Our own platform

Artificial Watson

The AI-powered operating platform that runs Triple-E's distributed telecom field workforce.

22
business modules in one platform
Web + mobile
browser portal, Android and iOS app
12
AI-assisted workflows in daily use
≈180
active user accounts every month
What Watson is

One platform for a workforce spread across an archipelago

The operating challenge

≈36,000 sites, Sumatera to Papua

Telkomsel BTS and power sites where Triple-E secures network availability.

Many companies, one job

Triple-E employees work alongside partner field teams and subcontractors — in Sumatera alone, 675 partner technicians.

24/7 work at remote sites

Tickets, fuel and repairs are judged on MTTR and availability KPIs, far from any office.

Watson's answer: one integrated system

≈180
active user accounts per month
3,300+
mobile app sign-ins per month
565
vendor-portal sign-ins per month
How Watson is built

Five engineering commitments behind the platform

Select a commitment to see how it is put into practice.

Security

  • Two-factor login on web and app
  • Role-based access, closed by default
  • Secrets kept out of code and out of the public web area

Life-cycle management

  • All code in GitHub with two-owner review
  • Docs-as-code: architecture, API spec, decision records
  • Logged server changes and a renewal calendar

Business continuity

  • Company-owned accounts in a shared vault, two owners
  • Register of 23 critical dependencies
  • Written takeover runbook for a successor team

Resilience

  • Encrypted backups to two off-site locations
  • Deletion-protected copies kept 30 days
  • Restore drill completed in 51 seconds

Adaptation to new technology

  • One Flutter codebase for Android and iOS
  • API layer decouples apps from the database
  • AI model swapped by configuration, not code
Key feature · AI at the core

12 AI-assisted workflows built into daily work

Select a workflow to see what the AI does and who benefits.

Operations

Finance & commercial

People

How it runs
Automatic jobs every 15 minutes and nightly
AI suggests, people approve
Documents read as data, never as instructions
Model chosen by configuration
Key feature · AI benefits

What AI delivers for the company and for every employee

For Triple-E

Speed

Tickets rated and sites briefed automatically, around the clock.

Control

Photo and receipt checks reduce fuel and expense leakage risk.

Scale

More sites and contracts without a back office that grows in step.

For employees

Less typing

Snap a receipt, PO or form; AI fills in the fields.

Ready at the site

A short brief of the site's history before each visit.

Fair recognition

One rubric for every ticket; skills visible for career moves.

≈10,000
AI rating suggestions on tickets
≈550
job applicants with AI skill readings
≈330
AI site briefs kept current
130
employees with AI skill profiles
Key functions

Nine functions, one platform

Each function runs on the same data, on web and mobile. Select a card for the full picture.

AI inside

Trouble tickets & field operations

Tickets on the phone, live maps, worst-site ranking and quality rating — built around MTTR.

AI inside

Fuel (BBM) management

Request, approve, refill with photo, AI meter check, handover report, PR/PO and payment.

Power, spare parts & assets

Hybrid power at off-grid sites, backup power, spare-part inventory and the asset register.

AI inside

Remote workforce management

Profiles, time reports, leave, team map, chat and the company handbook — on the phone.

AI inside

Recruitment, training & competence

Career site, AI-read CV database, best-fit positions, hiring, training and certifications.

AI inside

Expense & accounting

Snap the receipt, AI fills the claim, the manager approves and finance pays.

AI inside

Sales, sourcing & delivery

Customers and contracts, vendors and POs, execution and delivery progress in one flow.

Vendor portal

Partners see their own POs, acceptance, invoices and payments — on web or in the app.

Management insight & governance

Dashboards, Power BI, e-signature, organisation and projects, Microsoft 365 integration.

