Computer Vision · Fintech

AI Coin Counting Boosts ATM Efficiency

SLM Holdings needed a groundbreaking non-mechanical solution to modernize ATM banking. We delivered a high-speed vision system that replaced outdated hardware with real-time AI coin identification and tracking.

Client
SLM Holdings
Impact
Investor Funding
Start your project
SLM Holdings AI coin counting component for ATMs
Identification Accuracy
99%+
real-time coin tracking
99%+
Counting accuracy
real-time coin tracking
100%
Zero-mechanical
non-mechanical vision system
Instant
Display speed
optimised processing algorithms
Unlimited
Coin types
easily adaptable to any currency
The Challenge

Revolutionising banking required modernising hardware

SLM Holdings is on a mission to redefine the digital banking landscape. However, traditional ATMs rely on mechanical coin counting systems that are outdated, prone to breakdowns, and limited to just four coin types. They needed a non-mechanical solution to count money faster and more accurately.

Outdated mechanical systems
Existing ATMs were limited to just four coin types, leading to decreased utility and user frustration.
High maintenance & downtime
Mechanical coin counters are prone to breakdowns and require complex rebuilds when currency formats change.
Need for non-mechanical vision
SLM Holdings wanted to eliminate mechanical failures by using a camera for faster and more accurate counting.
Product validation for funding
The client needed to validate their product ideas and hypotheses to attract investor funding.
The Solution

Built for the edge — high-speed, camera-driven, perfectly seamless

Rather than accepting mechanical limitations, Gradient Insight developed a coin counting component using AI and Computer Vision. We implemented a customised object detection model that tracks coins in real-time under varying lighting conditions and angles, optimised for instantaneous display.

Implementation of a customised object detection model for accurate coin tracking in real-time

Optimisation of processing algorithms to meet the demands of real-time recognition

Robust performance across diverse lighting conditions and coin orientations

Selection and integration of embedded hardware: microcontrollers and high-speed cameras

Creation of an intuitive PiTFT touchscreen interface for seamless user control

End-to-end integration with the existing ATM system architecture

The Results

Outcomes that matter — from proof of concept to funded reality

The deployed system didn't just solve the mechanical failures — it provided the necessary validation for SLM Holdings to secure investor funding and accelerate their product development.

99%+
Accuracy rate achieved
Accurately identifying and tracking different coin types in real-time under varying lighting conditions and angles.
Zero
Mechanical breakdowns
The non-mechanical vision-based system eliminated the primary source of ATM maintenance and downtime.
Secured
Investor funding
Validated product hypotheses allowed SLM Holdings to secure funding and ramp up product development.
High
User satisfaction
Faster and more accurate coin counting led to a notable increase in user satisfaction with the ATM experience.
"Their expertise, paired with strong communication and collaboration skills, provided our team with great confidence in their ability to assist with our project and deliver high-quality results."
Andre Battles
Andre Battles
SLM Holdings
Technology Stack
Computer VisionObject DetectionReal-time TrackingEmbedded SystemsPiTFT TouchscreenCustom AlgorithmsMicrocontrollers
At a Glance
Client SLM Holdings
Industry Banking / Fintech
Delivery AI Coin Counting Component
Accuracy 99%+ validated
Status Secured Funding

Similar challenge?

We build bespoke AI systems for complex, high-stakes domains — from day one.

Book a free discovery call
Watch the Case Study

See the solution in action

"Gradient Insight delivered a robust vision system that transformed our ATM prototype. The accuracy and real-time performance were critical for our product validation and market differentiation."
Andre Battles
Andre Battles
Fintech Lead, SLM Holdings
Ready to build?

Your domain, your data — built to your exact requirements

We don't adapt generic models. We architect bespoke AI systems that meet your performance, licensing, and infrastructure constraints — from scratch.