AI engineering studio — Jakarta

Production AI,run by agents,governed end to end.

AIXOne Intelligence builds vertical AI systems for Indonesian founders and growing companies, staffs them with a fifteen-agent operations team, and audits every layer, so one founder can run what used to need a department.

Build
Vertical AI systems for five industries
Operate
A fifteen-agent team working scheduled shifts
Govern
Audit evidence for every model and agent call
Why AIXOne

Most agencies wire a chatbot to a prompt and call it AI. We build the way telco and banking systems are built: hexagonal architecture, deterministic pipelines, evaluation before anything ships. Then we hand you a team to run it, and a paper trail that proves it works.

Approach

Three jobs,
one engagement.

Every engagement covers all three: a system worth building, a team to run it every day, and a governance layer that shows it is safe.

Build

Vertical systems, not generic chatbots

Each system is designed for one industry's data, vocabulary and rules, then deployed to production on your cloud.

  • Five productised verticals
  • Deterministic-first: most steps are plain code
  • Evaluated on accuracy, cost and latency
View verticals →
Operate

AgentOps: a one-person company

You stay CEO. Five chief agents turn your direction into plans; fifteen worker agents execute on schedule and report back.

  • CTO, CMO, COO, CFO and CHRO agents
  • Working hours built into every agent
  • A live 2D view of the team at work
Ask for a walkthrough →
Govern

aigov: audit evidence on demand

A CLI-first audit layer across LLM output, agent tool calls, RAG and infrastructure. It produces evidence; it never claims to certify.

  • Six audit layers
  • Mapped to NIST AI RMF and ISO/IEC 42001
  • PII, prompt-injection and red-team checks
Ask about aigov →
Verticals

Five verticals.
One production standard.

Each vertical is a named product with a fixed scope and an architecture we have built before, so your project starts from a working blueprint instead of a blank page.

Fintech · Lending · Payments

AIXONELedgera

Risk and revenue intelligence for lenders, payment companies and fintech teams.

  • Credit scoringScores with SHAP explanations an analyst can read.
  • Fraud & anomaly detectionFlags unusual transactions and writes the case narrative.
  • Revenue forecastingWeekly forecasts with the drivers behind each change.
  • Ask-your-ledger analystPlain-language questions over your data, read-only by default.
Health · Hospitals · Clinics

AIXONEMedisyn

Operations intelligence for hospitals and clinic networks, from the ER to discharge.

  • ER command boardBeds, triage queue and staffing in one live view.
  • Patient flow forecastingLength-of-stay and admission forecasts per ward.
  • Clinical summariesSummaries of notes with citations back to the record.
  • PII redactionPatient identifiers removed before any model call.
Knowledge management

AIXONECorpus

Your documents and databases, answerable in plain language, with sources.

  • Multimodal agentic RAGAnswers from PDFs, slides, scans and images.
  • Natural language to SQLAsk the warehouse; get the query, the table and a chart.
  • Cited answersEvery answer links to the source page it came from.
  • Permission-awarePeople only see answers from files they can already open.
CRM · Sales · Support

AIXONERapport

One orchestrated team for sales, marketing and customer service.

  • Lead scoring & enrichmentRanks inbound leads and fills in the missing details.
  • Omnichannel serviceWhatsApp, email and web chat answered from one brain.
  • Churn alertsAt-risk customers flagged with a suggested next step.
  • Campaign draftingDrafts go to a human for approval before anything is sent.
Manufacturing · Industrial

AIXONEForgeline

Plant-floor intelligence for quality, safety and uptime.

  • Sensor drift detectionCatches failing sensors and process drift early.
  • Visual inspectionDefect detection and PPE checks from line cameras.
  • Predictive maintenanceMaintenance windows planned before breakdowns.
  • Automatic shift reportsEach shift ends with a written summary of what happened.
01 / 05
Stack

Open standards,
production tooling.

Agents connect to tools through MCP and to each other through A2A, so nothing we build locks you into one vendor.

Orchestration

  • LangGraph
  • Google ADK
  • MCP · A2A

Evaluation

  • Langfuse
  • Arize Phoenix
  • RAGAS · DeepEval

Safety

  • NeMo Guardrails
  • PyRIT · Garak
  • Presidio

Data

  • Weaviate · Qdrant
  • BigQuery
  • Redis semantic cache

Delivery

  • FastAPI · Next.js
  • Docker
  • GCP Cloud Run
Process

Four stages,
no black box.

You see the architecture, the evaluation numbers and the cost per request before anything goes live.

01

Discovery

Map the workflow and decide whether it needs a vertical agent or a simpler copilot.

02

Architecture

Hexagonal, vertical-slice design. LLM calls only where judgment is needed.

03

Build & deploy

Shipped to Cloud Run after passing accuracy, cost and latency checks.

04

Operate & govern

AgentOps runs it daily; aigov audits it continuously and keeps the evidence.

Pricing

Priced for founders,
not headcount.

Starting points for planning. Final scope and price are set after the strategy call.

Starter
Rp 15 jt/ month

The agent team for a solo founder, without a full platform build.

  • AgentOps fifteen-agent team
  • One workflow automated
  • Weekly CEO report
Book a call
Growth
Rp 45 jt/ month

AgentOps plus one vertical product, live in production.

  • Everything in Starter
  • One vertical: Ledgera, Medisyn, Corpus, Rapport or Forgeline
  • Evaluation and cost dashboard
  • Monthly architecture review
Book a call
Scale
Custom

Multiple verticals plus the aigov governance layer.

  • Everything in Growth
  • aigov audit and evidence reports
  • Dedicated architecture on-call
Book a call

Prices exclude cloud and model usage, which are billed at cost.

SimonFounder & principal engineer
Jakarta, Indonesia
“The person who designs your architecture is the person on your strategy call. There is no hand-off to a junior team.”

About six years in data engineering, five in software engineering and four in AI engineering went into the systems behind these verticals, across telco, finance, mining and knowledge management.

Contact

Your next hire could be fifteen agents.