Skip to content

From idea to an approved project

Most organisations are not short of ideas. They are short of decisions — and of somewhere to look up who decided what, when, and on what grounds.

Migreat! from Promigion holds that part: how a request arrives, how it gets assessed, and how a decision lands on it that still means something a year later. What follows is what the product does. What it does not do is set out halfway down.

What happens between an idea and a project

The route from request to project runs in a fixed order:

  1. Someone submits an idea.
  2. It is given a status, and can be typed on two axes.
  3. It picks up three scores, which say what is still missing.
  4. A decision is requested from a named person.
  5. They decide — approve, or reject with a reason. The requester can withdraw the request while it is still undecided.
  6. Only after an approving decision can the idea become a project.

That last step is a lock rather than a working agreement: without an approved decision, converting to a project is not available.

What you actually do with it

Two fixed axes, with your own values in them

An idea can be typed on two axes: the work form — project, change, operations — and the reason for it, such as improvement, innovation or legislation.

Those two axes are set by the product. The values underneath are not: which work forms and reasons your organisation recognises are yours to define. Both axes are optional, and there is deliberately no default — a default would stamp every idea with something nobody chose, and the reporting would then measure your configuration rather than your organisation. Anything untyped lands visibly in its own bucket.

Statuses that match how you actually work

The funnel has five fixed stages, from first sketch through to converted or rejected. What you configure are the statuses you lay over them: you create them, rename them, order them to match your own process, and mark one as the starting point. Each status maps onto one of them. One stage cannot be assigned by hand: converted follows from a project actually being created.

Deleting one is possible as long as it is not built in, not the starting status, and not in use. At least one always stays active. A status keeps its internal code once created — the name is what you can change, so renaming something later does not break its history.

A score that says what is missing

Every idea carries three scores: completeness, likelihood of success, and strategic contribution. The formula behind them is set by the product.

What matters more than the number is what sits under it. Each idea lists the fields still missing — an absent estimate, an unlinked strategic theme — so a low score reads as a short list of work rather than a verdict. There is also an AI assessment of the likelihood of success, available where the AI features are switched on. It sits beside the calculated scores, not in place of them.

A decision with a name attached

An assessment is not an email. You request a decision from a specific person, with a date by which it is due. That person can hand it to someone else, and that stays visible.

As the date approaches, a reminder goes out automatically — three days ahead by default. If the date passes anyway, it is escalated to the people accountable for the portfolio, not counting the assignee. That pass runs once a night and decides nothing itself; it only makes sure nobody can say they did not know.

At most one decision is open per idea at a time. A rejection cannot be recorded without a reason. And a decision stands as it was taken — you can request a new one, but you cannot rewrite the old one.

What Migreat! does not do

  • No benefits management. You record an expected value as free text, and nothing is steered or settled against it.
  • No probability-weighted forecast. An idea counts in a projection at one hundred per cent, or it does not count.
  • No scoring model you configure. The weighting is set by the product, and what a client submits is overwritten server-side.
  • No export from the idea funnel report. The other reports produce PDF; this one does not.
  • No portal for submitters or outside stakeholders. Whoever takes a decision is a user inside your own organisation.
  • No chatbot and no copilot. The AI features advise: they decide nothing and act on nothing.
  • No public customer API and no webhooks.

Which plan holds what

Ideas, the idea funnel, the classification and the decision workflow are part of the Enterprise plan — administering them, that is. The labels that come out of them are visible to anyone who can sign in.

Projects, timelines, dependencies and project health are in every plan.

Current prices are on the pricing page.

Further reading

How this connects to steering the portfolio as a whole is on the portfolio management software page, and every part is listed on what Migreat! does. For the discipline itself, separate from this product: what project portfolio management is. Why a decision so often stays informal in practice, and what a chain from theme to delivery actually needs, is in the guide from strategy to delivery. Why the distinction between a score and a decision matters once you have to account for which requests actually went ahead is covered in the guide prioritising projects. That AI here always advises and never acts on its own is the subject of the guide AI in portfolio management.