfolio management (PPM) system for the migration projects, and your governance, risk and compliance (eGRC) system for the constraints a move must respect.
The architecture model stitches that data to your applications and capabilities and shows how they relate. Nobody re-types values the source systems already hold, so the wave plan always reflects the estate as it is today. The migration list starts from your application portfolio.
From infrastructure to a sequenced plan
You get what you run as a model, related to the systems of record that hold its hosting, cost and project data.
You can trace the downstream impact of moving or retiring each component before you commit to a wave.
You get a migration order grounded in the real dependency graph, with the rationale captured in the model.
What a migration architect can now get done
Planning a migration is a set of jobs. For each one, the architect states the rules and judges the result.
Building the current-state picture
Today an architect exports inventories from the cloud consoles and the CMDB into spreadsheets, matches them to applications by name, and keys the result into the repository.
AI-augmented, the architect names the source of record for each fact (hosting, owner, run cost, migration project) and states the matching rules. AI connects to the ITSM, ITFM and PPM systems, relates the records to the applications in the model, and lists the ones it could not match.
The skill is knowing which system to believe for which fact, and writing matching rules that reflect how your organization names things.
Tracing what depends on each component
Today architects ask around the application and operations teams and draw a dependency diagram by hand for each candidate.
AI-augmented, the architect asks the model for everything that relies on a component (applications, integrations, data flows, processes) and reads the list with the evidence behind it.
The skill is knowing what a complete answer should contain, and noticing what the model has no way of knowing, such as a dependency nobody ever captured.
Sequencing the waves
Today the wave plan is built in a workshop on a spreadsheet, and the reasons for each placement end up in meeting notes.
AI-augmented, the architect states the sequencing rules: what must move together, what cannot move before its dependencies, and which workloads carry regulatory constraints. AI proposes a wave order from the dependency graph with the reason for each placement. The architect adjusts it and records the decision in the model.
The skill is setting the rules and judging the trade-offs the model cannot see, such as a business blackout period.
Keeping the steering group informed
Today the status slides are rebuilt for every steering meeting and are out of date by the time they are presented.
AI-augmented, the architect starts from the decision the steering group has to make and describes the view that would inform it. AI builds it on the connected data, and it updates as the waves move.
The skill is starting from the decision, and describing the view in terms of that decision.
Which tools help with which job
- Claude + AI Power Tools
Builds the current-state baseline from the ITSM, ITFM and PPM systems, traces dependencies on the live repository and proposes the wave order.
- GitHub Copilot + AI Power Tools
The same baseline and wave-planning work from VS Code, for architects who already work there.
- Kernaro AI Hub
Answers what depends on a component from the live model, read-only for the people who ask.
- Prolaborate 6 with AI
Gives the steering group a wave and status view on the connected data.
- Power BI
The migration dashboards leadership opens. Kernaro AI Hub is the recommended route for Power BI: it reads the model live through a GraphQL endpoint, with no intermediate database.
- Microsoft 365 Copilot
Lets the steering group ask questions of the model in Copilot. It is read-only.
How we help
Train and mentor your architects
Training and mentoring that teach your architects to plan migrations this way on your own repository, then stay with them until it sticks.
For ArchitectsBuild the model for you
Consulting services for organizations that are capacity constrained or need results quickly. We build the connected baseline and the dependency model and hand them over complete and current.
Build the model for youGet your EA data into the tools leaders use
Migration dashboards for the steering group in Power BI or Prolaborate, built on data that stays in the source systems.
Reporting and Power BIDecide where to start, with a plan management will fund
Using the NovoCircle Method, we get leaders and teams to decide where to start, with a plan management will fund and a team ready to deliver it.
Paralysis to a Plan