One AI talent engine that screens every applicant in minutes, keeps every candidate warm on WhatsApp, and hands your team only the best. Running 24/7, across all 10 brands.
The same hiring flow you showed us, with AI doing the heavy lifting at every step. Your team keeps every decision. The machine does the work.
CV in by WhatsApp or web, anywhere
500 applicants ranked into a shortlist
Assessments graded the moment they finish
Voice first round in Bahasa, 24/7
Letters drafted, signed, tracked
Docs collected, pushed into Talenta
A glimpse of what your team would open each morning:
This view is illustrative. Want the full picture, your workflow today vs with the engine, the architecture and the ROI in detail?
See how it works · full schemas & ROI →Built the usual way in Indonesia, a system like this costs Rp 600M to 1.5B to build, then Rp 420M to 1.2B every year to keep an AI engineer running it (see below), and you carry it all yourself. You get it built and run for you, one partner, for Rp 515M and Rp 10M a month. You approve one module at a time, starting with the cheapest, highest-return one, and only move on once you have seen it work.
Sources. Local build & talent rates: Lemon.io Indonesia (senior developer median $32/hr), Second Talent (AI engineer Indonesia $60 to 85/hr senior, Rp 35 to 100M / month), and Indonesian software-house guides KAZOKKU & Binar (a complex enterprise custom build starts at Rp 300M+).
Why not just hire someone to do this? An AI engineer in Indonesia earns Rp 35 to 100M a month, more than any role on your own team (your highest-paid role sits near Rp 16 to 20M). You get that capability for Rp 10M a month, and you never recruit, manage, or risk losing them.
You choose where to start. The prices show the full recruitment journey; you approve one phase at a time and add or remove pieces to fit. Most start with Phase 1, the Filter, because it is the fastest to a board-visible result. If you would rather start customer-facing, a WhatsApp customer-service audit is the other natural first step, the same audit-then-build method. After that, the bigger play is your core system, the one every department depends on.
Low-risk by design. You decide one module at a time, pay 50% to start and the other 50% only once Phase 1 is live on your real candidates, all backed by our written success-bar guarantee (below).
The short version for your board. Phase 1 costs Rp 95M once and Rp 10M a month. It frees Rp 65 to 145M of HR time every year and saves you a senior AI hire that would cost Rp 720M to 1.2B a year to find, manage and keep. You pay a fraction of what it returns, and a fraction of what it would cost to build or staff yourselves.
Proven in Indonesia: a Karawang manufacturer of 500 staff (8-person HR team, around 30 hires a month) ran an AI recruitment and HR program and cut time-to-hire from 45 to 18 days, freed about 120 hours of HR time a month, and halved probation turnover from 25% to 12%, reporting a 300 to 500% return within 6 to 12 months (published case study). At the global top end, Coppel hired 24,000 people in 75 days through WhatsApp AI. The same playbook, scaled to your hiring.
The brand upside: 93% of Indonesians live on WhatsApp, and your @declip.id audience is already 152k strong. Recruitment that replies to everyone, in minutes, turns every applicant into someone with a good impression of DeClip, Xuping and Meilyn, instead of silence.
Sources. Volume confirmed with your team (100 to 200 openings/yr). Benchmarks: manual screening 2 to 5 min/CV; AI frees 50 to 80% of screening time, cuts time-to-hire 30 to 50%, and lowers cost per hire ~27% (high-volume recruitment-automation studies). Salaries: Jobstreet data for PT Sumber Bintang Perkasa + SalaryExpert Jakarta, loaded BPJS + THR. Indonesian deployment: pakaiai.id case study (500-staff Karawang). Conservative, counted once.
Recruitment is what we audited and scoped in depth, that is the plan above. The same partnership extends across your business. Here is a preview of the highest-return modules we would scope with you next, module by module, each with its own ROI, so you choose where to go after the first win.
24/7 AI replies for shoppers and B2B resellers, handling 25 to 45% of routine messages and lifting sales. The fastest visible win.
AI titles, descriptions, SEO and images across your 10 brands and thousands of SKUs on Shopee, Tokopedia, Lazada and TikTok. Catalogue ~90% faster.
AI scripts, product discovery and affiliate optimisation for your live-streaming sales.
Forecasting and stock optimisation across your warehouses: fewer stockouts, less dead stock.
Reconciliation, invoicing and reporting on autopilot.
The company-wide system every department runs on, the goal you described to us.
Honest scope. We never price what we have not studied, so each of these starts with a short audit, then the same module-by-module build you see in the recruitment plan above. That recruitment plan is the worked example of exactly how we work with you, module by module.
Abdi Medicca: a WhatsApp AI agent answering patients 24/7 for a Jakarta clinic, running in production today. Apex Property: a custom CRM, dashboard and AI operations for a 115-villa portfolio. We ship working systems here, not slides. See more of our work →
Built compliant with Indonesia's data law (UU PDP) from day one, because for a market leader, trust is the brand. A sample Data Processing Agreement and a full audit log of every AI recommendation and human decision are ready for your legal team, so compliance is something you can see, not just a promise.
AI ranks and recommends. Your team makes every hire and rejection. No silent auto-rejects.
Photo, age and gender stripped before scoring. Clear consent on every application.
Stays in your systems. Vendors never train on it. Honest scope: we never overpromise.
This augments your HR team, it does not replace anyone. The same people, freed from reading 500 CVs by hand, spend their time on the candidates worth it and on the experience your brand is known for.
Our Phase 1 guarantee. Before we start, we agree one number with you in writing, your success bar: a usable, ranked shortlist from your next real opening, in minutes. You pay 50% to begin and the other 50% only once Phase 1 is live and hitting that bar on your real candidates. If we cannot get there, you owe nothing further, and you keep every asset we build (the CV parser, the WhatsApp flow, the dashboard). The risk of trying this is ours, not yours.
AI changes every month, and keeping up with it is our job, not yours. You get the ideas actually worth your time, and a straight answer whenever your team wonders "can AI do this?". Think of a senior AI team sitting next to yours, building the system and staying for the long run.
You mentioned every department now has Claude, but the value is not there yet. It is the most common gap in AI today: the tool is paid for, the habit is missing. This is the fastest return in the whole plan, because the licenses are already on your bill.
of AI comes from how people use it, not the tool itself (BCG).
on the right tasks for trained teams. Untrained teams barely move.
on every dollar put into AI, once people actually adopt it (IDC, Microsoft).
We have a few in-person training slots open until June 19, then we are fully booked on this program until mid-August. A hands-on day or two with your teams, on their real work, turns the Claude licenses you already pay for into hours saved every week. If it is useful, just say the word and we will send a short plan dedicated to it, kept separate from the recruitment build above. No commitment today.
On a short call we run it live for you: the demo on a real, anonymised batch of your own past applicants, the phases, and the full system diagram, module by module, so the decision rests on something you can see, not a promise.
You lead your market. In this era, the companies that put AI to work now pull ahead, and the ones that wait get overtaken. Phase 1 is the first, low-risk step to make SBP the leader that modernised early, not the one that catches up late. Every opening run the old way is hundreds of candidates never truly seen.