These are AI agent automation cases built by Dawell Service FDEs (Forward Deployed Engineers) embedded directly at customer sites. From investment and logistics to commerce, tax, healthcare, HR, and content — no matter the industry, agents take over the repetitive work.
Automation agents built at real customer sites — live cases alongside projects currently in active development.
AgentV automates investment management — from gathering portfolio company data to drafting and sending investment reports.
View case →AI agents automate content production and customer support, supporting 400K shipments a month.
View case ↗A settlement agent that automates group-buy reconciliation and payout calculation.
View case →A tax automation agent that drafts corporate income tax reports automatically.
View case →A healthcare CS agent that answers KakaoTalk inquiries 24/7 and routes sensitive questions to the front desk.
View case →A Dreamace logistics agent that automates import RFP generation and cost estimation.
View case →Automatically generates payouts and deliverables from e-learning content partner (CP) settlement raw data.
View case →Search draft documents in natural language, cross-check supporting amounts, and flag duplicate filings.
View case →An accounting automation agent that organizes and verifies receipts and carries them through to settlement.
View case →An agent that generates articles and content automatically — from research to first draft.
View case →Automatic work-hour calculation from attendance raw data, anomaly detection, and a dashboard.
View case →Automatically classifies overseas corporate card transactions with a rule engine and matches them to approval requests.
View case →Different industries, but every case above is built on one shared foundation — a common skeleton + proven parts + dev automation tooling + server infrastructure.
The shared foundation that compresses a months-long AX cycle into 1–2 weeks. See how skeleton, parts, and garage fit together, explained through a car analogy.
Tech foundation overview →The common Django skeleton every project rides on — auth, permissions, utilities, and tests built in.
Details →Proven shared parts for LLMs, Google APIs, and document parsing — plugged in via pip.
An AI-native dev flow that links create → setup → build → deploy with a handful of commands.
Details →Multi-tenant operation of many projects on one server, plus automatic domains and backups.
Details →Hands-on training that gets the whole team using this foundation equally well — education.popupstudio.ai.
Details →The AX we deliver to customers, first applied inside our own company — a system where agents actually do the work. Meeting notes, emails, and work files accumulate into an internal knowledge system (KMS), and agents work on their own on top of that context.
See the system that flows from meeting notes (MeetClaw) → KMS (Knowclaw) → email agent (MailAgent), and the agents working on top of it (company-wide project tracking, standarda dev automation).
Our own AX overview →An autonomous meeting-notes agent that records and organizes on its own, accumulating meetings as work context in the KMS.
Overview →A self-growing internal LLM wiki — the center of the work context shared by all three agents.
Overview →An autonomous agent that classifies email on its own and even drafts replies, leaving only the mail you actually need to see.
Overview →