The options, side by side.
Most comparisons online are written by whoever wins the comparison. Here are all four paths with their strengths, and Empley is not the right answer on every row.
Comparisons
We put the options side by side, not next to a strawman.
Comparisons
Four ways to build AI capability in HR
The same seven questions asked of all four. The numbers are typical ranges for companies with a hundred to a few thousand employees, not guarantees.
| Consultancy on retainer | Internal team | HR tech platform | Empley | |
|---|---|---|---|---|
| Time to first result | Two to six weeks for a recommendation | Four to nine months including hiring | Days to switch on, longer to get useful | Four to six weeks to a first agent in production |
| What you pay for | Hours and deliverables | Salary, employer costs and ramp-up | Seats and modules, licensed per year | A working data foundation plus agents in operation |
| Capacity when scope grows | More hours, higher cost | New hires, more headcount | Same platform, work still has to be configured | Same foundation, more agents on top |
| Who owns the work | The consultancy owns the method and often the model | You own everything, including the risk | The vendor owns the platform and the roadmap | You own the data, the models and the environment |
| Auditability | Varies, usually a report after the fact | Depends on who built it and whether they documented it | Logs exist, but rarely at the decision level | Logging per run from day one, built for the AI Act |
| Biggest risk | Dependency without a real handover | Key person leaves, knowledge leaves with them | Shelfware nobody configures or adopts | Wrong first process picked costs you a quarter |
| Fits when | You need judgment and outside perspective for a defined period | The need is broad, permanent and worth a full team | You already have the data and people to run it yourselves | The work is repeatable, HR data needs a foundation, and you want it running, not just advised on |
When not to choose us
If you need a permanent team with judgment embedded in the business every day, hire. If the work is different every time, there is no repeatable process to run as agents. And if nobody internally can define what a good outcome looks like, that is the problem to solve first.
Comparisons in detail
Skills taxonomy versus job architecture
How a skills taxonomy and a job architecture differ, and why most organisations need both, in order.
ComparisonWorkforce planning6 minStrategic versus operational workforce planning
A comparison of strategic and operational workforce planning, and why organisations need both running at once.
Not sure which path fits? We'll tell you if you shouldn't hire us.
A few lines about your situation is enough for us to answer.
What an agent in production costs and what's included.
Guides and analysis that go deeper on each path.