Moka AI Puts a Number on Its Recruiting Agent: 75% Faster, Self-Reported
The News
Moka AI has published deployment results for Moka Eva, its AI recruiting agent, at an unnamed international coffee and tea chain with more than 20,000 locations. The company says Eva cut per-store time-to-hire from 15.4 to 3.8 days, lowered cost per application by 60% and manual intervention by more than 80%, and lifted offer-to-onboarding conversion from 14.2% to 35.8%. It also claims the agent's assessments are consistent with human reviewers more than 92% of the time, and that Eva has screened over 5 million résumés and assisted in 600,000-plus interviews across more than 10 languages. The figures are self-reported with no independent validation, and the release itself cites KPMG data showing 55% of Asia-Pacific companies have delayed or scaled back AI agent rollouts because operating costs began to outstrip the value generated.Why It Matters
Vendors have been selling agentic recruiting on adjectives; this is one of the first times one has attached a before-and-after number to a large, distributed deployment. That is genuinely useful as a benchmark to interrogate — but only as a starting point. The same week's KPMG finding that most firms are quietly scaling AI agents back because the bills outpace the value is the necessary counterweight: a vendor's 92% 'consistency with human reviewers' tells you nothing until you know what the other 8% of decisions looked like, and a 75% time-to-hire cut on a 20,000-location retail network may not transfer to your ten-location, high-complexity hiring.Taking Action
All action is situated in your corporate context. These are starting points, not prescriptions — choose what fits.
Connect
Who to talk with in your enterprise
Your head of talent operations and your finance partner. Walking point: establish your own time-to-hire and cost-per-application baselines for high-volume roles before any vendor demo, so a claimed '75% faster' lands against a number you actually own. Your procurement or vendor-management lead: ask any AI-agent vendor for a reference deployment with pre/post metrics and independent validation, not a self-published case study.Inspect
What to take a closer look at
Pull your current time-to-hire and cost-per-application for your highest-volume roles and write them down as a baseline. For any AI screening or scheduling tool you already run, ask the vendor for its deployment telemetry — specifically its false-pass and false-fail rates, not just an aggregate 'consistency' percentage. Check whether your agentic-AI pilots have a defined cost model, because the KPMG pattern (costs outrunning value) is a rollout failure mode, not a launch failure mode.Reflect
What to think about more deeply
One vendor reported a 75% time-to-hire cut; the same week, KPMG found most firms scaling back AI agents because the bills outpaced the value. Both can be true — but only one gets a press release.