Insights
We Were Sending Too Many CVs. We Knew It. Here's How We Fixed It.
There is an uncomfortable truth in the recruitment agency business that most people do not say out loud.
Sometimes you send more CVs than you should. Not because you are trying to waste your client's time — but because the pressure to show activity is real, the volume is high, and manually filtering every submission to a precise standard takes more time than the business allows.
At TechTalent, a licensed recruitment agency in Malaysia specialising in technical and non-technical permanent and contract placements, this was the honest reality. Clients were pushing back. CVs did not always match the job. Hiring managers were receiving stacks of submissions and asking for something more useful — not just a CV, but a summary of why a candidate was worth their time.
The team knew the problem. What they needed was a way to fix it without adding headcount or slowing down delivery.
The Real Cost of CV Mismatch
When a recruitment agency sends CVs that do not match the role, the cost is not just one rejected shortlist. It is the hiring manager's trust. Every mismatched CV is a signal that the agency does not fully understand the requirement — or worse, that they are prioritising volume over quality.
For TechTalent, this was showing up in client feedback. Hiring managers wanted fewer CVs, not more. They wanted candidates who genuinely met the criteria. And they wanted a clear summary of each candidate's strengths — not a raw CV dump that required them to do the analysis themselves.
That last point is often overlooked in discussions about AI recruitment tools. Scoring a CV is one thing. Articulating why a candidate is suitable — in plain language a hiring manager can act on quickly — is another thing entirely.
What Changed With OPAL
Three months ago, TechTalent implemented OPAL — an AI CV scoring platform built for enterprise recruitment teams — across their placement workflow for both technical and non-technical roles.
The change was immediate and measurable.
Every CV submitted by a candidate is now scored against a role-specific rubric before any human recruiter spends time on it. Hard compliance gates filter out candidates who do not meet non-negotiable criteria. Weighted must-have criteria are then evaluated in depth — using contextual inferencing and ontology-based synonym mapping, not just keyword matching. A candidate who briefly mentions a skill in passing does not score the same as someone with hands-on experience. The system understands the difference.
But the feature that changed the client relationship most was the candidate summary.
For every CV that passes the scoring threshold, OPAL generates a structured summary of the candidate's strengths — mapped directly to the role criteria. Hiring managers no longer receive a stack of CVs and a covering email. They receive a shortlist with a clear, consistent explanation of why each candidate made it through.
The results after three months:
- Recruiter time saved on manual CV screening: 70%
- CV quality improvement reported by clients: 85%
- Client satisfaction improvement: 3x
"Our clients expect quality shortlists, not volume. Since implementing OPAL, the feedback from hiring managers has been completely different — they trust what we send them now. The candidate summaries make a real difference. Instead of asking why we included someone, they are asking when they can interview them." — Datin Malliga, Managing Director, TechTalent
Why This Matters for Recruitment Agencies Specifically
The enterprise buyer story is well understood — a large company with multiple vendors needs a quality gate before CVs reach the hiring manager. But the recruitment agency story is different and equally important.
An agency's entire value proposition is the quality of its shortlist. If a client can get the same volume of CVs from a job portal for a fraction of the cost, the agency has to justify its fee with something better — better matching, better insight, better speed to quality candidate.
OPAL gives TechTalent that edge. Not by replacing the recruiter's judgment, but by giving the recruiter a consistent, defensible scoring framework to work from — and a structured summary to present to the client.
The recruiter still makes the relationship call. They still manage the candidate experience. They still negotiate the offer. OPAL handles the part that was eating their time and undermining their credibility — the CV screening and qualification step that requires technical depth the recruiter may not always have for every role they fill.
The Summary Problem Nobody Talks About
Most discussions about AI in recruitment focus on speed — how many CVs can you process, how fast can you shortlist. The TechTalent story surfaces something more nuanced.
Hiring managers do not just want fewer CVs. They want to understand quickly why a candidate is worth their time. A well-structured candidate summary — mapped to the specific criteria of the role — turns a CV review from a reading exercise into a decision exercise.
That is a fundamentally different experience for the hiring manager. And it is a fundamentally different service proposition for the recruitment agency delivering it.
What TechTalent Does Now
With 70% of manual screening time reclaimed, TechTalent's recruiters spend more time on what agencies are actually paid to do — building client relationships, managing candidate expectations, and closing placements faster.
The quality signal has changed the commercial relationship too. Clients who previously pushed back on shortlist quality are now receiving consistent, well-documented candidate submissions. Trust has been rebuilt. Repeat business follows trust.
For any recruitment agency in Malaysia asking whether AI has a place in their workflow — the TechTalent experience suggests the question is not whether to adopt it, but how quickly.
OPAL is an AI recruitment screening platform built by Oxydata for enterprise hiring teams and recruitment agencies in Malaysia. To see how OPAL can work for your agency or organisation, request a demo at oxydata.ai.