# DefoAI: AI-assisted cyber risk quantification | Coresentry > DefoAI connects conversation, evidence and financial simulation to quantify cyber risk, understand uncertainty and compare security investments. Official page: https://www.coresentry.fr/en/solutions/ai-risk-analysis Contact: hello@coresentry.com ## DefoAI Available DefoAI helps you understand what a cyber incident could cost your business. An AI-guided conversation gathers context and evidence; reviewed assumptions then feed financial simulations to compare your options. A decision you can explain: which risks to address, with what budget and on which assumptions. - Resumable guided conversations focused on missing information - Document reading and suggestions linked to their sources - Financial scenarios based on approved assumptions - Annual loss simulations with visibility into severe events - Action comparisons with costs and estimated reductions - Traceable reports, exports and ownership tracking ## Talk about your business. Understand your exposure. You know your customers, your operations and what needs to keep running. DefoAI starts with that knowledge. Leaders, finance teams and business owners can contribute in their own words, without specialist cyber or statistical knowledge. The assistant structures the information; your analyst reviews the evidence and financial assumptions. ### A conversation that moves with you Describe your organisation and concerns. The assistant identifies relevant gaps, asks one question at a time and lets you resume later. You do not need to navigate a long checklist to find the right topic. ### Evidence you can revisit Your notes and documents can support the assessment. Each suggestion links to the passage behind it. You review and accept changes; missing information remains an explicit gap to investigate. ### Numbers from an explicit model AI helps with collection and understanding. A versioned calculation engine produces results from approved parameters. You can trace the sources, choices and limits behind each estimate. ### Understand what matters Define the decision, activities, countries and scope. Conversations and documents help identify dependencies, existing controls and evidence to complete. ### Build credible scenarios Define the incident and included costs. With your analyst, document low, likely and high estimates for frequency and loss, their sources and applicability to your company. Approve these assumptions before calculation. ### Compare and decide Simulate possible years, examine loss and uncertainty, then compare actions with their costs and estimated effectiveness. Assign owners, track progress and recalculate when assumptions change. ### A view of typical years and severe losses The median marks the middle of simulated years; the mean reveals the influence of expensive events. P95 is a threshold below which 95% of simulated annual losses fall, under the chosen assumptions. ### The assumptions that shape your decision Exceedance curves show how often simulated losses pass a threshold. Sensitivity analysis shows how changes in frequency or cost affect mean loss; convergence checks flag results that need review. ### A record to share and develop Sources, approved assumptions and model versions accompany the results. Printable reports and CSV and JSON exports support conversations across leadership, finance and security. Earlier results are preserved. ### Break risk down to explain it The approach separates incident frequency from cost, drawing on quantitative approaches such as FAIR. DefoAI currently uses a variant based on estimated total loss per event, allowing each scenario and assumption to be discussed separately. ### Make uncertainty visible Monte Carlo simulation explores many possible years from documented ranges. It can represent incidents clustered in some years and rare losses exceeding the estimated base cost. A single average cannot fully describe that exposure. ### Adapt to the actual context Sector, countries, activities and the observation period define the scope. Parameters are contextualised for your company, and their sources remain available for review. ### Learn from new evidence Bayesian network research examines how to connect risk factors and revise estimates as new evidence appears. DefoAI incorporates automatic Bayesian updates so estimates evolve with new information. ### Understand cascading effects Risk network research sheds light on dependencies across functions, suppliers and activities. DefoAI incorporates scenario dependencies and correlated losses to represent their combined impact on your business. ### Connect security with business consequences Recent work, including FABRICS (2026), combines cyber and business expertise to estimate incident consequences. This informs our intake method: teams contribute evidence and context; the analyst makes included costs explicit. DefoAI incorporates the FABRICS model to connect threats, control failures and financial consequences. ## References - [Wang, Neil & Fenton · FAIR et réseaux bayésiens / FAIR and Bayesian networks · 2020](https://doi.org/10.1016/j.cose.2019.101659) - [Orlando · Cyber Value at Risk · 2021](https://doi.org/10.3390/risks9100184) - [Sheth & Sinfield · Quantified Risk Networks · 2024](https://doi.org/10.1038/s41598-024-71764-x) - [Slapničar & Joshi · FABRICS · 2026](https://doi.org/10.1016/j.cose.2026.104900) - [Thevaratnam & Rezaeifar · Bayesian cyber threat analysis · 2026](https://doi.org/10.1007/s10207-026-01220-6) Financial estimates depend on approved data and assumptions; they do not predict the next incident. AI proposes information for review, and calculations remain under human oversight. Our literature review guides product development; it is not an independent validation of its results. ## Can I contribute without technical expertise? Yes. DefoAI gathers context through a guided conversation with accessible questions about activities, outages and costs. An analyst then reviews evidence and financial assumptions. ## Which methods does DefoAI use? DefoAI combines Monte Carlo simulation, the FABRICS model, Bayesian updating and scenario dependencies. Results remain linked to reviewable assumptions and sources. ## Which results can I examine? The assessment presents simulated annual losses, median and mean losses, severe-loss thresholds and exceedance curves. It also supports comparing actions with their costs and estimated effectiveness. ## Does AI make investment decisions for me? DefoAI helps structure and understand the assessment. Assumptions are approved, numbers come from the calculation engine and decisions remain under human control.