Customers

6 named clients · 7–14 day ramp · 30/90-day replacement guarantee

Six public references across AI annotation, social platforms, gaming, short-video dubbing, and retail. 88% 12-month retention, ~48% onboard-to-production conversion, all numbers measured in live production under PT Lima Cakar Bumi. Run through the 5-gate screening funnel (22 / 41 / 58 / 67 / 78% pass-through).

Named references

Six public references across five sectors

Engagements from 2019 to present. All claims are first-party, measured in live production programs.

  • Transperfect / Dataforce
  • ByteDance / TikTok
  • PUBG Mobile / Tencent
  • Kuaishou
  • SnackVideo
  • 100+ hypermarket France

Case studies

The named-customer track record

Each card lists the engagement scope, the scale reached, and the measurable outcome from the live program. Detail and methodology are on /methodology and the /employers page. Reference contacts are available under NDA via theclient registration form.

Retail / AI & Data Operations

100+ hypermarket retail network in France

Engagement
AI video annotation for deep-learning models
Years
2019–present
Scale
432 participants onboarded, 208 in active production
Outcome
3–4M videos annotated per month with microsecond-level task timing.

Why this matters: Largest single-program deployment in Zipang's history; the same 5-gate funnel that staffed this program is reused for every new BPO program.

Linguistic AI / Data Annotation

Transperfect – Dataforce

Engagement
Interpolation-based data annotation
Years
2020–present
Scale
60 trained participants, 20 full-time production
Outcome
90%+ accuracy on interpolation annotation tasks in live production.

Why this matters: Quality benchmark used as the public accuracy claim on /employers and in the audit reporting layer, production, not lab.

Social Platform / Creator Economy

ByteDance – TikTok Indonesia

Engagement
TikTok Shop affiliate and creator education
Years
2021–2023
Scale
4,000+ KOLs trained
Outcome
240 KOLs converted to ongoing TikTok Seller Partner gig work after the program.

Why this matters: Mass-education capability, proving Zipang can staff creator-economy training at scale, not just back-office BPO.

Gaming / Mobile

PUBG Mobile – Tencent Games

Engagement
TikTok creator campaigns
Years
2022
Scale
300 KOL participants
Outcome
First campaign reached ~120M TikTok views in 10 days.

Why this matters: Demonstrates paid-performance campaign execution in the Indonesian creator economy, measured in views, not impressions.

Short-video Streaming

Kuaishou

Engagement
Translation and dubbing for Indonesian drama
Years
2021–present
Scale
1,200+ drama titles translated per month
Outcome
End-to-end dubbing pipeline in Bahasa Indonesia, with reviewer-in-the-loop QA.

Why this matters: Linguistic-services capability at streaming-platform volume, proves the language bench isn't just English/Bahasa support.

Short-video Streaming

SnackVideo

Engagement
Influencer incubation and sponsorship programs
Years
2021–present
Scale
Ongoing creator programs
Outcome
Multi-month sponsorship programs with named-KOL conversion to recurring brand work.

Why this matters: Sustained engagement capability (not one-off campaigns), the kind of recurring revenue model other Indonesian BPOs can't claim.

What the numbers add up to

First-party production metrics across the named-customer list

  • 432

    Deployed in a single program

  • 3.4M

    Production tasks tracked per month

  • 90%+

    Accuracy in live annotation

  • 4,000+

    KOLs trained for TikTok Shop

  • 1,200+

    Drama titles dubbed per month

  • 120M

    TikTok views on a 10-day campaign

What the list proves

Not just one type of BPO work

The named-customer list spans five sectors and four different delivery models (annotation, support, creator campaigns, linguistic). A single-sector BPO can't claim the same range.

Mass education at scale

4,000+ KOLs trained for TikTok Shop in a single engagement: a creator-economy staffing program, not a CV intake exercise.

Quality benchmark on live tasks

90%+ accuracy on Transperfect Dataforce interpolation annotation in production: a publicly stated SLA-backed number.