Adapting to new technology

A managed modernisation plan

Delivered in 2026

  • New Flutter app on Google Play and the App Store
  • Two-factor login for web and app
  • Encrypted off-site backups and restore drill
  • Full technical documentation programme
  • Company accounts moved into a shared vault

Next, in plan

  • Runtime and framework upgrades to current supported versions
  • A dedicated hosting environment for Watson
  • Continued security hardening
  • A successor takeover drill

Continuous

  • Extend AI to more workflows as models improve
  • Two-owner review on every change
  • Nightly live-system snapshot and renewal calendar
  • Documentation updated with each release
Summary

Why Watson makes Triple-E different

Field to finance in one system

Tickets, fuel, people, vendors and money share one database, on web and mobile.

Built for a remote, multi-company workforce

Employees, partner field teams and subcontractors work on the same platform.

AI in everyday work

12 live workflows; AI suggests, people decide.

Owned, not rented

In-house platform shaped to Triple-E's contracts and ways of working.

Engineered to last

Security, backups, continuity and documentation for a successor team.

The bottom line

Watson turns Triple-E's operating know-how into a scalable, auditable and AI-ready platform: an advantage that generic tools and spreadsheets cannot easily copy.

Figures: Watson production data, October 2026; values marked ≈ are approximate. Site and workforce figures: July 2026.

Want to see Watson at work in your operations?

Talk to Our Team
How Watson is built

Security

Watson holds operational, people and commercial data for a company that is itself ISO 27001:2022 certified. Security is designed in at every layer, from login to the AI prompts.

In practice

How Watson is built

Life-cycle management

Every change to Watson is versioned, reviewed and documented, so the platform can evolve quickly without losing control.

In practice

How Watson is built

Business continuity

Watson is designed so that the business never depends on a single person. If key people change, a successor team supported by AI can take over.

In practice

How Watson is built

Resilience

Backups only count if they can be restored. Watson's data is copied off-site every night, protected against deletion and tested by real restore drills.

In practice

How Watson is built

Adaptation to new technology

Watson is built to absorb new technology — new devices, new channels and new AI models — without a rewrite.

In practice

AI workflow · Operations

Ticket quality rating suggestions

After a ticket is closed, the AI reads its history and suggests a quality rating against Watson's shared rubric. A person confirms or adjusts it, so every engineer is judged by the same standard. Runs automatically every 15 minutes.

AI workflow · Operations

Site briefs from each site's history

For each site, the AI keeps a short brief of recent faults, repairs and recurring issues, so an engineer knows what to expect before arriving. Briefs refresh automatically every 15 minutes.

AI workflow · Operations

Structuring free-text ticket notes

Engineers write notes in their own words. The AI turns them into structured data — cause, action, parts — that can be counted, ranked and analysed across thousands of tickets.

AI workflow · Operations

Note assistant in the mobile app

Before a ticket note is sent, the assistant checks it and suggests improvements in Indonesian or English. It advises; it never blocks the engineer from sending.

AI workflow · Operations

Genset meter-photo verification

Refill photos carry the running-hour meter, time and GPS text burned into the image. The AI reads them and Watson compares the photo's location with the site, so suspicious refills stand out before they are paid.

AI workflow · Finance & commercial

Expense receipt reading

Staff photograph a receipt or upload a PDF; the AI fills in merchant, date, total (including tax and service) and a suggested category with a confidence level. Images that are not readable receipts are rejected rather than guessed.

AI workflow · Finance & commercial

Customer PO and bill-of-quantities extraction

A customer purchase order is read straight into Sales: PO number, subject, contract reference, issuing customer, value, dates and every line of the bill of quantities.

AI workflow · People

CV reading into a structured profile

Every CV — from the career site or imported — is read into a structured profile: contacts, education, experience, organisations, soft and technical skills, and certifications.

AI workflow · People

Skill scoring and best-fit position

Skills are scored 0–100 and matched against Triple-E's positions to suggest where a candidate fits best. Recruiters see the reasoning and make the decision.

AI workflow · People

Employee skill profiles, refreshed daily

A nightly job keeps each employee's AI skill profile current, which supports training plans and internal moves.

AI workflow · People

Hand-filled application form reading

Scans of hand-filled candidate application forms are read into Watson's candidate data, removing manual re-typing.

