Get to know our methodology

Clear, value-driven process

We use AI to transform dynamic market data into actionable, unbiased recommendations. Our approach centres on analytical transparency and respect for user privacy under South African law. All methods are detailed so users can review data sources, oversight, and understand that results may vary. Past performance is not indicative of future outcomes.

How we process your data

We analyse public and permitted proprietary data streams to identify patterns and relevant signals. Our AI applies multiple model layers to extract key insights for timely recommendations.

Data is processed securely and in accordance with POPIA. Personal information is never sold or transferred to unauthorised parties. Information is only retained for as long as necessary for service delivery.

All system outputs undergo continual review using feedback loops and independent audits. This ensures our signals remain reliable, unbiased, and transparent. However, decisions based on these outputs remain user responsibility.

AI model process flow illustration

The human side of automation

Thandi Nkosi

Thandi Nkosi

Director of Analytics

"Our mission is to provide accessible, transparent automated recommendations supported by reliable data—empowering users to make decisions that suit their unique needs. While we stand by our process, outcomes always depend on the user's choices."

1

Jan 2024

Foundational data framework

Initial collection and verification of diverse market data sources, with strict POPIA-aligned standards from the outset.

2

May 2024

AI methodology launch

Our analytical models go live, offering transparent market signal generation and robust user privacy controls.

3

Oct 2025

Oversight process expanded

Feedback-driven updates introduced, enhancing the reliability and auditability of signal recommendations for every user.

4

Feb 2026

Continuous improvement

Incorporation of new data sources, ongoing client feedback, and regulatory changes to keep our process up-to-date.