Blue Fire AI

Frequently asked questions

BlueFire AI covers listed companies only. Approximately 10,000+ companies globally including 1,000+ Chinese onshore companies.

Emmalyn machine reads 10,500+ news wires, bulletin boards, industry journals, exchange announcement and publicly available media. Emmalyn scans the web daily to ensure important information is included in the analysis of the target company. In addition, if our clients have specific URL’s that inform their investment process we can include these sources into Emmalyn’s content feed.

Inputs include:

1) English/Mandarin reading

2) Forensic Balance Sheet and Cashflow Analysis (earnings quality, asset quality, etc.)

3) Street Perception and Herd Analytics

4) Liquidity and Transaction Profiling (such as collapse risk, historical liquidity trends, short pressure, etc.).

Incumbent risk systems are primarily point-in-time, based on linear rules based logic.  BlueFire AI provides forward-looking (predictive) triggers based on probability thresholds and are non-linear by design. Our signalling typically provides from 6 to 9 months information value.

No. Data availability and source design are different from that of listed companies.

We read Chinese-centric media (in Mandarin) as part of the 10,500+ news wires (and counting) daily.  However, due to strict licensing content agreements we are unable to surface these articles to the user. Scraped information is factored into the company’s unique risk sentiment index.

Significant historical data is available for the purpose of signal validation.

BlueFire AI’s core objective is to identify any event or piece of information that will likely impact a company’s intrinsic value. ESG is an increasingly important aspect of a company’s performance and risk exposure. Unlike ESG data vendors, we do not focus on scoring a company, instead we aim to understand the future impact of negative ESG events in the context of a company’s wider operations as and when they occur. 

No. BlueFire AI provides a high degree of context in our Risk Narrative feed illuminating the driver of the signal. In addition we compare and visualise the history of the company and its peers. This level of context is innovative in Market Risk.

Emmalyn is a deep-learning neural model. She is constantly learning and evolving and is not rules-based. She searches for “context” to the risk, benchmarked to the 000’s of companies under her coverage.

BlueFire AI signals and contextual information in relation to the driver of the signal provides the starting point of your risk investigation. Our role is to point our clients in the direction of where additional investigation may be required to validate (or invalidate) the signal.

Upon receiving a signal from Emmalyn, investigate the UI and ask:

    1. Is there a known fundamental issue in this company that could be crystallising into the Asset Price breakdown/Credit Spread widening?
    2. Is there a new/emerging issue that could also be relevant alongside the event risk?
    3. Revisit fair-value model and recovery rate assumptions to assess potential impact to expected performance.
    4. Actions post review – Keep/Add, Trim/Hedge, Exit, Do Nothing.

Note: BlueFire AI provide a data driven “outlook” on the target company. Importantly, this is not a “view” as we make no subjective judgement on whether the company is good or bad. The data drives the insight. We simply flag anomalies and abnormal patterns.

No. Emmalyn is founded in probability and enriched by oceans of data. Her insights are non-linear by design. Emmalyn “thinks” cognitively mimicking the human brain’s processing of “new” information to inform decision-making. 

Both. Emmalyn looks at a company holistically from a risk perspective. She is not biased and is sector-agnostic providing a comprehensive forward-looking 360-degree view of the unforeseen risk (to the extent they exist) in the target company.

Analysts are anonymised. That is how BlueFire AI receive the raw data. The smart consensus chart provides context regarding historical accuracy of the analyst leader cohort.

BlueFire AI are not able to provide underlying data due to licensing agreements with our various data providers (e.g. FactSet, S&P, S3, etc.).

No. Neutral simply means Emmalyn does not see any unpriced risk at that moment in time i.e. the market price reflects known risks. Equally, a move in status from “Severe” to “Watch” is not a positive indication in our framework. There may be information value in these moves, however, we have not performed any degree of statistical testing on this element.

Risk-as-a-Service: pricing is based on the number of companies covered. A base cost per month plus additional cost per company per month with substitution available, ie: clients ability to add and delete securities from their coverage lists.

Active-Investment-as-a-Service: negotiated on a case-by-case basis.

Risk-as-a-Service: pricing is based on the number of companies covered. A typical contract is 24 months, with walk-away gates addressing the trial component of 6 months. Coverage begins at USD $50 per company/per month with a minimum per month cost of USD $5,000.

Active-Investment-as-a-Service: negotiated on a case-by-case basis.

BlueFire AI define Peer Group as a function of geography and sector.

Opinions differ significantly from person to person. If however you value a different perspective, talk to Emmalyn.
Perspective vs Opinion – Probability vs Certainty.

If less than 3 analysts cover the company, BlueFire AI do not consider this cohort to portray bias-free insight. We will not surface the earning outlook.

BlueFire AI require a minimum of 3 analysts with an historical accuracy of better than 70% in forecasting EPS.

BlueFire AI treat these companies as two separate companies/entities and may have different risk statuses on each satellite, a function of market dynamics in the respective markets.

Emmalyn reads at sentence level, not article level. Emmalyn is trained to “event extract” and is able to recognise multiple “events” in the one article.

This depends on the data set. For structured financial data it is a lengthy time frame, typically 10-15 years. For events (i.e. news, media), it is closer to 3-5 years. Faster moving data (i.e. market data) 1-2 years. Herd conviction (Holdings) data 4+ years.

Blue Fire AI