AI’preneurs participants develop a digital geologist

The second stage of the 4th cohort of AI’preneurs, an Astana Hub programme focused on creating AI-based startups, is currently taking place in Astana. One of the teams formed within the programme, Mineral Intelligence, is developing a digital geologist designed to accelerate the analysis of archival data and assess the potential of prospective areas.

The programme consists of two stages. During the four-week Founder Lab, participants study market challenges, validate their ideas, and form teams. This is followed by the 10-week Core Accelerator, where teams develop initial product versions, test them with potential customers, and prepare for market launch. The current stage in Astana is the Core Accelerator phase.

How to find gold in a “dark room”

Imagine a dark room covering an area of 10 of square kilometres. To find a small object inside, one option would be to illuminate the entire space, requiring significant time and resources. Another approach is to identify areas with a higher probability of discovery in advance and avoid investigating locations that appear unpromising.

This is the principle behind Mineral Intelligence. The system analyses historical geological reports, transfers extracted data onto modern maps, and supplements it with Earth remote sensing data. This helps geologists identify areas requiring further investigation.

A significant portion of Kazakhstan’s subsurface information is still stored in Soviet-era archives. These include typewritten reports, tables, maps, and separate appendices. Data on samples, and laboratory studies may be scattered across different documents, while coordinates may be recorded using outdated reference systems.

According to the team’s estimates, a preliminary analysis of a single area may require four geologists working continuously for two to four weeks. The digital geologist is designed to perform the initial processing of such datasets within approximately one day. However, the system does not discover mineral deposits within a single day: it takes over the process of searching for and structuring information, while interpretation of the results remains the responsibility of specialists.

“We do not aim to replace geologists with AI. Our goal is to eliminate weeks of manual archival work for specialists. The system helps collect and verify initial data faster, while the final decision always remains with the professional”, Damir Kyzyr, Team Lead at Mineral Intelligence said.

Project team

The head of Mineral Intelligence, Damir Kyzyr, is a Kazakhstani engineer and entrepreneur with over 10 years of experience in the mining sector, subsurface use, and mineral exploration. He graduated from Satbayev University and obtained a Master’s degree in Petroleum Engineering from the University of Alberta under the Bolashak international scholarship programme. He has worked at Kazgeoinform, Aurora Minerals Group, Eurasian Resources Group, and NovaDrilling.

From an old report to a modern map

The team’s core product is an Operational Geological Brief, or Prospectivity Report. The system recognises archival documents, extracts information on drill holes, samples, and laboratory studies, converts historical coordinates into modern formats, and plots relevant objects on a digital map.

Every significant conclusion must be supported by a reference to a specific page, table, or archival document. This enables geologists to quickly verify sources and reduces the risk of artificial intelligence-generated errors.

The final report is designed for two audiences. Geologists receive detailed technical information, while executives and investors obtain clear insights into the area’s potential, the quality of source materials, and associated risks.

Global market

Artificial intelligence is already being applied in mineral exploration worldwide. The US-based company KoBold Metals uses machine learning and geological data to identify critical mineral resources. In early 2025, the company raised $537 mln at a valuation of $2.96 bln. While it is not a direct analogue of Mineral Intelligence, it demonstrates strong investor interest in the sector.

Mineral Intelligence is currently at an early stage, validating its hypotheses within the AI’preneurs programme. The next step is to test the solution on real-world projects.

The 4th cohort will conclude with a Demo Day for investors and corporate partners. The best-performing teams will have the opportunity to receive pre-seed investment through SAFE agreementsfrom Astana Hub Ventures and receive scaling support through Astana Hub’s international network.

Meanwhile, recruitment is currently underway for the sixth cohort of AI’preneurs, providing new participants with an opportunity to progress from an initial idea to launching their own AI startup.