JCon

AX consulting · Build · Operate

Connect the context.Automate the core.

We connect the data and know-how scattered across your company with AI agents, and build products and systems for your teams to use every day.

  • 01Founder: 20+ yrs in enterprise systems
  • 02In-house AX platform, running daily
  • 03Two products, real screens
01 — JCon ERP ChatReal screens
JCon ERP Chat — Just ask your ERP

01 — Products

Built by us, shown on screen

We turn repetitive work into products. Every product is published with real screen recordings.

01 — Conversational ERP queries

JCon ERP Chat

Just ask your ERP

Query data in Korean ERPs such as Douzone in plain language, without hunting through menus. It follows the conversation and answers with tables and charts.

View product

02 — Receipts · draft entries · approvals

JCon Bookkeeping Agent

Snap the receipt. Review only what needs review.

Employees upload receipt photos; AI reads merchant, business number, amounts and approval number, then drafts the journal entry. Only items that hit company policy go to bookkeeping and the CEO.

View product

Tools we already run in-house

Running daily in JCon’s own operations, not yet sold externally · as of 2026-09-19

  • 01

    Knowledge chat

    Ask in a chat app and get an answer from the internal knowledge base, with the source documents.

  • 02

    Scheduled reports

    Market and indicator data collected, summarized and sent to chat at fixed times.

  • 03

    Sales lead collection

    Prospects gathered from public information and organized into a daily list.

  • 04

    Knowledge distillation

    Decisions and know-how extracted from the day’s conversations and moved into the wiki.

02 — Solutions

Every company gets stuck in a different place

The same AI solves different problems depending on who uses it. We start from one of three places.

01Enterprise AX teams

From pilots to daily operations

Where it gets stuck

You have run several pilots, yet no AI is used by the business every day. It is hard to see which data goes to which model, and what it costs.

What JCon does

  • Install an AI agent operating model with roles, review and approval steps
  • Rules that route data to models and mask values by sensitivity
  • Audit records of calls, cost and outputs, plus pre-send checks
  • Take one business workflow all the way to automation that runs daily

What you keep

Internal AI governance rules and one workflow the business uses every day

02Small and mid-sized companies

The AX team you do not have to hire

Where it gets stuck

You cannot afford a dedicated team. Work moves between ERP, spreadsheets and chat because people copy it by hand.

What JCon does

  • Find the one task that stalls most often and start there
  • Connect to the ERP, spreadsheets and chat you already use — no new system
  • Query in one chat message and receive reports at fixed times
  • The same team keeps running and improving it after launch

What you keep

A work agent you query in chat, plus scheduled reports

03Knowledge-asset LLM

An internal AI that answers in your company’s words

Where it gets stuck

Know-how lives in people’s heads and scattered files, and leaves when they do. Uploading internal material to an external AI feels risky.

What JCon does

  • Structure documents, conversations and work logs into knowledge with source and date
  • An internal knowledge chat that shows the source documents with every answer
  • Sensitive material handled by a local LLM on your own server, with nothing sent out
  • A human approval step so wrong content does not build up in the knowledge base

What you keep

An internal knowledge base and knowledge chat that stay when people change

03 — AX platform

We prove it on our own work first

JCon built the AX operating model we propose to clients inside our own company first, and runs it every day. Client systems are built separately inside the client’s environment; client data is processed in the client’s environment as a rule and is kept separate from JCon’s internal system.

20
AI agents with defined roles
144
Automation scripts
2,040
Automated tests
2,100+
Internal work conversations on record

Measured on JCon’s own operations · as of 2026-09-19

  1. L1

    Command

    Hand off work in one chat message and get a report back when it is done.

  2. L2

    AI operations team

    Agents split into planning, building, review, legal and reporting roles check each other’s work.

  3. L3

    Knowledge base

    JCon’s own conversations, decisions and documents accumulate as context for the next task.

  4. L4

    Automation pipelines

    Scheduled collection, reports and alerts run on time without a person in the loop.

  5. L5

    Quality gates

    Nothing counts as done until automated tests and pre-send checks pass.

04 — Consulting & delivery

One team from diagnosis to operation

We do not start by picking tools. We find what stops the work, build the structure, and automate only where that structure is in place.

  1. 01Find the problem

    Collect, case by case, where work stalled this quarter and who filled the gap by hand.

  2. 02Gather the parts

    Pull in every piece of data, system, person and rule that touches the problem.

  3. 03Structure

    Define how the parts relate, and attach a source and a date to every value.

  4. 04Connect & automate

    Hand the repeating paths to agents, then keep improving them in operation.

01

AI agents & automation

We design and build AI agents that read your internal systems and take over repetitive work, including overnight and scheduled jobs.

Agent design · Workflow automation · LLM API integration · Local LLMs

02

Knowledge assets

We organize documents, conversations and work logs into searchable knowledge, so the decisions and reasons you have recorded can be found again.

Internal knowledge base · Document search · Structured work logs

03

AI data dashboards

AI agents gather and clean shop-floor data, then draw it into charts, dashboards and HTML reports. Charts are built with open-source libraries such as ECharts and D3. The founder built Power BI, Tableau and SEEQ dashboards in their personal career (each is a trademark of its owner; JCon is not affiliated with them).

AI-built dashboards · HTML reports & decks · Open-source charts (ECharts, D3) · KPI dashboards

04

Web & app development

From planning to deployment and operation with Next.js, React and Python FastAPI.

Next.js · React · FastAPI · Cloud operations

How to start

  1. 01AX assessment

    We review your work and systems together and deliver a report on which problems to solve first, and in what order.

  2. 02Pilot build

    We pick one workflow, build automation that actually runs, and measure the effect with you.

  3. 03Operating partner

    We keep what we built running and add new workflows one at a time.

05 — Company

AI, applied by people who know the systems

JCon founder Zoo Young Lee has led systems projects for manufacturers and distributors in Korea and abroad for more than 20 years. Knowing how ERP, MES and the rest fit together, we start by finding where AI actually removes work.

The person — founder’s career

20+ yrs
Systems projects in Korea and abroad
8 domains
From ERP to smart factory
KO · EN
Global team collaboration

* These figures describe the founder’s personal career, not the company’s years in business.

The company — JCon

We turn the founder’s field experience into AI products and an operating model.

  • 01Two products (JCon ERP Chat · JCon Bookkeeping Agent) — real screen recordings published
  • 02AX operating model built and used in-house every day
  • 03The same team owns diagnosis through operation

Systems we work with

ERP · SCM · WMS · MES/POP · CRM · PLM · Smart factory · BI

Selected experience

Enterprise systems projects at HD Hyundai Construction Equipment, Yuhan-Kimberly, Samsung Electronics and others (founder’s career)

How we work

  1. 01

    We use it ourselves first

    Anything we recommend to clients runs inside our own company first.

  2. 02

    Measured is not estimated

    We label what we measured and what we estimated, and say so when we could not check.

  3. 03

    Data leaves only as needed

    Sensitive values are masked or handled in-house, and we show you where data flows before we start.

  4. 04

    Small first, then layer by layer

    We ship a small version that works end to end, then add one feature at a time.

Contact

What work would you like to reduce, and what would you like to expand?

Tell us which systems you use, which work you want to reduce and which you want to expand. A person reads every message and replies directly.

You can also reach us by email or phone.

jconitai@gmail.com+82-10-3353-5253