# Portfolio Analytics with AI Agents > Financial Services · AI Course · FIN-8 > Source: https://ibl.ai/solutions/financial-services/course/portfolio-analytics-with-ai > Last updated: 2026-08-25 **AI across performance attribution, benchmarking, and reporting — with the numeric verification that keeps a generated figure out of a client statement.** ## The Short Answer **In portfolio reporting every number must come from a deterministic calculation, never from generation, and the two must be architecturally separate. ibl.ai runs analytics inside the firm where you own all the code and the data, with computed figures and generated narrative kept apart and lineage traceable end to end.** On ibl.ai you own all the code and the data, run it model-agnostic across any LLM, and pay with no per-seat pricing — so you can deploy anywhere, from your own cloud to a fully air-gapped network. [Request Access](https://ibl.ai/contact) · [Explore Financial Services](https://ibl.ai/solutions/financial-services) ## Course facts - **Level:** Intermediate - **Duration:** 5 hours across 8 modules - **Format:** Technical workshop with reporting labs - **Modules:** 8 - **Catalog code:** FIN-8 - **Frameworks covered:** GIPS, SEC Marketing Rule, FINRA communications rules ## What is this course about? Portfolio analytics is computation plus narrative, and the two must be kept strictly apart. This course covers where AI genuinely helps — attribution narratives, benchmark commentary, risk reporting — while enforcing that every number comes from a deterministic calculation with lineage traceable from source system to client statement. ## Who is this course for? - Performance and attribution analysts - Portfolio reporting teams - Investment operations - Client reporting and communications staff ### What do I need before starting? - Performance measurement or reporting experience - Familiarity with your reporting systems ## What will I be able to do afterwards? - Separate deterministic computation from generated narrative architecturally - Generate attribution narratives from computed results - Produce benchmark and peer analysis at scale - Personalize client reporting within compliance limits - Trace lineage from source system to client statement ## What does each module cover? ### Module 1 — Where must the arithmetic stay deterministic? The architectural separation between computation and narrative, and why it is not negotiable. _(45 min)_ **Objectives** - Separate computation from narrative architecturally - Identify every place a number could be generated - Design the separation into the system **Topics:** Computation versus narrative · Generation risk points · Architectural separation · Enforcement **Activity:** Audit a reporting workflow for every point a number could be generated rather than computed. ### Module 2 — How do you generate an attribution narrative? Explaining computed attribution results in language a client understands. _(50 min)_ **Objectives** - Generate narratives from computed attribution - Ensure the narrative matches the numbers - Explain at the right level for the audience **Topics:** Attribution narratives · Number-narrative consistency · Audience level · Verification **Activity:** Generate attribution narratives and verify each statement against the computed results. ### Module 3 — How do you handle benchmark and peer analysis? Comparative analysis at scale with correct benchmark selection and honest comparison. _(45 min)_ **Objectives** - Select appropriate benchmarks defensibly - Produce peer comparisons at scale - Avoid misleading comparison **Topics:** Benchmark selection · Peer comparison · Comparison honesty · Scale **Activity:** Produce benchmark commentary for a portfolio set and check for misleading framing. ### Module 4 — How do you narrate risk and stress results? Risk reporting narratives that convey uncertainty accurately. _(45 min)_ **Objectives** - Narrate risk metrics without false precision - Convey uncertainty honestly - Explain stress test results accessibly **Topics:** Risk narration · False precision · Uncertainty communication · Stress results **Activity:** Narrate a stress test result and check it for false precision. ### Module 5 — How much can you personalize a client report? Personalization within the compliance limits that govern client communication. _(40 min)_ **Objectives** - Personalize within compliance limits - Avoid crossing into individualized advice - Maintain consistency across clients **Topics:** Personalization limits · Advice boundary · Consistency · Review requirements **Activity:** Personalize a report and check it against the advice boundary. ### Module 6 — What is the verification gate? The check that makes it impossible for a generated number to reach a client. _(45 min)_ **Objectives** - Implement a numeric verification gate - Verify every figure against its calculation - Make the gate non-bypassable **Topics:** Verification gate · Figure verification · Non-bypassable design · Exception handling **Activity:** Implement the gate and attempt to publish a generated figure. ### Module 7 — How do you trace lineage to the statement? Data lineage from source system through calculation to the number on a client statement. _(45 min)_ **Objectives** - Implement lineage from source to statement - Reconstruct any figure's derivation - Support audit and client inquiry **Topics:** Lineage implementation · Derivation reconstruction · Audit support · Client inquiry **Activity:** Trace one client statement figure back to its source system records. ### Module 8 — Building the reporting agent The lab module: a reporting agent where numbers come from calculation, not generation. _(50 min)_ **Objectives** - Build the reporting agent with strict separation - Verify no figure is generated - Measure production time savings **Topics:** Agent build · Separation verification · Time measurement · Quality review **Activity:** Build the agent and prove no number in its output was generated. ## What is the capstone project? **Client reporting agent with verified numeric integrity.** Build a portfolio reporting workflow with architectural separation of computation and narrative, verified attribution narratives, defensible benchmark commentary, compliance-limited personalization, a non-bypassable verification gate, and full lineage to the statement. _Deliverable:_ A working reporting agent with proof that no published figure was generated. ## How are learners assessed? - Verification gate tested — a generated figure must be impossible to publish - Lineage test — trace a statement figure to source records - Narratives verified statement by statement against computed results ## What ships with the course? - **Facilitator guide.