A migration moves the form without answering the question the form was asking. See the three patterns behind poor intake adoption, and how to tell from your own data which one you have.
What usually gets blamed, and why it’s wrong
The project ran properly. Eighteen months, a real budget, a steering committee. The new system went live, and adoption is worse than the thing it replaced.
So you book more enablement sessions, set up a champions network, send another comms round. Adoption ticks up for a fortnight, then settles back exactly where it was.
Sound familiar? None of those things changed the work. A new system moves where the work happens. It doesn’t change how much of it there is.
A migration moves the form. It doesn't answer the question the form was asking.
The work your system still asks people to do
Look at what somebody has to produce before a requisition is complete. Then look at what they had to produce in the old system. Spot the difference?
Pick a category
From a taxonomy you wrote for reporting, not for the person buying. They're guessing, and you know they're guessing.
Find a commodity or GL code
Which means knowing a coding scheme they see maybe four times a year.
Choose a supplier
Without being able to see who's preferred, who's contracted, or who's even set up in your system.
Attach the right documents
A quote in the right format, from the right contact, with the right terms on it.
Answer your policy questions
About thresholds and approvals that exist to protect the company, and that they've got no way to reason about.
There isn’t one. The fields look different and the buttons moved, but their job didn’t change, so neither did what they do about it.
The three patterns behind poor adoption
- They route around you.
Somebody needs it this week, got it wrong last time, and buys it on a card or asks the colleague who knows the shortcut. The spend still happens. It just happens outside the system you spent eighteen months implementing.
- They submit and wait for you to fix it.
They fill it in as best they can and let your team correct it. Adoption looks healthy in the dashboard while the correction work moves quietly onto you.
- They give up halfway.
The request gets started, hits a field nobody can answer, and dies there. You'll never see these in an adoption report, because an abandoned request isn't a request.
The second one is the one that fools people. Submission rates recover, the team absorbs the gap, and the project gets written up as a success while cycle times quietly get worse.
What actually changes adoption
Adoption changes when the request stops being a form to complete and becomes a sentence to say.
Your colleague describes what they need in plain language, in Microsoft Teams or in the system you just migrated to. Pactum’s Intake Agent does the rest.
It works out the coding
Category, commodity code, and GL code come from your own data, not from somebody's best guess at 4pm on a Friday.
It finds the right supplier
From your preferred list, your contracts, and your past orders for the same thing. Rather than asking somebody who can see none of that.
It only asks what it genuinely needs
And it asks in the conversation, in plain words, instead of as a required field with no explanation next to it.
It arrives complete
The Alignment Agent checks the finished requisition against your policies, your contracts, and your approved suppliers before money is committed. So it doesn't come back to you a week later.
If you want the difference between this and simply building a better form, Intake Agent vs. Intake Form goes through it.
If the capture step is still largely manual behind the new system, what to automate first in a manual intake process covers the step to fix before adoption is worth measuring.
How to tell which pattern you’ve got
Three questions, all answerable from data you already hold.
- Compare card and off-contract spend before and after go-live. If it went up, people are routing around you.
- Count corrections your team makes per requisition. If that went up, you’re absorbing the gap and your adoption numbers are flattering you.
- Count started and abandoned requests. Most systems log this and almost nobody looks. A high abandon rate on one specific field tells you exactly where requests are dying.
What to do next, in order
- Stop adding training. If three rounds haven’t fixed it, the fourth won’t either.
- Find the field that’s killing requests. Your abandon data will tell you in an afternoon.
- Remove the question instead of redesigning it. Anything an agent can resolve from your own data should never reach your colleague in the first place.
- Leave the approvals exactly as they are. The adoption problem is in the capture step, not in governance. Pactum works inside the workflows and approval paths you already have in Coupa and SAP Ariba.
Adoption usually gets tested hardest in your busiest weeks, so it is worth reading how agents absorb a volume peak next. For what each agent does across the whole intake process, start with the guide to intake agents in procurement.