Spatial data AI startup Deeping Source is stepping up its push into the South Korean market this year and expanding its business with a target of 17 billion won in annual sales.

At a news conference held on April 2 at Deeping Source's headquarters in Seoul's Gangnam district, CEO Tae-hoon Kim (김태훈) said anonymisation technology has been competitive in Japan, where personal data protection standards are high, generating more than half of the company's total revenue. He said the company will raise the domestic share this year and achieve 17 billion won in annual sales.

Deeping Source focuses on technology that uses cameras to collect offline spatial data and uses AI to analyse and optimise it. It anonymises personal information at the first stage of collection and then analyses it.

Its overseas business is delivering results in Japan. The company said it supplies solutions to local retail companies based on a partnership with Japanese telecommunications company KDDI. It participated as a key technology partner in the first store of the next-generation smart convenience store 'Real×Tech LAWSON', jointly operated by KDDI and convenience store chain Lawson (LAWSON). It provides real-time analysis of shelf inventory status and customer behaviour, as well as an AI agent-based system to support store operations.

It is also running a project in Japan linking store patrol robots with its AI solution. The system is designed so that if a robot detects low stock or pricing label errors while moving around the store, the AI makes a judgment and issues instructions to tidy up.

Based on this, it is accelerating its expansion of results. Kim said the company posted 3 billion won in sales last year and is certain of more than 10 billion won this year based on confirmed contracts. It aims to expand the number of deployed stores from the current level of several hundred to around 10,000 this year.

Deeping Source also unveiled three AI agent solutions on Thursday.

StoreCare is a real-time monitoring tool for store staff and managers. It uses cameras to detect low shelf inventory, refrigerator abnormalities and whether tables need tidying, and sends alerts to a tablet. It displays store conditions using traffic-light colours. Kim said operations can start from the same day by plugging a device about the size of three fingers into a store. It reuses existing camera infrastructure and supports both on-premises and cloud environments.

StoreInsight is a product that collects and analyses customer data in stores. It distinguishes between customers and employees and tracks them, compiling in real time estimates of gender and age group, movement paths, interest by zone and whether items are picked up. Age estimates are based on clothing and behaviour patterns, not faces.

StoreAgent is a product in which AI receives goals and proposes execution plans to optimise stores. It automatically generates order quantity recommendations, display rearrangement simulations and strategies for handling items nearing disposal. It uses simulations that learn customer behaviour to forecast sales changes before and after display changes. Test results showed a case in which changing the placement of a specific product doubled sales volume and lifted revenue by 96 percent. The company said applying order recommendations could cut disposal by about 90 percent.

Deeping Source last month hired Sang-ho Keum (금상호), formerly of SK Broadband, as chief business officer. It is the company's first external C-level hire since its founding. Keum oversaw B2B telecommunications business strategy, sales and new business development at SK Broadband for 26 years. Keum said he will contribute so that Deeping Source's distinctive technology becomes a new business standard in the global market.

Kim set a target of turning profitable on a quarterly basis in the second quarter of this year. The company is also preparing for a listing and plans to review additional fundraising in the second half.