How to hand over website integrations

For each live integration, write down where it starts, what it should produce, how the new owner can verify the result and what they should do if it fails. Ask the receiving owner to run an agreed safe test before they sign it off.

The Telecoms World rebuild shows why a tool list only gets you part of the way. The website has HubSpot enquiries and a phone-number ordering flow that connects selection and checkout to the backend. The case study documents work with ECCE on the ordering connection and ongoing website maintenance.

Make a row for each outcome someone relies on

Start with the flows whose failure would leave a real customer waiting or the team missing something. Walk each one through from the page or trigger to the system where work actually happens. Record the person who owns that destination and the person who can fix or escalate problems on the website side. Keep them distinct if responsibility is split.

If different forms use the same automation, group them only if their destination, owner and pass check really are the same. A different sales queue, required field or alert recipient is a good reason to keep a separate row. That helps the next person see which business process will be affected when they edit a form or rotate access.

Agree what counts as a pass

Check the result at the destination. For a lead, verify that the receiving team can find the record, that the intended fields arrived and that it landed in the right queue. Agree how to tag and remove the test record. Use a sandbox or an approved test process for anything that might charge money, reserve stock, contact a customer or start fulfilment.

Ask the receiving owner to do this with their own permissions. If they can see the confirmation page but cannot check the destination or access failure logs, there is an access gap to resolve. Record the check date and keep any supporting evidence in a restricted location with personal information removed where possible.

Write the first step for a missing result

For each flow, record who receives the failure alert and where they look first. Say which system to open, what time window to look at and what reference helps find the attempt. Give the owner a way to reconcile website submissions against destination records where the systems allow it.

Agree when it is safe to retry. An automation error can mean the final confirmation was lost after the receiving system already created the record. If the owner cannot verify the result or establish that replay is safe, route it for manual reconciliation. Avoid documenting a blanket instruction to resend everything.

Make access part of the handover

Record which account owns the integration, who can change it and how the receiving team obtains access. Use the team’s approved access or password management process for credentials. Keep tokens, personal data and screenshots with sensitive details out of the register itself.

Before signing off, ask whether the named owner could still run the check and escalate a failure if the person who built it were away. If the answer is no, keep the row open and write down the missing access or decision.

Start with the flows that would hurt to lose

Use the editable integration handover register for the most consequential leads, bookings or orders first. Expand it once the receiving team can actually use it.