Recruitment agencies pay a hidden tax: fragmented software stacks

Nearly half of organisations in talent acquisition still manage disparate systems, while only 43 per cent rely on integrated suites, according to Aptitude Research's 2025 CRM Index. That single statistic frames a quiet cost problem now hitting recruitment agencies: they run one system for almost every step of the job, and the people doing the job are paying for it in lost hours and bad data.

The argument comes from the founder of Klose Recruit, an Australian company built, in his words, by people who have done the recruiting job. Writing from the recruitment desk, he describes how small and mid-market agencies have accumulated a separate tool for applicants, another for client relationships, a spreadsheet for fees and forecasts, a separate outreach product, another for reporting, and a growing pile of AI add-ons that do not share a view of the business. Each tool can be justified on its own, he notes. Together they create a tax that rarely appears on a profit and loss statement.

The stakes are economic as much as technological. Recruitment margins are tight, and consultants are the most expensive resource in the firm. When software forces them into manual integration work, the cost lands directly on the bottom line.

Why does tool sprawl cost agencies so much?

The mechanism is simple. When the applicant tracking system (ATS), the CRM and the financial picture live in different places, the consultant becomes the integration layer. The same person is expected to develop clients, run a pipeline, maintain a candidate community, and judge in the moment whether a placement will be profitable. The founder's verdict is blunt: "The most expensive person in the room is doing work a connected system should already have done."

The economic consequences break down into three patterns he identifies:

• Data quality declines. A client record updated in one system is stale in another. A candidate's status can be current in the ATS and wrong in the CRM. Forecasts built on that foundation, he writes, "are a matter of faith."
• Reporting arrives too late to be useful. Owners and directors often learn what the business actually earned after the month closes, once someone has stitched exports together. Live revenue, margin and pipeline data should sit next to the work, not in a weekend spreadsheet.
• New tools stop paying for themselves. An AI writing assistant that cannot see the pipeline, or a scheduling tool that does not know the week's priorities, adds another login without changing the underlying economics.

The last point is where the AI angle bites. The promise of automation is real, the founder argues, but the practice of bolting another application onto a fragmented stack leaves agencies busier, not faster.

The takeaway: fragmented stacks don't just cost licence fees — they corrupt the data every business decision depends on.

Is this unique to recruitment?

No, and the source is careful to say so. The Aptitude Research 2025 findings cover talent acquisition broadly, and the pattern will be familiar to any CIO who has inherited a stack that grew by exception: a new gap, a new vendor, another integration to maintain.

But agencies feel the cost more sharply because the firm is simultaneously a sales organisation and an operational one. The dual role means every disconnected system touches both revenue generation and delivery. For agencies, tool sprawl is a compounded tax, not a flat one.

What is agentic AI, in plain terms?

The founder calls agentic AI "the part of this shift that is easiest to oversell and most useful when it is described plainly." His definition is concrete: software that can take a sequence of actions, not just generate text.

On a recruitment desk, that can look like:

• Planning the day around the roles that will actually move
• Drafting the next email with the full context of the relationship
• Booking a meeting without detouring into a separate calendar product

Used well, he argues, agentic AI gives time back to the work that still requires a human being: judgement about fit, honesty with a client, and the trust that closes a placement. The framing matters because it positions AI as coordination-removal, not output-generation — automation that deletes busywork rather than adds a new surface to watch.

In short: agentic AI works when it acts on the same picture the consultant sees, not on a context-free prompt.

Why is the market structure changing now?

The shift, according to the source, is in the shape of the system rather than the feature list. The emerging operating model puts applicant tracking, client relationship management, revenue and financial insight, and agentic AI in one environment.

That combination sounds obvious. For a long time it was not available to small and mid-market agencies at a price they could reach. The choice was to stitch together several specialist tools, or to buy an enterprise platform designed overseas for a much larger firm. The founder credits the change to systems designed for this market segment — which is where his own company enters the story.

He and his co-founder Zach, the company's Commercial Director, started Klose Recruit because, as he puts it, "we were tired of watching good recruiters lose hours to tools that did not talk to each other." The company is built in Australia for the agencies that make up the backbone of hiring in that market. His stated lesson: "The advantage is not more software. It is fewer, better-connected systems, with insight and AI in the same place as the work."

The disclosure is relevant context: this is a vendor's analysis of a market problem his product addresses. The Aptitude Research data, however, stands independent of that interest.

What should buyers and IT leaders watch?

The founder proposes a simple test for agency principals and the technology leaders who advise them — three questions that expose whether a stack is unified or fragmented:

• Can a consultant start the day in one place and see the pipeline, the relationships that need attention, and the revenue implications of their decisions?
• Can an AI assistant act on that same picture, rather than on a prompt with no context?
• When something changes — a candidate withdraws, a client delays, a fee gets renegotiated — does every relevant view update, or does someone have to go and tell four other systems?

If the answer is the latter, his conclusion is firm: the firm does not have a productivity problem. It has an architecture problem.

The point extends to the businesses that hire through agencies. A hiring partner drowning in systems shows it as slower shortlists, weaker feedback, and less visibility on where a brief actually stands. The agencies that will serve clients better in the next few years, he writes, "will not be the ones with the longest software list," but the ones whose technology gives the consultant a single, current view of candidates, clients and commercial reality — and uses automation to remove coordination rather than to add theatre.

The economics of recruitment, the founder argues, are too tight and the work too human to keep asking consultants to be the middleware. His closing bet: the agencies that take integration seriously will see the payoff in time, in margin, and in the quality of service they deliver to the businesses that rely on them.