Mercer | Mettl is one of the most established online assessment and proctoring platforms in the world — a deep test library plus AI-powered remote proctoring, now backed by a Fortune 500 parent. If your problem is measuring skills at scale — screening thousands of applicants, running secure certification exams, standardizing a hiring funnel — it's a genuinely strong, credible choice. This review is for the person weighing it and wondering whether it still holds up in 2026. Mettl was founded in 2010 by Ketan Kapoor and Tonmoy Shingal, out of Gurgaon, India. It built a talent-assessment platform — psychometric tests, aptitude tests, job-specific skill tests, and coding evaluations — and grew it across dozens of countries before being acquired by Mercer in 2018. Mercer is part of Marsh McLennan (NYSE: MMC), so behind the product sits one of the largest professional-services firms on the planet. That backing matters. When you buy Mettl, you're buying enterprise-grade stability: security posture, uptime, compliance, and a company that will still be here in five years. The acquisition took Mettl from a fast-growing startup to a globally distributed assessment business.

The scale is real. Mercer | Mettl reports 6,000+ clients across 90+ countries and more than 12 million assessments administered. Its test library runs to 100,000+ questions covering 800+ skills across 25+ industries, with 400+ pre-built tests you can deploy out of the box. On the technical side, the coding assessments include auto-graded coding simulators that run and score real code across 30+ programming languages, with analytics on logic, structure, and correctness. The AI-based proctoring suite — SecureProctor plus a secure browser — is what lets organizations run high-stakes exams remotely without a room full of invigilators. Its client roster includes Accenture, Capgemini, Wipro, SAP, Cognizant, and Samsung.

Breadth of the library. Very few platforms match 100,000+ ready-to-use questions across 800+ skills. If you need to assess a role, there's almost certainly an off-the-shelf test for it — you're not building from scratch. Proctoring at scale. This is arguably where Mettl leads. If you're running certifications, entrance exams, or high-volume screening where integrity is the whole point, AI proctoring plus a locked-down browser is exactly the tool for the job. Auto-graded coding evaluation. The simulators execute and score real code, so you get a consistent, objective signal on programming ability without a human grading every submission. For funneling a large applicant pool down to a shortlist, that's efficient and defensible. Enterprise trust. The Mercer / Marsh McLennan parentage means procurement, security review, and legal all get answered easily. For a large org, that's not a footnote — it's often the deciding factor.

It measures skills — it doesn't source or hire people. Mettl is a testing layer. It tells you how well a candidate performed on an assessment; it assumes you've already found the candidate and will handle everything after the score. That's the right shape if you own the top of your funnel and just need a rigorous filter. A test score is a snapshot, not a working relationship. Auto-graded assessments are excellent at range and consistency. They're best paired with real-world signal — how someone actually builds — when the stakes on a single hire are high and you want more than a number. It's built for volume standardization. The platform shines when you're screening many people the same way. If your need is one senior, AI-native engineer who's already employed and not taking tests from strangers, a high-throughput assessment engine is solving a different problem than the one you have.

Here's the honest gap. Mettl is a test you run on people who show up to take it. The best AI engineers don't show up. The ones you actually want are already working — booked on a contract, heads-down on something hard, not sitting in an applicant pool waiting to be assessed. Nextdev reaches the top 1% of AI engineers — the ones who aren't looking. We go find them. Then we prove them. Instead of a score on a question bank, we give each one a real problem to build and watch how they actually work — so you know they're AI-native because you've seen it, not because a report said they passed. A proctored test tells you someone didn't cheat on a quiz. It doesn't tell you they can ship in an AI-native codebase. And then we're done — we employ them for you. One named engineer, working on your problem. The contract, the payroll, the compliance never touch your desk. Mettl hands you a result and leaves the sourcing and the hiring to you. We do the whole thing. You're not scoring who applied. You're getting the engineer who never would have applied.

Mercer | Mettl is a strong, credible platform, and if your problem is measuring skills at scale — big applicant pools, secure exams, standardized screening — it's one of the best tools you can buy, with the enterprise backing to match. Just be clear about what it is: an assessment layer, not a way to find and hire the person. If you already have the candidates and need to test them, Mettl earns its place. If you need the rare engineer found, proven on real work, and employed for you, that's a different job — and the one we do.