Intelligent Operating Models

What It Is

The capability to design operating models where AI agents are first-class organizational participants — not tools bolted onto existing processes, but structural elements of how work gets decomposed, decisions get made, and value gets attributed. This covers agent POD design around Centers of Gravity (decision density nodes), process decomposition methodology, EBITDA attribution at the CoG level, and the formal proof that the 95% enterprise AI failure rate is an equilibrium problem, not a capability problem.

Why It Matters

Most organizations deploy AI into existing operating models. The result is narrative, not EBITDA. Redesigning the operating model around AI produces the opposite.

The dominant approach — point-solution automation layered onto existing processes — reduces labor cost in a task while leaving decision latency, coordination overhead, and exception rates unchanged. The operating structure was not designed to absorb AI capability. Decision flows assume human velocity. Coordination patterns assume human bandwidth. Exception handling assumes human judgment at fixed review cadences. Deploying agents into this structure accelerates individual tasks without addressing the structural bottlenecks that determine end-to-end throughput and value creation.

Platform investments create the opposite failure: technical capacity that goes unused because the operating structure cannot absorb it. Both approaches — point automation and platform investment — look good in board decks. Neither moves EBITDA. The Cost of Rationality paper proves why: multi-agent systems converge to Pareto-inferior Nash equilibria through perfectly rational behavior. Each agent optimizes locally. The system underperforms globally. This is not a technology failure — it is a coordination architecture failure that no amount of better models, better prompts, or better tooling addresses.

The Center of Gravity methodology identifies where decision density concentrates — 6-10 CoGs in mid-market companies versus 20+ in enterprises — and deploys agent PODs around those nodes. In PE-backed portfolio companies, the concentrated decision density makes this tractable within a 100-day operating review window. The methodology bridges strategy (what to transform) with operations (how agents actually run), with EBITDA attribution at the CoG level — not aggregate ROI, but specific value creation per decision density node.

Proof Points

Agentic OpModel: 8 presentations delivered, 18 HTML demos built, personal value chain pilot with 23 JSON canon schemas — the methodology is not theoretical but demonstrated across multiple delivery formats
Proforma Intelligence: 5-step diagnostic deployed — Steps 2-3 are deterministic TypeScript (zero AI calls for cost control), Steps 4-5 use AI Gateway for enrichment. Production-grade PE financial diagnostic
ATLAS: 6-stage methodology wizard (Intake → Processes → Recommend → Configure → Output → Review) with Fractal Canvas map and PDF export, deployed at entat.vercel.app — live delivery tool, not a prototype
CoG mapping: Tractable within 100-day PE operating review window — 6-10 CoGs in mid-market, 20+ in enterprise, each with EBITDA attribution
12 client engagements: PHH (6 Excel models, operating model analysis), Ares PE (value creation playbook, fund-level AI program), Navacord (2 agentic dashboards), NiSource (comprehensive dashboard), Fortis Finance (Agents vs. CustomGPTs brief, operating model), UES (RFI deck, business state of play)
R² = 0.850 WTP regression: 1,820 pricing opportunities analyzed — engagement type and buyer persona drive pricing more than practice ownership (empirically validated, not assumed)
Cost of Rationality proof: 95% enterprise AI failure rate is a Nash equilibrium, not a capability gap. Technology improvement alone cannot change the outcome
Coordination Failure formalization: three-layer crisis model with multiplicative compounding that predicts empirical success rates (adoption, integration, value capture each fail independently and compound)
Patent: USPTO 19/418,922

Market Position and IP

Patent-protected methodology (USPTO 19/418,922). The CoG-based transformation approach is structurally different from what the major consulting firms offer. McKinsey's Portfolio Excellence, BCG's Value Creation Plan, and standard PE operating partner toolkits treat AI as subordinate to the existing operating model — a capability to be deployed within existing decision structures. This methodology redesigns the operating model around decision density, deploying agent PODs at CoGs with EBITDA attribution per node.

The competitive distinction is formal: the Cost of Rationality paper proves that the dominant approach (better technology within existing operating structure) converges to Pareto-inferior equilibria. The intervention is not better models — it is better coordination architecture. No competing framework provides this formal proof, which means competitors are optimizing within a paradigm that the mathematics shows cannot produce the outcome they promise.

Production evidence: Proforma Intelligence deployed (5-step diagnostic with deterministic + AI hybrid), ATLAS deployed at entat.vercel.app (6-stage methodology wizard), 6 Excel models delivered for PHH alone, value creation playbook for Ares PE. The methodology has been delivered to 12 clients across PE, utilities, insurance, and financial services.

Novel Research Contribution

Three papers formalize different aspects of intelligent operating model design:

The Cost of Rationality proves that the 95% enterprise AI failure rate is an equilibrium problem, not a capability problem. Multi-agent systems converge to Pareto-inferior Nash equilibria through rational local optimization. The formal contribution: demonstrating that technology capability improvement (better models, better data, better infrastructure) cannot change this equilibrium — only coordination architecture redesign addresses the actual failure mode. Target venue: Management Science.

The Coordination Failure paper formalizes the three-layer crisis model: adoption, integration, and value capture each have independent failure probabilities that compound multiplicatively. If each layer has 70% success probability, end-to-end success is 34% — matching empirical failure rates. The contribution: converting "AI projects fail" from an observation into a predictive model with intervention points at each layer. Target venue: Strategic Management Journal.

The PE AI Thesis introduces Center of Gravity methodology as a PE-native value creation framework — not a scaled-down enterprise AI playbook, but a methodology designed for the concentrated decision density and compressed timelines of PE-backed portfolio companies. Target venue: Journal of Private Equity.

Implementation and Impact

Clients receive an operating model diagnostic and redesign. The typical engagement follows a structured sequence:

Phase 1 — Diagnostic (2-3 weeks): CoG mapping identifies where decision density concentrates. The Proforma Intelligence app runs the financial diagnostic (5-step pipeline). Output: CoG map with decision density quantification, current-state EBITDA attribution, and identification of where AI agent deployment will create measurable value versus narrative.

Phase 2 — Design (3-4 weeks): Agent POD configuration per CoG. ATLAS delivers the methodology wizard. Process decomposition identifies which decision flows redesign around agent velocity. Output: target operating model with agent POD specifications, EBITDA attribution per CoG, and implementation sequencing.

Phase 3 — Attribution (ongoing): CoG-level EBITDA tracking through the change climate framework. Leading indicators (5 climate dimensions) provide 2-3 quarter advance signals of financial trajectory shifts.

For PE operating partners, the full diagnostic fits within a 100-day operating review window. Measurable outcomes: EBITDA impact attributed at CoG level (not aggregate), decision latency reduction quantified per CoG, and coordination overhead eliminated through operating model redesign rather than technology improvement.

Connections

Builds: Proforma Intelligence, ATLAS, Practice OS
Frameworks: Agentic OpModel, Process Dictionary, Proforma Engine, Value Chain
Papers: cost-of-rationality, pe-ai-thesis, coordination-failure, balancing-intelligence, leadership-imperative
Imperatives: Fractal Design, 10x10 Domain Intelligence, Exploit-Proofing Triad
Capabilities: Financial Value Creation, Agentic System of Systems
Clients: PHH, Ares PE, Navacord, NiSource, Fortis Finance, UES