Not Every AI Feature Needs an LLM. Here’s What I Built Instead.

2026-07-12 · Source: Artificial Intelligence on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, FinTech & Digital Financial Services · Depth: Intermediate, medium

Summary

A portfolio copilot's Health Score feature provides a deterministic assessment of investment portfolios, assigning a score from 0 to 100. This score is derived from a weighted average of five sub-scores: diversification (30%), concentration risk (25%), ETF overlap (20%), volatility (15%), and expense efficiency (10%). Instead of using a Large Language Model (LLM) for scoring, the system employs a pure Python engine to ensure consistent, explainable, and testable results. Sub-scores utilize financial metrics like the Herfindahl-Hirschman Index (HHI) for diversification and weighted beta for volatility. The system also generates plain-English explanations and risk alerts based on specific thresholds, with an LLM then interpreting these deterministic outputs to provide personalized, natural-language narratives to the user.

Key takeaway

For AI Engineers building financial applications, prioritize deterministic algorithms for core calculations like portfolio health scores. Your teams should use pure Python for math, consistency, and explainability, reserving LLMs for natural language synthesis and interpreting these concrete results into personalized narratives. This approach ensures auditability and reduces hallucination risk, providing a robust and trustworthy user experience.

Key insights

Combine deterministic logic for core calculations with LLMs for personalized, natural-language explanations.

Principles

• Deterministic logic ensures consistency and testability.
• Domain-informed defaults enhance transparency.
• Hybrid AI leverages strengths of both approaches.

Method

Calculate a composite health score from five weighted sub-scores (diversification, concentration, overlap, volatility, expenses) using pure Python, then generate deterministic explanations and risk alerts, which an LLM can interpret.

In practice

• Use HHI for sector diversification scoring.
• Implement piecewise linear scoring for concentration risk.
• Generate deterministic explanations via templates.

Topics

• Hybrid AI
• Deterministic Algorithms
• Portfolio Analysis
• Financial Modeling
• Risk Management
• Python

Best for: AI Engineer, Data Scientist, Director of AI/ML

Related on AIssential

• Fin-Analyst at FinMMEval 2026 Task 3: A Live Hybrid Trading Agent with LLM Specialists and Rule-Based Signals — Hybrid LLM trading agents demonstrate strong equity performance but require memory and broader asset application.
• Making AI Financial Models Deterministic: A Skill That Generates Editable DCF Excel Workbooks — AI should automate mechanical tasks in financial modeling while preserving human control and model transparency.
• Evaluating Financial Intelligence in Large Language Models: Benchmarking SuperInvesting AI with LLM Engines — Evaluating financial intelligence in LLMs requires a multi-dimensional framework beyond simple factual accuracy.
• Artificial Intelligence in Ship Finance: Applications, Opportunities, and a Case Study in AI-Augmented Loan Origination — Agentic LLM systems can automate complex, data-intensive financial workflows, enhancing efficiency and auditability.
• Coinbase for Agents: Automating portfolio trading with AI — Coinbase for Agents integrates AI with financial execution channels to automate portfolio trading and payments.
• Toward Expert Investment Teams:A Multi-Agent LLM System with Fine-Grained Trading Tasks — Fine-grained task decomposition in multi-agent LLM systems significantly improves financial trading performance.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence on Medium.