The hidden cost of your HR tech stack (even when it works fine)

Ask most HR teams how their system is doing and you'll hear the same answer: "it works fine." Payroll runs, people can book holiday, the reports come out eventually.

I think of it like a restaurant. Out front, the room looks lovely, the staff are friendly and the food arrives, a little late but it arrives. Open the kitchen door and it's a different story: people running around, three cooks working from three different recipes, and the only person who knows them all is on leave. That's how many HR teams operate, and it's nobody's fault. When you're busy keeping the business running, there's rarely time to zoom out.

On 27 October I joined Shane Hudson and Federico D'Orsi at Factorial for a webinar on exactly this: The Hidden Cost of Your HR Tech Stack. Here's a recap for anyone who missed it, and a refresher for anyone who didn't.

Where the time goes: what the room told us

We started with a quick poll: which task takes up the most time in your team? The top answers were:

• Updating employee data across several systems
• Answering repeat questions from employees and managers
• Chasing approvals
• onboarding and offboarding

These answers matched what I see on transformation projects. Take an integration between an HR system and payroll. It might create the employee profile and send part of the record across, but not timesheets or payslips. The integration "works", yet someone still has to move the rest by hand.

Repeat questions are just as telling. Questions about policies, entitlement and payslips usually mean the data exists but isn't reachable by the people who need it. These are symptoms of the work happening behind the scenes. You can have a good system, even some AI on top of it, and still need people to carry a process from start to finish.

Four places the work hides

When I run vendor selection and implementation projects, the business case often starts with replacing the system. Once we review the current state, the gaps usually fall into four groups:

• Process gaps. Are the steps repeatable, clear and owned? Is everyone using the same tool for the same request, or is one person using the system while another uses a Slack or Teams message?
• Integration gaps. Does data move securely and in both directions between systems, or are people acting as the integration?
• Platform gaps. Can the platform meet your compliance needs and grow with you? Do you have confidence in the roadmap and the support?
• Visibility gaps. Before a board meeting or business review, does someone glue together data from several spreadsheets, only to find another team has sent different numbers?

Replacing the system is often the tip of the iceberg. Scratch the surface and there is usually much more than a tool gap underneath.

What the workarounds really cost

The licence fee is easy to see. The cost of the workarounds is not, and it lands in three places:

• Time and manpower. Skilled people spend their week on repeatable tasks, such as moving data between systems and Excel, instead of strategic work or advising the business.
• Risk. Once data is exported and copied around, it's harder to show it's handled securely and in a compliant way.
• Credibility. If HR can't answer a simple question quickly, employees and managers stop asking.

Put a number on it

You don't need a big review to start. Pick one task that feels painful, such as job requisition approval, offer approval or a monthly dashboard, and work out three things:

1. How often it happens
2. How many people are involved
3. Roughly how long it takes each time

Multiply them and you have the weekly cost of that workaround. That's the hidden cost, in your own numbers.

Shane asked whether there's a ratio to compare against licence cost. There isn't a magic one. I'd weigh value against effort: the impact of the workaround on one side, and the time to build or improve the integration or automation on the other. In my experience, tackling the right workarounds can cut the time spent on them by roughly 20 to 30 per cent. That depends on change management and training, because people have to adapt for the new way to stick.

Four questions, in this order

The framework I shared works as a sequence, and the order matters:

1. Optimise. Is the process clear, owned and transferable? If you automate a broken process, you simply get a faster broken process.
2. Integrate. Do you have the APIs or exports to connect your systems securely, so nobody re-keys data? AI-assisted options such as MCP servers (a standard way of letting AI tools connect to other systems) are now part of this conversation.
3. Automate. Are there repeatable steps that run on clear "if this, then that" rules? Those are great candidates, as long as you can audit them.
4. Replace. Is the platform itself the gap? Sometimes it is, but replacement is a big project with its own process mapping, change management and adoption work.

Most people jump straight to question four. If you haven't answered the first three, a new system can leave you in exactly the same place, just with a different interface. Don't let anyone herd you into a new platform before you've checked.

Where AI fits

Federico then showed what this can look like inside Factorial. The AI agent works within the platform and respects permissions, so it only shares what a given user is allowed to see. He also showed it connecting to other tools in a single conversation, including a deck built for sales training using Notion and Slack, a monthly clock-in report sent to his inbox, and a PDF of next week's shifts.

The self-service use case is the one I'd highlight. Upload your policies and let employees and managers ask questions in chat or in the app, instead of emailing HR for the third time that week.

There's a catch, which we also discussed. AI only works well on data you can trust, so timesheets need to be complete and adoption needs to be high. If your policies or processes are weak, the AI will inherit the problem. It's another reason to start with question one.

What to do this week

• Map the high-level steps and note where data is copied, delayed or chased through other channels.
• Pick one task that creates repeat effort, such as approvals, onboarding routing or data updates across systems.
• Calculate frequency, people and time to get the weekly cost.
• Choose the shortest route through the framework. Optimise, integrate or automate before you consider replacing.

Federico summed it up well: "Start small, get confident with these tools and then transform." I'd add one thing. Begin with the high-value, low-effort wins, and let the results build trust.

Thank you to Shane, Federico and the Factorial team for hosting us. If you'd like an independent view on where your own workarounds sit, or help working through the framework and the numbers, get in touch with Udder. If you'd like to see Factorial's product in action, their team will be glad to walk you through it.