L1.6 Behavioral Intelligence

The Maxim Moat

Any prompt library can name behavioral frameworks. Maxim enforces them. That single distinction is the entire architecture of this pack.

Most AI writing tools optimize for fluency. Maxim optimizes for behavioral effect: the psychological structure that turns a sentence from informational into actionable. AI writing tools do not fail because they lack frameworks. They fail because the frameworks they cite are never enforced.

The replication barrier is not framework count. It is the registry. Anti-patterns are catalogued. Critiques are noted against each framework. Outputs are scanned against the Fogg Behavior Model (B.J. Fogg, 2009) B=MAP composition before ship, via the behavioral_audit MCP hook that runs against every external-facing paragraph. Competitors can list 64 frameworks. They cannot replicate the enforcement without rebuilding the pack-engine audit mechanism, and they cannot replicate the audit without the anti-pattern ledger that makes every scan non-trivial. Enforcement is the moat. Everything else is literature review.

Business Outcome

Behavioral Intelligence applies Fogg Behavior Model (B=MAP) as the compositional anchor across every Maxim output. The mechanism: Motivation, Ability, and Prompt are each addressed explicitly before a sentence is allowed to ship; the 77 other frameworks overlay as modifiers on one of the three levers. Outcome claims accumulate in the living ledger at documents/ledgers/MOAT_TRACKER.md § MOAT-06.

Primary measurement metric: framework-tagged output shipping rate. Target post-v1.0.0: at least 95% of shipped outputs cite an applied framework, compared to an untagged baseline. Secondary metrics logged per canary run: behavioral-outcome proxy (clicks, decisions, completions) between tagged and untagged arms, and reader-reported clarity on a 1-to-5 scale across side-by-side comparisons.

Primary Behavioral Framework

Framework: Fogg Behavior Model (B=MAP).

Mechanism: Behavior equals Motivation multiplied by Ability multiplied by Prompt. Collapse any single factor toward zero and the behavior does not occur, regardless of the other two. Motivation makes behavior possible. Ability makes behavior reachable. Prompt makes behavior immediate. The equation is not a metaphor. It is geometric.

Source: B.J. Fogg, Stanford Persuasive Tech Lab, foundational paper 2009; refined in Tiny Habits (Fogg, 2020).

Application: every external-facing Maxim output is composed against B=MAP before ship. Motivation gets named and amplified, usually via one of Cialdini's seven principles or Prospect Theory framing. Ability gets protected through EAST's friction-removal pass; if the next action demands cognitive load the reader cannot afford, the copy is rewritten before the prompt fires. The prompt itself is timed, placed, and sized so it arrives when Motivation and Ability are already sufficient, never earlier.

Critique: Fogg is elegant for instigation and weaker for habit maintenance. Lally et al. (2010) showed the roughly 66-day habituation curve that Fogg does not address. Maxim uses Fogg for trigger moments and composes the Transtheoretical Model on top for stage-matched follow-through.

Behavioral → Persuasion Translation

The 15 frameworks below map onto one or more of Fogg's MAP variables. None of them replaces Fogg. Each modifies one lever in a specific direction.

Framework	Mechanism	Maxim positive	Maxim anti-pattern	Boundary
Fogg Behavior Model (Fogg, 2009)	B = M × A × P	/mxm-cmo scans MAP before ship	CTA fired with zero Ability	instigation-only (Lally, 2010)
COM-B (Susan Michie et al., 2011)	Capability × Opportunity × Motivation	Compliance Shield culture rollout	training absent opportunity	motivation layer overlaps Fogg
EAST (UK Behavioural Insights Team, 2014)	Easy, Attractive, Social, Timely	skill friction-removal pass	symmetric added steps	factor order contested (Benartzi, 2017)
Cialdini's 7 (Robert Cialdini, 1984; Unity 2016)	seven influence principles	Authority on moat copy	ungrounded scarcity	Cialdini's own warning against manipulation
Prospect Theory (Kahneman & Tversky, 1979)	loss aversion, reference point	"avoid $200K fine" frame	loss frame on gain-oriented offers	framing sensitivity varies culturally
SCARF (David Rock, 2008)	Status, Certainty, Autonomy, Relatedness, Fairness	enterprise onboarding protects Certainty	scarcity destroys Certainty	model not peer-reviewed at aggregate
Self-Determination Theory (Deci & Ryan, 1985)	Autonomy, Competence, Relatedness	Founder OS respects operator autonomy	gamification without competence signal	workplace application contested
Hook Model (Nir Eyal, 2014)	trigger, action, reward, investment cycle	Proactive Watch investment loop	hook without user-consented reward	Eyal's 2019 dark-pattern critique
Nudge Theory (Thaler & Sunstein, 2008)	default plus salience plus framing	safe defaults in project-manifest.json	hidden opt-outs	paternalism critique (Glaeser, 2006)
Transtheoretical Model (Prochaska & DiClemente, 1983)	six stages of change	operator-stage detection in MemPalace	stage assumed, not detected	crisp-boundary critique (Sutton, 2001)
Dual Process Theory (Kahneman, 2011)	System 1 vs System 2	pre-commit friction triggers System 2	slow copy where fast decision is needed	dichotomy oversimplified (Evans, 2014)
Jobs-to-be-Done (Christensen & Moesta, 2016)	functional, emotional, social job	persona-tier naming (Founder OS)	feature list without job named	quantitative validation disputed
MINDSPACE (UK Cabinet Office, 2010)	nine behavioral levers checklist	voice-profile messenger calibration	applying all nine simultaneously	empirical support uneven (Dolan, 2012)
Anchoring (Tversky & Kahneman, 1974)	reference-point bias	$79.99 anchors bundle discount	anchor unrelated to the decision	weak under expertise (Mussweiler, 2002)
Endowment Effect (Richard Thaler, 1980)	owned outweighs unowned	free-tier ownership feel	forcing forfeiture of "owned" features	fades with experience (List, 2003)

