The Three Places HR Consultants Still Type the Same Thing Twice
Drafting with ChatGPT was never the expensive part of HR consulting. The same data typed into three systems that do not talk to each other is. Here is where it hides.
For independent HR consultants, the costly AI gap is not drafting letters, it is the same employee data re-keyed into systems that do not talk to each other. It hides in three places: onboarding a new starter, policy and contract versions, and records both you and the client keep. Fixing it means one source of truth that the other systems read from, which also shrinks your data protection exposure. Start with the task that annoys you most and map where the time goes before changing any system.
Most independent HR consultants did not get into this work to spend their afternoons copying a new starter’s details from an offer letter into one system, then a second, then a third. Yet that is where a real slice of the week quietly goes. The writing, the advice, the difficult conversation with a client about a tricky exit, that is the work you trained for. The re-keying is not.
Plenty of consultants have started using a tool like ChatGPT to draft a letter or tidy a policy, got a decent result, and decided they are now doing AI. The drafting was never the expensive bit. The expensive bit is the same data being typed into two or three systems that do not talk to each other, and no chatbot has touched that. There are three usual places it hides.
Place One: Onboarding a New Starter
A new employee joins one of your clients. Their details go into the HR platform. Then into payroll. Then into the pension portal. Then onto a contract template, and maybe a right-to-work record after that. Same name, same dates, same national insurance number, keyed in four or five times.
Every one of those re-keys is a chance to fat-finger a start date or transpose two digits of a salary. And when one system says one thing and another says something else, you are the one who has to work out which is right.
The fix is simple to describe. Pick one system as the source of truth for a new starter. Enter the details once, there. Let the other systems be populated from it automatically, rather than by hand. Enter it once, use it everywhere.
Place Two: Policies and Contracts That Live in More Than One Place
A policy you maintain for a client tends to exist in at least three forms at once. Your master version. The copy the client keeps on their own drive. The email thread where you sent them the latest update. When that client asks whether they are covered for a recent change in the rules, you cannot always say with confidence which version they are working from.
That gap is version drift. It is the quiet reason a client ends up acting on a policy you revised two updates ago.
The fix is to keep one live version that everyone reads from, generate the client-facing copy from that single source, and keep a simple record of what was sent and when.
Place Three: The Records You and Your Client Both Keep
Holiday balances. Sickness. Timesheets. You update them in your system. The client updates them in theirs. By the end of the quarter the two disagree, and someone spends an afternoon reconciling numbers that should never have diverged.
Both sides are diligent. Both are entering the same facts twice. The way out is to decide, per record, who owns it, then let the other side read from that record rather than re-enter it.
Where the Data Protection Risk Actually Sits
There is a reason to care about all this beyond the lost time. Every time an employee’s details are copied into another system, you have created one more place that data can leak, go stale, or sit somewhere it should not. Re-keying does not just cost minutes. It widens the area you have to protect.
That cuts both ways, and in your favour. If you do bring in automation or AI to remove the double-handling, done properly it leaves the data in fewer places, not more. Strip out any field the tool does not need to see. Use the enterprise tier of whichever AI provider you choose, the one that does not train on your data and will sign a data processing agreement. And keep the data in the UK or EU, or on systems you control.
Before you change a single system, list every place a new starter's details get typed. The length of that list is usually the whole argument.
Where to Start
Do not try to fix all three at once. Pick the one that irritates you most on a Monday morning and start there.
The re-keying rarely gets named, because it feels inevitable. It is too small to bother auditing and too spread across the week to notice. But three of these loops, at a quarter of an hour each, is close to two hours a week. That is the gap between the work you trained for and the work that quietly pays for your software.
If you would rather not start from a blank page, I have turned the five usual places into a one-page checklist. Drop me an email and I will send it over.
And if you want to know what fixing it is actually worth, that is what the AI Opportunity Snapshot is for. Two hours, remote. I ask the questions, so you do not need to know what to ask. Within two working days you get a short written map of your three biggest AI and automation opportunities, what each one is worth, and where to start. It is £500, all-in. Drop me an email and I will send the booking details.
- The costly AI gap for HR consultants is not drafting, it is the same data keyed into systems that do not talk to each other.
- Double-handling hides in three places: onboarding a new starter, policy and contract versions, and records both you and the client keep.
- The fix in every case is one source of truth that the other systems read from, rather than re-entering by hand.
- Done properly, removing double-handling shrinks your data protection exposure, because the data lives in fewer places.
Curious how this could work for your business?
Take the 2-minute assessment, or send me an email. I'll come back with something useful, not a sales pitch.
2-minute assessment