AI workflow · People

Indonesian → English career pages

Position descriptions written in Indonesian are translated to English automatically, so the public career pages stay bilingual and up to date.

Key function · Trouble tickets & field operations

Every network fault tracked from alarm to closure

Trouble tickets are where network availability is won or lost. Watson manages every fault on the sites Triple-E maintains, from the moment it is raised to the moment it is closed and rated.

What it covers

AI inside

AI suggests a quality rating for each closed ticket, keeps a brief of every site's history and turns free-text notes into structured data. In the app, a note assistant helps engineers write clear notes — without ever blocking them.

≈10,800
tickets recorded
62
API operations
10
mobile screens
Read: Performance management that shows its work →
Key function · Fuel (BBM) management

Genset fuel controlled from request to payment

Gensets keep off-grid and backup sites running, and fuel is one of the largest and most leakage-prone costs in field operations. Watson turns fuel logistics into fuel governance.

What it covers

AI inside

AI reads the running-hour meter, time and GPS text burned into each refill photo, then Watson checks the photo's location against the site. Suspicious refills stand out before they are paid. A browser-based reader acts as a fallback.

≈1,250
fuel-managed sites
≈9,500
fuel transactions
76
web pages: the largest module
Read: Turning fuel logistics into fuel governance →
Key function · Power, spare parts & assets

Physical resources linked to the sites and tickets they serve

Availability depends on power and on having the right part in the right place. Watson keeps these physical resources visible and tied to the work they support.

What it covers

4
connected modules
35
web pages
8
mobile screens
Key function · Remote workforce management

Managing people you rarely see, in one app

Most of Triple-E's people work at remote sites across Indonesia. Watson gives managers a clear view of who is working where and on what, and lets staff handle their HR tasks without travelling to an office.

What it covers

AI inside

Employee skill profiles are refreshed by AI every night, supporting training plans and internal moves.

≈180
active user accounts per month
3,300+
mobile sign-ins per month
149
API operations
Read: Turning remote field teams into one connected workforce →
Key function · Recruitment, training & competence

From applicant to skilled engineer, with AI reading every CV

For a services company, skilled people are the product. Watson runs the whole journey from first application to trained, certified engineer.

What it covers

AI inside

AI reads each CV and hand-filled application form, scores skills from 0 to 100 and suggests the best-fitting positions. Recruiters see the result and decide.

≈850
job applications
≈550
AI-assessed CVs
5
AI workflows for people
Key function · Expense & accounting

Field expenses claimed from a phone, settled with a clear trail

Field staff spend money far from the office. Watson makes claiming fast for them and keeps every rupiah traceable for finance.

What it covers

AI inside

AI reads the receipt photo or PDF: merchant, date, total including tax and service, and a suggested category with a confidence level. Unreadable images are rejected, not guessed.

≈2,800
expense records
50
API operations
20
mobile screens
Key function · Sales, sourcing & delivery

The commercial chain from customer PO to delivered work

Watson connects what Triple-E sells, what it buys and what it delivers, so commercial and operational teams work from the same numbers.

What it covers

AI inside

AI reads a customer PO — including its bill-of-quantities table — straight into Sales: PO number, subject, contract reference, customer, value, dates and every line item.

125
API operations across the chain
≈110
vendors in the sourcing base
42
mobile screens
Key function · Vendor portal

Subcontractors work inside Watson, not around it

Partners and subcontractors sign in to their own view of Watson, on the web or in the same mobile app. They see exactly their own orders and money, so fewer e-mails and calls are needed to close work and get paid.

What it covers

565
vendor sign-ins per month
61
API operations
22
mobile screens
Key function · Management insight & governance

Dashboards, signatures and structure in one place

Leaders need a reliable view of the business and a controlled way to sign and keep documents. Watson provides both on the same data the teams use every day.

What it covers

9
e-signature pages
30
organisation & project API operations
1
shared database behind every view