Where others see a prompt library, I see an enforcement hook. The frameworks above are catalogued. The anti-patterns are registered. The outputs are scanned. Concrete invocations follow.

1. Apply Fogg to a landing-page hero via /mxm-cmo:

/mxm-cmo draft-hero \
  --framework fogg \
  --target "Founder OS waiting list" \
  --variant-id 1551386
# Output: hero copy where Motivation, Ability, and Prompt are each
# addressed in three sentences before the CTA fires.

2. Declare the moat in an Maxim skill's frontmatter (ADR-007):

---
name: "Example Maxim skill"
business_outcome: "Reduces CTA abandonment; tracked in MOAT_TRACKER § MOAT-06"
primary_framework: "Fogg Behavior Model (B=MAP)"
---

3. Enforce the moat via the behavioral_audit MCP call:

{
  "tool": "mcp__mxm-behavioral__behavioral_audit",
  "args": {
    "text": "Your landing copy here.",
    "required_frameworks": ["Fogg", "Cialdini", "Prospect Theory"]
  }
}
// Returns: { "pass": false, "missing": ["Ability lever"],
//            "suggestion": "..." }

Anti-Patterns

Five failures catalogued from real Maxim pack development, documented so the mechanism is not silently eroded.

1. Framework theater. Naming a framework without applying its mechanism. Violates mechanism visibility per ADR-007 § 8.2. The reader detects the surface citation and trust collapses across the entire output.

2. Over-stacking. Three or more persuasion layers trigger psychological reactance (Brehm, 1966). Cialdini Authority plus Fogg Prompt plus Nudge Salience applied simultaneously read as manipulation. Two frameworks applied deliberately beat five applied at once.

3. Context mismatch. B2C scarcity urgency deployed in enterprise procurement violates SCARF's Certainty signal. Enterprise buyers need stability markers, not countdown timers. The framework itself is correct; the deployment context was wrong.

4. Ungrounded scarcity. "Only three left" without a real constraint. Cialdini himself warns that scarcity without truth produces trust collapse. Readers detect the hollow claim and discount every downstream signal from the same source.

5. Friction-stripping where friction protects. Removing confirmation on irreversible actions inverts the nudge. Thaler and Sunstein's core rule is "preserve freedom." Stripping the pause that lets a user undo a mistake is the nudge turned into a trap: same mechanism, opposite ethics.

Pack Integrations

Three sibling packs deepen Behavioral Intelligence in specific ways. Pack L1.1 through L1.5 are the live ecosystem; no future packs referenced.

• L1.5 Brand & Design Pro: framework-tagged visual application. Cialdini Authority becomes typography hierarchy; SCARF Certainty becomes grid stability; Dual Process System 1 recognition is encoded in color semantics.
• L1.3 Proactive Watch: the behavioral-moat-drift checker (drift class 12) fails commits where Maxim skills stop citing frameworks. Behavioral Intelligence without continuous enforcement drifts toward generic prompt-library output within weeks.
• L1.2 MemPalace Pro: per-audience framework memory across sessions. Transtheoretical Model stage is tracked per operator in the knowledge graph; the next session resumes at the correct behavioral stage instead of defaulting to pre-contemplation.

Persuasion is not style. It is architecture. Behavioral Intelligence makes that architecture visible. Visible means auditable. Auditable means defensible.

Team-tier operators get L1.6 Behavioral Intelligence inside Founder OS, Growth Stack, Professional OS, or Agency All-In at 47% off individual L1 pricing; the pack catalog has the full matrix.

---

Copyright (c) 2026 iSystematic Inc. Maxim product. BSL 1.1.