How the platform will work.

It takes in company data, ranks the opportunities, suggests position sizes within your limits and records every step.

Five steps, from data to decision.

Step 1

Collect

Gather company accounts, share prices, filings, results calls and news.

Step 2

Rank

An AI model scores each company and shows what drove the score.

Step 3

Size

Suggest how much to hold, within limits you set.

Step 4

Record

Log every suggestion and the checks behind it.

Step 5

Share

A dashboard for your team, and a data feed (API) for your systems.

These are planned. None is available to customers yet.

What goes in.

Share prices, company accounts, regulatory filings, results-call transcripts and news.

You choose which companies to cover, so the rankings fit what you can actually buy.

Good data comes first

On any past date, the model only sees what was known on that date. That stops it "cheating" with hindsight. We also check for gaps and late corrections.

Data licence costs are separate and not yet confirmed.

Two engines, kept apart.

1. The ranking engine

An AI model ranks companies by opportunity, adjusted for risk. It uses factors from Minerva Money Management's stock-picking process, the kind an experienced manager would recognise.

Each ranking shows what moved it up or down, and by roughly how much. That explains the answer. It doesn't prove it's right.

2. The sizing engine

This suggests how much to hold and when to rebalance, within limits you set: maximum position size, sector limits and minimum trading volume.

Why keep them apart?

You can check and improve each one without breaking the other. Limits help control risk, but they can't guarantee against losses.

What comes out.

Your team gets a dashboard showing rankings, reasons and limit checks in one place, a data feed (API) for your own systems, and a full log of every suggestion and review. All of these are planned and not yet available to customers.

People stay in charge.

Software supports decisions. People make them. You set the companies, the mandate and the risk limits, your oversight team watches how the system behaves, and you can pause or override it at any time, with the reason recorded.

Limits and security.

Bad data, changing markets, wrong models and system failures can all affect results. The software won’t remove bias, guarantee compliance or make outcomes certain. The platform will run in the cloud and is being designed with role-based access, separate client data, encryption and a log of admin actions.

Tell us about your process.

Which factors, limits and records would your team need? Talking now helps shape what we test first.

Discuss the platform  or  how working with us would look