Inside the HR Builder Revolution: Why 91% of global teams are building, not buying
Ask HR teams where their software came from and the answer has almost always been the same: they bought it.
A vendor decided what the product could do. The customer made that process work for their needs. Remote's latest survey suggests that dynamic is becoming a thing of the past.
In the past 12 months, 91% of HR leaders worldwide said their team has built a custom tool or automation for at least one need instead of buying software. Nearly half, 46%, now treat building as a deliberate strategy rather than a one time fix for a missing feature. And they are not slowing down: 90% plan to build more over the next two years, 42% of them significantly.
What they are building tells the real story. HR teams are not just speeding up routine work like scheduling interviews and pulling together status updates. They are building the workflows their software never offered, shaped to fit the way their company actually runs.
The pattern holds across regions and company sizes, though how fast it is moving, and how far teams are taking it, differ sharply from one place to the next.
Custom HR software went mainstream before most companies noticed
Custom building has spread because so little of HR work is standard. Performance management differs from company to company. Employee relations workflows depend on local law, internal policy, manager expectations, and culture. Workforce planning questions change year to year, compliance processes vary by jurisdiction, and reporting needs rarely match a vendor's dashboard. For a long time HR teams accepted those constraints as the price of using software.
AI changes the math. Getting started with a tool no longer needs a procurement cycle, an implementation project, or an engineering backlog. It can begin with someone who understands the process describing what they need and building a prototype themselves. That does not turn HR into an engineering team. It simply puts more of the system in the hands of the people running the process.
The payoff is that a custom workflow can reflect how the company actually operates: the language managers use, the approval paths Legal expects, the local rules People Ops has to check, the metrics leadership wants, and the risk thresholds the company will not cross.
It helps too that most HR processes pull from data held in different places: the HRIS, payroll, policy pages, case history in an email chain, leadership questions raised in meetings rather than tickets. AI gathers that scattered context, drafts from it, and handles the repetitive assembly, leaving the person to apply judgment where it counts.
That makes the best early targets the recurring tasks that draw on context from several places at once: looking up country-specific employment guidance before advising a manager, drafting versions of the same communication for different audiences, summarizing employee relations case notes, building a performance tracker, or compiling policy answers from internal documents. Each one takes the manual assembly out of the work while leaving the judgment to HR.
Personal spending shows adoption is outrunning the rollout
Another big finding from the survey: 59% of respondents have spent their own money on AI tools and subscriptions, most of them between $25 and $100 a month.
For HR leaders that cuts both ways. On the upside, the curiosity is already in the building. Employees are trying things out without waiting to be told to.
The risk is fragmentation. When people rely on personal accounts and disconnected tools, the company loses data security, visibility into data usage, quality, and measurement. One team's helpful workflow can become another team's compliance nightmare.
The response is to make safe experimentation easier than the unsafe kind, with clear guidance on what data can be used, which tools are approved, where human review is required, and how workflows should connect to trusted company systems.
The building is not slowing down
The survey also exposes a gap between confidence and proof. Ninety one percent of respondents say AI has helped their business goals, but only 62% have data that shows it.
Belief is enough while AI adoption is cheap and experimental. It stops being enough the minute the CFO wants a return on the spend, the auditor wants proof the governance held, or the regulator wants to know what data those tools touched.
As custom tools get embedded in everyday workflows, companies will need to measure their impact more meaningfully. Some benefits are simple to count, like time saved, fewer manual steps, and shorter response times. Others are just as real but harder to put a number on, such as steadier manager confidence, more consistent documentation, and better decisions when the right context is finally easy to find.
The teams that get ahead of this will measure at the workflow level rather than the program level, tracking the specific work a tool changed instead of the rollout as a whole. Useful signals include:
Time saved on recurring administrative work
Faster responses to manager and employee questions
Fewer handoffs between HR, Legal, Finance, and Operations
Less reliance on spreadsheets for repeatable processes
The test is whether the workflow got better.
Payroll is where teams move most carefully
The gap between general AI use and payroll AI use is wide. 76% of respondents use AI somewhere in the organization, but only 21% use it for payroll workflows.
That delta reflects deliberate care rather than reluctance. Payroll is critical infrastructure. It touches tax, statutory benefits, employment law, filings, local deadlines, currency, and the basic expectation that people are paid correctly and on time. Because so much depends on getting it right, teams are introducing AI here with more validation, review, and testing than they apply elsewhere.
