ITSpace credits a repointed factory AI with cutting Incheon market electrical fires 60%

The redeployment is literal. UYeG was watching current, temperature and arcing at industrial sites before ITSpace pointed it at traditional markets in Incheon, and it watches the same three signals there ^[3][4]. The site changes and the sensor count changes. The research carries over, so the cost of the second deployment is installation.

The two percentages come from the vendor. Yoon Hye-jin, vice president of ITSpace, told the forum that the project cut the rate of electrical fires at the markets by 60% and reduced maintenance costs by 40% ^[5]. Yoon did not give the number of markets, a baseline fire count, an installation price or the period over which the comparison was made ^[11].

Take the 40% as given, and maintenance spend is running at 60% of its old level ^[12]. For a market association the sum that decides anything is install cost against that saving. A system priced at one year of the previous maintenance bill repays itself in two and a half years. One priced at six months of that bill repays in fifteen ^[13]. Fires avoided are a public benefit. The 40% is what a market association can put in a budget ^[5].

If most of the fire reduction came from rewiring and repairs that the warnings triggered, the software is a diagnostic layer on a conventional electrical retrofit, and the credit belongs to the remediation spend. Yoon credited the project as a whole ^[5].

Two of the other presentations follow the same substitution pattern without producing outcome figures. Dolbomdream's Caerly reads heart rate, respiration and sleep through an air-pressure ballistocardiography sensor and gives early warning of conditions such as arrhythmia and heart failure. Its chief executive Kim Ji-hoon described AI deployed instead of additional staff ^[8][9]. Dalgu, according to chief executive Park Si-eun, is still developing technology to share the vital signs of patients in ambulances with hospitals in real time, in place of radio and phone calls. In that system, vision AI analyses the first-aid procedures ^[10]. Of the three companies, only ITSpace brought percentages ^[14].

What the forum supports is the pattern. One deployed case with vendor-supplied numbers is enough to show that an industrial model can be moved into a public setting for the cost of sensors and labour. It is not enough for a city procurement officer to treat 60% as a specification. Je Young-jun, head of the startup division at the Ulsan Center for Creative Economy and Innovation, said: "There is a lot of room to move technology proven at industrial sites into areas that citizens can feel directly." ^[7]