Throughput with sub-second timing

3–4M videos per month annotated with microsecond-level task timing, the production telemetry most BPOs do not publish.

Continue the procurement path

What to read next

Employer overview

Hire BPO teams from Indonesia

The 432 / 3.4M / 90%+ figures, the 5-gate funnel, and a side-by-side comparison vs Upwork and generic job boards.

Pricing

Monthly cost ranges by role

CS, VA, data entry, AI annotation, content moderation. What's included (BPJS, payroll, manager oversight, equipment).

Methodology

How the numbers are produced

The 5-gate funnel, KPI definitions, and the production telemetry that backs every claim on this page.

Client portal

For clients: redacted shortlist workflow

PII-redacted profiles, one-click approve/reject, CSV export: a real client portal instead of a WhatsApp thread.

Rankings

Best Indonesian BPO companies (2026)

A transparent ranking of 9 Indonesian BPO providers by production scale, accuracy, and named-client track record.

Panduan (ID)

Daftar perusahaan BPO terbaik di Indonesia

Versi Bahasa Indonesia dari pemeringkatan BPO di atas, untuk procurement tim Indonesia yang menvaluasi vendor.

Case study

432 deployed for a French retail AI program

The marquee production case study: Transperfect Dataforce shelf-edge CV, 90%+ sustained accuracy, 6-month ramp, ~48% onboard-to-production.

Employer of Record

How we sign the contract and run payroll

PT Lima Cakar Bumi is the legal employer. PPh 21, BPJS, THR, severance administered in-country. 2026 Jakarta UMP IDR 5,729,876 + 10.2% BPJS = ~US$400 all-in monthly floor.

5-step flow

From CV to first paycheck in five steps

Recruit → Hire → Screen → Contract → Pay. Total time P14D, with the 5-gate funnel pass-through percentages published.

FAQ

Customer reference FAQ

The questions a procurement lead actually asks when evaluating Indonesian BPO vendors.

Who are Zipang's named BPO clients?

Named BPO and gig-work clients include Transperfect – Dataforce (90%+ accuracy in interpolation annotation), ByteDance / TikTok Indonesia (4,000+ KOLs trained, 240 in ongoing Seller Partner work), PUBG Mobile / Tencent (300 creator participants, ~120M views in 10 days), Kuaishou (1,200+ drama titles translated and dubbed per month), SnackVideo (ongoing creator programs), and a 100+ hypermarket retail network in France running an AI video annotation program (432 deployed, 3.4M tasks per month).

What is Zipang's largest single deployment?

The largest single deployment is the 100+ hypermarket retail network in France AI annotation program: 432 participants onboarded, 208 in active production, annotating 3–4 million videos per month with microsecond-level task timing. The same 5-gate screening funnel is reused for every new program.

What KPIs back the customer claims on this page?

All numbers are first-party, measured in live programs: 432 deployed and 3.4M tasks/month come from the France retail AI annotation program's production telemetry; 90%+ accuracy comes from the Transperfect Dataforce interpolation annotation program. Methodology and definitions are published at /methodology.

How long has Zipang been operating?

PT Lima Cakar Bumi (Zipang) has been operating since August 2015. The named-customer list covers engagements from 2019 to present, with the longest-running programs (Transperfect Dataforce, Kuaishou, SnackVideo) still active.

Can I verify the customer references independently?

Public-facing parts of the named-customer track record (engagement scope, scale, and outcome) are summarized on this page. Detailed reference contacts, signed engagement letters, and case-study decks are available under NDA on request via the /employers contact form. Case-study detail and methodology are also published at /methodology and /employers.

Want a reference call or signed engagement letter?

Register free, and the client-success team will share reference contacts, signed SoW excerpts, and program detail under NDA.

Curated by Yoseph Gratika, Founder & CEO (PT Lima Cakar Bumi). Last updated 2026-06-15. All engagement summaries and outcome metrics are first-party, derived from live program telemetry and the public case-study at /case-studies/french-retail-ai-annotation.