Done well, AI in payroll rests on the same foundation as everything else: accurate data, correct permissions, and reliable systems. At many companies, teams are establishing confidence in lower stakes areas first, then carrying that experience into the processes where the cost of an error is highest.
Compliance and integration are the real blockers
When leaders name what is holding their AI efforts back, two answers dominate. Thirty three percent of respondents name regulatory compliance concerns as a top obstacle to adoption, and 29% point to poor integration with existing systems.
Both barriers come down to the nature of HR data. Employee, pay, and legal information is too sensitive to hand to a tool a team does not fully trust, since the cost of getting it wrong is so high. And because that data is scattered across systems that were never built to connect, a tool's usefulness is capped, since nobody will rely on one that cannot dependably get the full picture. Until both challenges are solved, most AI stays at the edges of the real work.
This is the gap that a newer class of tools is built to close. Remote MCP gives AI agents a secure, permissioned connection to employment data such as payroll, contracts, compliance information, and org structure, without exports, API keys, or custom engineering. Remote Agent works on top of that same trusted foundation, handling employment tasks with the data and rules of global employment already built in. Handled this way, compliance and integration stop being blockers and start becoming the reason the tools can be trusted with real work.
How different countries compare
The survey tracked two behaviors by market: whether teams built custom tools instead of buying, and whether individuals spent their own money on AI. Building was high across the board, ranging from 74% in the Netherlands to 97% in Australia. Personal spending was more uneven, running from 34% in France to 72% in Australia.
Here is how the eight markets stack up:
Market
Built custom tools
Spent personal money on AI tools for work
Australia
97%
72%
France
90%
34%
Germany
95%
71%
Netherlands
74%
52%
Singapore
96%
69%
Spain
87%
50%
UK
94%
60%
US
93%
69%
Global average
91%
59%
Company size makes far less difference than geography. The 91% building rate holds across every band surveyed, from 500 to 749 employees up through 1,000 and above, so this is not a shift confined to the largest enterprises. Smaller teams use it to solve a specific workflow problem without a long implementation, while bigger ones adapt workflows across existing systems and regions, but the impulse to take control of the tools is the same.
What the shift looks like inside one People team
The survey captures the macro trend. Remote's People team shows what it looks like up close.
In a recent conversation about building HR tools with AI, Remote Chief People Officer Barbara Matthews described starting the year anxious about it. Despite a career in Tech, she did not think of herself as especially technical, and her first move was not a formal transformation program. She experimented with low stakes personal projects until AI felt approachable, and that changed her sense of what was possible at work.
From there, her People team began building custom tools and workflows in plain language, including systems for their company’s specific performance and case management needs. Buying the equivalent off the shelf could have cost €100,000, Matthews noted, and still would not have matched how the company actually works.
That logic is why building is becoming the dominant strategy: the people who run a process know exactly what the tool needs to do, and AI finally gives them a way to build it without waiting on engineering or a vendor.
Why building it yourself is the real advantage
AI tends to move HR work up a level. When it drafts the first version of a policy summary, the HR professional spends the freed time on nuance, local context, tone, and risk. When it builds a workforce analytics view, the leader spends less time making the chart and more time deciding what to do about it. The result is a more strategic version of the same function.
But that payoff only lands if HR builds the capability itself. When adoption is owned entirely by IT, Legal, or outside vendors, HR inherits tools designed around someone else's read of the work. The teams engaging directly are the ones learning how a people function should operate in an AI native environment, and building the judgment, governance, and operational muscle to do it, while slower organizations are still debating whether HR should experiment at all.
The takeaway is straightforward: start building. Pick one repetitive, context heavy workflow, automate a piece of it, and learn from what happens before moving to the next. If the confidence is not there yet, the training is: Remote's free AI training course is a place to start.
The question is no longer whether HR will build, but how much it will reshape the work.
In May 2026, Remote surveyed 3,250 People leaders across the UK, the US, Spain, Singapore, the Netherlands, Germany, France, and Australia. More than 90% had already built their own AI tool or automation rather than buy software for the job. This is what it looks like when the people who run HR stop waiting for their systems to catch up.
Subscribe for the latest updates
Sign up for our newsletter to get the inside scoop on all things remote work and global employment.