** Session-by-session running order, discussion prompts, and the questions that reliably derail a room. - **Learner workbook.** Exercises, checklists, and the templates each module's activity produces. - **Hands-on lab environment.** A sandboxed ibl.ai deployment so exercises run against real agents, not screenshots. - **Assessment bank.** Scenario questions and rubric criteria mapped to each stated learning outcome. - **Source bibliography.** Every primary regulation and standard cited on this page, linked and dated. ## Which AI agents does this course use? - [Portfolio Analysis Agent](https://ibl.ai/solutions/financial-services/agent/portfolio-analysis-agent) - [Client Advisory Agent](https://ibl.ai/solutions/financial-services/agent/client-advisory-agent) - [Operations Agent](https://ibl.ai/solutions/financial-services/agent/operations-agent) - [Compliance Agent](https://ibl.ai/solutions/financial-services/agent/compliance-agent) ## Where does the course material come from? Every module is grounded in primary sources — the regulation, standard, or research itself, not a summary of it. Each was resolved at authoring time. - [U.S. Securities and Exchange Commission](https://www.sec.gov/) — SEC. Performance presentation and marketing rule requirements. - [Artificial Intelligence](https://www.finra.org/rules-guidance/key-topics/artificial-intelligence) — FINRA. Guidance on AI in client-facing analytical output. - [Supervision and Regulation Letters](https://www.federalreserve.gov/supervisionreg/srletters/srletters.htm) — Federal Reserve. Model risk expectations applying to analytics used in reporting. - [AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework) — NIST. Verification and monitoring structure for the reporting pipeline. ## Delivery notes Binding guidance for anyone preparing and delivering this course: - Module 1's separation is the entire course. A generated number in a client statement is a regulatory and reputational failure, and the architecture must make it impossible rather than unlikely. - Module 6's gate should be tested adversarially. Participants should genuinely try to get a generated figure through, and the ones that succeed are the design flaws. - Module 4's false precision problem is subtle and important. Models narrate uncertainty poorly by default and will state a risk figure with unwarranted confidence. - Use synthetic portfolios with known correct attribution. Verification requires ground truth and real portfolios cannot provide it in a workshop. - Coordinate with FIN-5 — personalization limits appear in both and should be consistent. ## Why run AI training on a platform you own? - **You own the course, not a licence to it.** Course content, learner data, and the platform run inside your perimeter — you own all the code and the data. - **Model-agnostic delivery.** Run the course's AI components on any LLM — Claude, GPT, Llama, Gemini, Command — and switch anytime. - **No per-seat training licences.** Usage-based or self-hosted, so cost tracks actual use rather than headcount. - **Deploy anywhere.** Cloud, private VPC, on-premise, or fully air-gapped — including for cohorts that cannot use public AI tools. ## Frequently asked questions ### What does the Portfolio Analytics with AI Agents course cover? Portfolio analytics is computation plus narrative, and the two must be kept strictly apart. This course covers where AI genuinely helps — attribution narratives, benchmark commentary, risk reporting — while enforcing that every number comes from a deterministic calculation with lineage traceable from source system to client statement. It runs 5 hours across 8 modules across 8 modules, at intermediate level, and closes with a capstone: Client reporting agent with verified numeric integrity. ### Who should take Portfolio Analytics with AI Agents? It is written for Performance and attribution analysts, Portfolio reporting teams, Investment operations, Client reporting and communications staff. Prerequisites: Performance measurement or reporting experience; Familiarity with your reporting systems. ### Can we run this course on our own infrastructure? Yes. ibl.ai is model-agnostic and deploy-anywhere — cloud, private VPC, on-premise, or fully air-gapped — and you own all the code and the data. Cohort data, submissions, and any material learners upload stay inside your perimeter, which matters for financial services teams that cannot send work to a public AI tool. ### How do we get access to Portfolio Analytics with AI Agents? Request access and we will set it up for your cohort — hosted by ibl.ai, or running against your own deployment. Tell us the group size and timing you need, and whether it should run inside your own perimeter. ### How much does AI training for financial services cost on ibl.ai? There is no per-seat pricing — you pay for usage or self-host and pay only for the infrastructure, so a 5,000-person rollout does not cost 5,000 licences. 1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University. ## More Financial Services courses - [AI Model Risk Management for Financial Institutions](https://ibl.ai/solutions/financial-services/course/ai-model-risk-management): Extend model risk governance to generative AI — inventory, validation, challenger testing, and the documentation examiners expect for a non-deterministic model. - [KYC and AML with AI: Screening, Alerts, and SAR Support](https://ibl.ai/solutions/financial-services/course/kyc-aml-with-ai): Apply AI across the BSA/AML program — name screening, alert triage, and narrative drafting — without weakening the audit trail a regulator will examine. - [AI Supervision Under FINRA and SEC Recordkeeping Rules](https://ibl.ai/solutions/financial-services/course/ai-supervision-finra-sec): Supervise AI in a broker-dealer or RIA — communications review, books and records obligations, and what happens when an agent talks to a client. - [Fraud Detection with AI: Anomalies, Alerts, and False Positives](https://ibl.ai/solutions/financial-services/course/fraud-detection-with-ai): Build AI-assisted fraud detection where a false positive is a blocked customer — anomaly detection, adaptive fraud, and fair-lending exposure. - [AI for Client Advisory Without the Compliance Risk](https://ibl.ai/solutions/financial-services/course/ai-client-advisory-compliance): Research and client content generation inside a regulated advisory business — sourcing, review workflow, disclosure, and the line before personalized advice. - [Regulatory Reporting Automation: SOX, PCI DSS, and Audit Trails](https://ibl.ai/solutions/financial-services/course/regulatory-reporting-automation): Automate regulatory reporting and control testing with AI — evidence collection, narrative drafting, and a control environment that keeps the automation auditable.