In development — early access not open yet

Data Migration AI Toolkit

AI-assisted field mapping and transformation for data migration projects: the tool proposes, a human decides. It turns the weeks of mapping spreadsheet work at the front of every migration into a reviewed, versioned specification.

Status, stated plainly

This is in development. Early access has not opened, there is no trial, no download and no purchase. The waitlist is a list — joining it tells you when there is something to try, and nothing else happens in the meantime. If you have a migration running now, the consulting engagement is the thing that exists today.

The problem

Mapping is the part of a migration that eats the calendar.

It is also the part that carries the most silent risk — a wrong mapping does not fail loudly, it produces plausible data in the wrong column.

What it looks like today

A spreadsheet with a few thousand rows. Someone works through it field by field, guessing at intent from column names, chasing the one person who remembers why a legacy field is called what it is called, and re-doing a chunk of it when the target schema changes in week six.

It is slow, and by the end nobody can reconstruct why any individual decision was made.

Why it goes wrong quietly

A missed mapping fails loudly and gets fixed. A wrong mapping loads successfully, looks reasonable in the target, and surfaces months later inside a report somebody already made a decision on.

The defence against that is not more care. It is making every mapping decision explicit, attributed and reviewable.

How it works

The tool proposes. The human decides. Nothing moves on a proposal.

This is the same human-in-the-loop pattern as every AI workflow EthanCorp builds — for the same reason: the organisation has to stand behind the result.

Data Migration AI Toolkit: proposed mapping, human decision Source and target schemas are profiled, then a proposal step suggests field mappings and transformations with a confidence score and a stated reason for each. Every proposal goes to a human reviewer who accepts, edits or rejects it. Only accepted mappings enter the versioned mapping specification that drives the migration run. Source schema Profiled Target schema Profiled Candidate matching Names · types · samples What the AI does Proposes mapping + transformation States confidence and its reason What the human still decides Accept · edit · reject Every single field Signs off the spec Before anything moves Versioned mapping specification Only accepted mappings — diffable, reviewable, reusable Migration run Executes the signed-off spec, never the raw proposal
Every proposal carries a confidence score and a stated reason. Only accepted mappings enter the versioned specification, and only the signed-off specification drives a migration run.

What the AI does

  • Profiles both schemas — names, types, nullability, sample values, distributions
  • Proposes a field mapping and, where needed, a transformation
  • Attaches a confidence score and a stated reason for the proposal
  • Flags the cases it is least sure about, so review effort goes where it matters

What the human still reviews

  • Every single field — accept, edit or reject, with no bulk auto-approve
  • The transformation logic, not just the source-to-target pair
  • The final specification, which is signed off before any migration run
  • Anything the tool flagged as low confidence, first

Who it is for

People who own the mapping spreadsheet.

Data migration consultants

Running the mapping phase on someone else's estate, against a cutover date that will not move.

Business analysts

Holding the source-to-target spreadsheet that everyone else treats as the specification.

Project managers

Needing to know how much of the mapping is genuinely agreed, not how many rows are filled in.

Solution architects

Reviewing mapping decisions they did not make and cannot easily audit.

Consulting vs. toolkit

Two different things. Worth being clear which you need.

EthanCorp's primary work is consulting: custom enterprise engagements, scoped to your estate, delivered and handed over. That is what exists today and what the case studies describe.

The toolkit is a productised slice of one phase of that work — self-service, aimed at people running their own migration who want the mapping phase to go faster without giving up control of it. It is not a replacement for the engagement, and it is not available yet.

Consulting — available now

Custom integration and migration delivery on your systems, with runbooks and handover. See the service.


Toolkit — in development

Self-service mapping workflow for your own migration project. Early access has not opened.


Indicative pricing

US$99–299, early access, indicative only. The pricing model — one-time or subscription — is not decided, and this range may change before launch.

Want to know when early access opens?

Join the waitlist and you will hear once, when there is something real to try. If your migration is running now, say so — the consulting engagement is the thing that can help this quarter.

Response time
Within two business days
Based in
Ho Chi Minh City, Vietnam — working across Asia and remote