Kore Geosystems’ cloud-based platform uses AI to extract geological and geotechnical data from core imagery to improve workflow efficiency for geologists. The system processes both new and legacy imagery, allowing users to view, analyze, and export results directly into existing workflows.
Our team faced the challenge of executing smooth Windows-to-web migration, upgrading core features and strategically showcase AI integration to attract new market segments.
How Logging fits after organizing images
Following image upload and organization, users transition to core rock analysis utilizing AI capabilities. I spent time learning how geologists track their data the traditional way and how they use our legacy app. I found out our AI technology is still quite new and is being progressively trained to enhance its predictive capabilities.
My job involved making sure the AI technology supports geologists in their logging activities, while making sure the users could still step in and clean up any errors from the AI's predictions.
Who are the main users?
After operators handled image uploads and organization, I researched who primarily uses our Logging system. By speaking with subject matter experts and our sales and support teams, I identified two main proto-personas: Loggers and Reviewers.
Loggers
Loggers split time between core sheds, temporary field offices, and the lab, and is open-minded about learning and adopting new tools.
They carefully examine the details and make corrections to any errors identified in the AI predictions.
Needs and Goals
Needs to efficiently log large volumes of core samples with consistent accuracy
Wants to build professional reputation through quality work and reliability
Aims to minimize manual data entry errors that could affect analysis
Wants to learn industry best practices through the tools they use
Seeks to demonstrate value to senior team members through efficient work
Motivations
Career advancement and professional development
Learning from more experienced geologists
Contributing meaningful data to exploration projects
Being recognized for thoroughness and attention to detail
Interest in using modern technology to improve traditional processes
Pain Points
Struggles with inconsistent terminology between different logging systems
Frustrated by software that crashes or loses data during field use
Feels pressure when logging speed expectations conflict with accuracy requirements
Limited training on software tools leads to underutilization of features
Worries about making mistakes that more experienced loggers would catch
Challenges with connectivity in remote locations affecting real-time data upload
Reviewers
Reviewers operate mainly from office settings, with occasional field trips to monitor drilling operations.
They focus on the big picture, supervising data across various projects to maintain consistent quality standards.
Needs and Goals
Needs to efficiently review and validate core logs from multiple field teams
Wants integrated data systems that connect logging to modeling and resource estimation
Aims to make timely, data-driven decisions about exploration activities
Needs to ensure compliance with reporting standards and company protocols
Wants tools that facilitate collaboration between field teams and office-based analysts
Seeks to minimize time spent on data formatting and maximize analysis time
Motivations
Improving exploration success rates through better data management
Reducing exploration costs while maintaining quality
Mentoring junior staff to develop their skills and judgment
Building organizational knowledge through standardized data collection
Advancing projects from exploration to resource definition efficiently
Pain Points
Time wasted reconciling data formats between different systems
Difficulty tracking the progress of multiple drilling programs simultaneously
Frustrated by inconsistent data quality from different logging teams
Struggles with software that doesn't integrate well with other corporate systems
Customer feedback
While gathering feedback on Ingest, I also reviewed our original Windows logging software. While users appreciate the efficiency from AI-assisted logging, several concerns have been identified.
There is not enough context when looking at the UI.
The AI makes a lot of wrong predictions, around 10% of the project.
They need more flexibility in the UI to accommodate their desired information.
Loggers analyze images sideways, while reviewers look at graphs up and down.
Loggers want more collaborative features, such as note sharing and comparing logs.
Reviewers want to see an overview of the projects.
Reviewers want integration to their other geology-related software.
Product goals
In addition to customer feedback, our experts and Senior Director wrote a thorough documentation of our goals on the Confluence platform.
Most archived images and logging systems are misaligned, making them AI-unfriendly. We need to guarantee that logging data remains properly synchronized with the corresponding images throughout our product.
Enable remote logging so geologists can log core from the comfort of their homes using our software and Ingests' high-resolution images.
Speed up the core logging process by making data entry much faster and more enjoyable.
Feature comparisons
Since AI in the mining industry is still emerging, I analyzed our competitors to identify the key features they're offering to our target market.
MX Deposit is a geological modeling and resource estimation software developed by Seequent, a company specializing in geoscience software solutions.
Their platform has the best integration for other popular geology-related apps.
They also provide different features and pricing for loggers and reviewers.
Even though they do not use AI or scanned images for their logging data., they provide more advanced features and calculations to cover a larger market.
Half of their UI is using a table to log geological data, while the other half is showcasing the graphs and visual data.
Datarock is a company specializing in providing advanced geological data analysis and machine learning solutions for the mining and exploration industries.
Their AI-driven platform is designed to help geologists and mining professionals extract valuable insights from geological data more efficiently and accurately.
They have very similar services to ours and GeologicAI. They also utilizes imaging with OCR and AI to assist geologists.
Their content framework is similar to our current app where the top part is used for navigation and tools.
The main content shows rows of rock cores and designated lines of logging data.
There is a left sidebar to edit logs and annotate images.
GeologicAI is a technology company that specializes in providing advanced geological data analysis solutions, particularly for the mining and mineral exploration industries.
Their logging solution leverages artificial intelligence (AI), machine learning (ML), and high-resolution imaging to automate and enhance the process of geological logging.
Their UI framework inspired our experts to follow their layout for future designs.
The left half of the screen consists of geological logs and other data that are useful for reviewing.
The middle section shows the scanned images of the rocks for context and more details.
The right sidebar has all the settings and adjustments needed for logging and analysis.
Sketching concepts
Although loggers analyze rocks horizontally while reviewers scan data vertically, the responsibilities of these two roles are gradually converging in the field today. I pitched several ideas to the team about mixing these two approaches for geological data.
The original design prioritized logging data, while minimal reviewing features were on another page.
When logging and reviewing share a horizontal layout, loggers must scroll extensively, introducing workflow friction and reducing efficiency.
Vertical layout for both logging and reviewing data is very effective only when the AI doesn't make any wrong predictions.
Combining logger and reviewer features
While our AI significantly accelerates work processes for users at every level, we recognized that AI accuracy improvements are still needed. The product team has determined that combining logger and reviewer functionality represents the optimal middle ground.
This strategy increases feature bloat and complexity but effectively addresses the requirements of both our key user personas.
Initial MVP release
In November 2022, our team launched the Logging feature MVP. While it addressed many logger concerns, it still fell short of meeting reviewer requirements.
Ongoing updates
For the next year, we added features for new users, such as help tips, rulers and measurements, more UI context on data, better zoom functionality, comparing logs, sharing notes and descriptions, graphs and tables for reviewers, and other minor UI adjustments.
Our AI has also improved dramatically, causing loggers to gradually transition to reviewer roles. As their need for traditional logging diminishes, they increasingly request more editing capabilities to correct minor AI errors.
After a year of improvements, we captured additional market share and strengthened relationships with our existing customers.
Support Ticket Reduction
Improved monthly geological data exports to third-party apps and labs within 6 months.
Feature Adoption Usage
Grew monthly active users from fewer than 50 to several hundreds in less than two years.
Future and thoughts
As I pointed out earlier, as AI capabilities improve, the logger's role is evolving from active logging to primarily reviewing. We should be designing to minimize logger features while enhancing reviewer capabilities. But since AI improvement is a slow, resource-intensive process, our current balanced but complex product meets our market's needs well enough.
I've learned that developing a highly specialized product requires significant time investment in understanding the subject matter. I had to review numerous geology documents to properly understand what we're offering. Connecting more closely with our new customers would have been a fantastic opportunity, as their invaluable and fresh insights continue to shape and enhance our product in exciting ways.
