Table of Contents
Portfolio Extension: Solving the Multi-Category Problem
Most organizations don’t compete in just one category. This creates an AI Portfolio Strategy challenge. A professional services firm may serve clients across strategy consulting, financial advisory, and technology implementation. A manufacturing company may compete in industrial equipment, replacement parts, and maintenance services. A media company may operate across news, entertainment, and analytics products.
Each of those is a separate AI representation challenge within an AI Portfolio Strategy. The queries your buyers ask about your strategy consulting practice differ from the ones about your technology implementation work. AI systems that find info about your industrial equipment might use entirely different ways of weighting data than those handling your maintenance services.
Theorem 3 — the Portfolio Extension of Byrum’s Law — applies to organizations in this situation. It says that managing AI representation as a portfolio isn’t just a practical convenience — it’s a formal requirement. Your total representation across all categories is a weighted sum. The weights come from how often buyers actually search in each category. Theorem 3 provides the mathematical foundation for AI Portfolio Strategy.
Why Category Prominence Matters for AI Representation
The big idea behind Theorem 3 is category prominence. Category prominence is a key metric in AI Portfolio Strategy. Every competitive category you operate in has a prominence weight — the fraction of all AI queries across your categories that falls into that specific one. A category where buyers search frequently has high prominence. One where they rarely search has low prominence.
Total AI representation isn’t the average of your per-category authority. It’s the prominence-weighted average. That matters a lot in practice. Imagine you achieve near-perfect representation in a low-prominence category but neglect a high-prominence one. Your total representation will be poor — even if that neglected category’s score looks fine on its own.
You can’t control category prominence. It’s determined by buyer behavior — how often your actual and potential buyers search in each category across AI platforms. And it changes over time. Markets evolve. New competitors and technologies shift attention. AI platform query distributions change with adoption patterns. Your best move is to monitor it and reallocate investment when it shifts significantly.
What Is the Portfolio Extension Protocol?
Theorem 3 gives a clear management approach for AI Portfolio Strategy. Monitor each category’s prominence every quarter using search volume as a proxy. Then trigger a reallocation review whenever any category’s prominence weight changes by more than 15 percentage points since your last review.
The satisficing allocation rule is straightforward: invest in each category’s AI representation in proportion to its prominence weight. If 60% of your buyers’ AI queries fall in your primary product category, roughly 60% of your AI representation investment should go there. If a secondary category’s prominence rises from 15% to 30% over two years, your investment allocation should follow.
This protocol sounds simple. Yet most organizations don’t implement it. They lack visibility into how their buyers’ query distributions are shifting across AI platforms. They invest based on internal priorities, not external buyer patterns. The result? Systematic under-investment in high-prominence categories and over-investment in low-prominence ones.
Cross-Category Spillover in Portfolio Extension
Categories that share vocabulary, buyer types, and competitive structure can show positive spillover. Authority signals you build in one category can reinforce your strength in adjacent ones. That’s the cross-category benefit. Managing spillover is part of an effective AI Portfolio Strategy.
If you’re deeply established in industrial equipment, your authority signals carry weight in industrial maintenance too. The vocabulary, buyer types, and competitive context overlap a lot.
But spillover can backfire when categories are too far apart in vocabulary and competitive structure. If you claim authority in both financial advisory and consumer food products, you create vocabulary and attribution conflicts. That can suppress AI representation in both categories. AI systems faced with contradictory signals from the same source tend to hedge — exactly what Theorems 1 and 2 are designed to prevent.
The formal rule for managing this risk is the cross-category correlation of authority signals. When two categories have high correlation, you can manage them together. The same signals, vocabulary, and corroboration sources are relevant to both. Investment in one helps the other. But when correlation is low, manage them separately to avoid conflicting attribution.
Investment Decisions Under the Portfolio Extension
The practical implications of Theorem 3 for investment decisions are direct. First, before allocating any AI representation budget, map your categories and estimate their prominence weights. The best available proxy is search volume across the queries your buyers use — this is available through standard search analytics tools. Build this map first; it determines how every subsequent dollar should be allocated.
Second, set a calendar reminder to update this map quarterly. Category prominence shifts. Markets that are AI-query-heavy today may consolidate or fragment. New categories may emerge as AI platform adoption changes what buyers search for. The allocation that was correct last year may be wrong today.
Third, audit your highest-prominence categories first. Whatever your current AI representation investment looks like, it’s probably misallocated. If you haven’t done a formal prominence-weighted assessment, you’re likely misaligned with actual buyer query weight. The highest-prominence categories give you the biggest return. Improvements there have the largest impact on total representation and commercial results. These investment decisions are the core of an AI Portfolio Strategy.
Theorem 3: Why It’s a Portfolio Extension, Not a New Law
It’s worth being clear about the formal status of Theorem 3. After rigorous adversarial review across 17 scientific committees, Theorem 3 got a VERDICT B: a formally correct extension of Theorem 1, not a new theorem on its own. That means every per-category representation requirement comes from Theorem 1 — each category must satisfy Theorem 1’s conditions independently. What Theorem 3 adds is the portfolio aggregation function and the prominence-weighting framework. These are the building blocks of a complete AI Portfolio Strategy.
This isn’t a demotion. A formally correct extension of Theorem 1 that tells the vast majority of organizations — those competing in more than one category — how to manage their AI representation is a big, practical contribution. The VERDICT B simply means the math comes from Theorem 1, applied across multiple categories with the portfolio aggregation function on top.
How Confident Are We in the Portfolio Extension?
Theorem 3 is theoretically strong — three independent mathematical derivations (Lyapunov stability theory, Little’s Law queuing theory, and Price equation analysis) all confirm the extension’s structure. Empirically, it depends on Theorem 1’s primary field confirmation, which hasn’t been executed yet. But the portfolio management protocol is actionable right now as part of your AI Portfolio Strategy. You don’t need to wait for empirical confirmation.
Next in this series: Theorem 4 — The Adversarial Extension. What happens when competitors launch coordinated attacks on your AI representation, and how to build defenses that cannot be overcome.
Joseph Byrum is an accomplished executive leader, innovator, and cross-domain strategist with a proven track record of success across multiple industries.
