Data

Introducing Falcon

Falcon is a model monitoring tool that guides you to the root cause one layer at a time. Every investigation feels like a story waiting to be discovered.

Falcon is an intuitive model monitoring tool, built by model developers for model developers. With so many components to track, finding what matters usually means juggling several filters and views at once. Falcon removes that cognitive overload, taking you to the answers one step at a time.

A guided journey

Filters, drop-downs and drill-throughs are now standard across model monitoring, and banks have built numerous dashboards with them. Even the best of them still leave you to work out what’s behind the deterioration. Falcon is different. It turns each investigation into a guided journey, and that journey is what makes people enjoy using it.

You begin at the top, with a clear overview of the whole portfolio. One area needs attention, so you open it. Falcon displays only the relevant charts, each chosen to answer the question you’ve just asked. When something needs a closer look, you step down a layer, to the model, then to its components. Each step reveals a little more, and each answer leads naturally to the next question.

Less searching, more thinking

This rhythm changes how monitoring feels. With only the charts that matter in front of you, there’s more room to think. Built with simplicity in mind, so you can work through each item properly before moving on to the next. The view is always contained, always relevant, and always pointing somewhere useful.

There’s also a quiet satisfaction in it, the feeling of navigating a well-laid path. Every click moves you forward, from IFRS 9 to LGD to a single component. Seeing how each step fits with the last, you know where you started, how far you’ve come and what’s left to explore.

When you reach the root cause, the path you took becomes a story. That story turns into slides for your committee with a single click.

That’s why users look forward to using Falcon. Monitoring becomes something you explore with curiosity, and a tool people enjoy is a tool they use often, which is exactly what good model oversight needs.

What sits underneath

Behind the simplicity sits real depth. Falcon goes from the whole model estate down to a single component, such as the recovery curve in an LGD model or the term structure behind a lifetime PD, which standard monitoring rarely reaches.

Throughout, Falcon applies your own thresholds and runs inside your systems.

From line charts to waterfalls, you can move a threshold or change an assumption and watch the results update in real time.

As part of every build, we set Falcon up to export the views you choose straight into a pack, in PowerPoint or whichever format your committee works in. Your team can write the commentary and narrative for each slide, or have the AI your bank has already approved draft it automatically and then edit it before it goes to committee.

Made for your team

Falcon is an example of what we build at Gini. Every tool we make is bespoke to your portfolios, your policy and your systems, and we hand it over, so your team owns it. We start with a single model, so your team sees the difference quickly before anything grows.

Book a demo

Sources

  1. IASB: IFRS 9 Financial Instruments, project summary. 2014. Records the G20's "too little, too late" concern that drove the move from incurred loss to expected loss.
  2. IASB: IFRS 9 Financial Instruments. 2021 consolidated edition, mandatory for IFRS reporters from 1 January 2018. Paragraph B5.5.43 defines what a 12-month allowance does and does not cover. View source ↗
  3. EBA: IFRS 9 implementation by EU institutions, 2023 monitoring report. November 2023. Source for the scenario-weight averages across the sample and for the observed dispersion in staging practice.
  4. EBA/GL/2017/06: credit risk management practices and accounting for expected credit losses. Effective 1 January 2018. Paragraph 135 on the 30-days-past-due backstop, and Principle 3 on the governance a post-model overlay must carry.
  5. Botha, Oberholzer, Larney and de Jongh (2023): Defining and comparing SICR-events for classifying impaired loans under IFRS 9. Tests competing SICR definitions against real mortgage data and finds the choice materially changes which loans transfer, and when.
Receive updates directly in your inbox

Stay connected