iPaaS readiness assessment

Is there a current inventory of every interface in the estate?

Current means it matches what is deployed, not what was deployed when it was written.

Does every interface have a named owner?

A person, not a team mailbox that forwards to a distribution list.

Do you know which interfaces are still actually used?

Most estates carry interfaces that have not moved a message in a year.

Are message volumes and peaks known per interface?

Peak matters more than average — platforms are sized and priced on peak.

Is the data crossing each interface classified?

Personal data, payment data, and anything with a statutory retention period.

Is there a network path from the new platform to the systems it must reach?

Including the on-premises ones behind a firewall nobody wants to touch.

Is there a firewall change process that runs in days rather than weeks?

Measured by how long the last one actually took, not the SLA.

Do trading partners allow-list your IPs, and do you know which ones?

Moving platform usually changes egress addresses.

If on-premises systems are involved, is a gateway or agent approach agreed?

Most iPaaS reaches on-prem through an agent that someone has to host and patch.

Is it clear who approves a new integration going live?

The named function, and what evidence they require.

Is there a secrets management approach for credentials and keys?

Not a configuration file, and not a spreadsheet.

Are data residency constraints known and documented?

Which data may not leave which jurisdiction, and who decided.

Do you have an inventory of certificates and keys used by interfaces?

Expiry dates, owners, and which partner each belongs to.

Is there an agreed audit and retention requirement for message data?

How long payloads are kept, and who can read them.

Are integrations held in source control today?

Actually in it, with the deployed version traceable to a commit.

Is there an automated build and deployment pipeline?

Or is promotion still a person with a checklist and a maintenance window?

Are there non-production environments that resemble production?

Same integration points, comparable data shape, not just the same platform.

Is representative test data available?

Including the awkward records — the partner with the 40-character name, the negative quantity.

Is there a regression suite that would catch a behaviour change?

Something that compares old and new output for the same input.

Does the team have hands-on experience of the target platform?

Delivered on it, not trained on it.

Is there monitoring that would tell you an interface has stopped?

Before the business tells you.

Do alerts reach a rota rather than a mailbox?

Someone accountable at 3am on a Sunday.

Are there runbooks for the interfaces that matter?

Enough for someone who did not build it to recover it.

Are SLAs defined per interface, or at least per tier?

What 'down' means, and how long is tolerable.

Is it agreed who supports the new platform once it is live?

Including out of hours, and including the first six months.

Is there a business sponsor who will still be there in a year?

Migrations outlast reorganisations more often than sponsors do.

Is budget secured beyond the first year, including dual running?

Two platforms run in parallel for months. Both are paid for.

Is there genuine appetite to decommission the old platform?

Someone must own switching it off and absorb the argument that follows.

Is there a named product owner for the integration platform itself?

Someone whose job is the platform, not a project that ends.

Are business freeze periods known and built into the plan?

Year end, peak season, regulatory reporting windows.