Chapter 1. Market Reality
The agentic pricing market has not settled
Five concurrent pricing patterns dominate the 22-vendor inventory: pure outcome-based (Intercom Fin, Sierra), hybrid platform-fee plus meter (Decagon, Writer, Glean, ServiceNow, Microsoft Copilot), pure per-seat with absorbed inference (Harvey, Hebbia, Notion AI, Perplexity Enterprise), per-case or per-interaction flat (EvenUp, Cresta), and pure token-rate pass-through (Cursor, Cognition Devin, GitHub Copilot after June 2026). Salesforce alone runs five SKUs simultaneously for Agentforce. That is the cleanest market-level signal that no single pattern has won.
The 14 named published voices on agentic pricing (a16z, Bessemer, Battery, OpenView, Tunguz, ICONIQ, Sequoia, Klarna, Gartner, YC, Madrona, Salesforce, Microsoft, Menlo + Foundation) converge on six consensus claims:
| # | Claim | Strength |
|---|
| C1 | Per-seat is structurally inadequate for agents at the application layer | 8 of 14, strong (application-layer only) |
| C2 | Hybrid dominates the transition | 6 of 14, strong |
| C3 | Application-layer AI gross margins compress to 40-65%; supplier-layer at 70-85% | 5 of 14, strong with numeric agreement |
| C4 | Outcome-based pricing is the destination | 7 of 14, contested on execution |
| C5 | "Agent washing" is real and significant | 4 of 14, strong |
| C6 | Value metric must proxy customer-perceived value | 5 of 14, strong |
The consensus describes the destination. The published material is largely silent on the operational discipline of getting there without a Cursor-class trust event or a Klarna-class deployment failure en route. That is the JustKeenAI white space.
Where AI gross margin compression lives
The framework's predictions apply specifically to the application layer of the AI stack. Across the four-layer architecture in 2026:
| Layer | Representative vendors | 2026 GM range | Compression direction |
|---|
| Foundation model | Anthropic, OpenAI, Google | 40-70% current, 75-85% projected 2028 | Short-term capacity-investment driven, not structural |
| Infrastructure / orchestration | Vercel, LangChain, Pinecone | 60-75% | Stable; not the framework's target |
| Application layer | Intercom Fin, Cursor, Klarna's OpenAI app, copilot products at portco scale | 40-65% | Where the compression hits. Framework's prescriptions apply here. |
| Vertical labor anchor | Harvey, EvenUp, Hebbia | 50-75% (labor cost arbitrage dominates) | Intact; inference variance is structurally absorbed |
Two implications matter. First, the framework's compression narrative is specifically about the application layer; it does not describe Anthropic-class supplier economics or Harvey-class vertical labor anchors. Second, a portco that is application-layer in business model but vertical-labor-anchor in customer base (a legal AI product priced anchored to associate cost) inherits the vertical-anchor protection automatically. Diagnosing the layer is Phase 1 work.
The hyperscaler trap, with a scope clarification
GitHub Copilot is the most credible cost-coverage signal in the inventory. Owned by Microsoft, backstopped by Azure. Two repricings in twelve months. If Microsoft's vertically-integrated economics cannot maintain flat per-seat for an agentic product at the application layer, most mid-market vendors will not.
Two named scaling exceptions show where the thesis does not generalize. Perplexity Enterprise Pro runs flat per-seat ($40-$325/user/month) for an agentic research product and has not repriced through May 2026; its agentic surface is structurally bounded. Notion AI's core seat ships at $20/user/month with unlimited core features for 2+ years; Custom Agents moved to a credit model, the core seat did not. The framework's flat-per-seat warning is calibrated to unbounded-amplifier products, not all per-seat products.
Five vendor archetypes in market today
| Archetype | Wave 1 anchor | Best-in-class pattern |
|---|
| Pure outcome, absorbed cost | Intercom Fin ($0.99/resolution) | Works at Intercom scale; structural margin compression without it |
| Hybrid platform fee + outcome | Sierra, Decagon, Cresta | Six-figure platform fee absorbs cost-coverage |
| Pure per-seat, vertical anchor | Harvey, Hebbia, EvenUp | Labor-cost anchor absorbs inference variance |
| Hybrid seat + meter (knowledge) | Writer (best-architected) | Lite/Pro split + Palmyra API meter |
| Pure per-seat, bounded agent surface | Perplexity Enterprise, Notion AI core | Agent scope structurally bounded |
| Pure pass-through (post-pivot) | Cursor, GitHub Copilot 2026 | Cost-coverage pivot; trust friction even when business absorbs the pivot |
The pattern that survives every margin-collapse scenario in the DPO 4 model is the Writer-class hybrid: free Lite tier absorbs low-usage employees, Pro seat captures active users, Palmyra API meter charges programmatic and power consumption. It is the template the framework points portcos toward when other archetypes do not apply.
Chapter 2. The Value-Metric Decision Framework
Pick the value metric first, then design the contract
The most common pricing mistake at PE-backed B2B SaaS portcos is to copy the headline metric of a successful vendor without copying the structural cost-coverage architecture underneath it. Fin's $0.99 per resolution works because Intercom has Atlassian and Shopify-class scale spreading fixed costs. A $50M ARR portco copying $0.99 with a similar agent stack faces the Support Agent Volume Trap (Chapter 3) the moment query mix shifts.
The decision tree below sorts agentic SaaS products into 6 archetypes and 8 terminal value-metric nodes. Each terminal carries dual vocabulary: the internal mechanism for PE operating partner diagnostics and CFO modeling, and the customer-facing sell in Predictability Architecture vocabulary for the AE pitch and the contract preamble.
The decision tree
ROOT: "How does the customer measure success?"
A. Volume of interactions → Branch A
A.1: Per-interaction cost variance > 3x across cohorts?
YES → A.2 Hybrid seat + meter (Writer / Outreach Amplify)
NO → A.3 Per-interaction + platform floor (Decagon / Cresta / Fin)
B. Specific outcomes → Branch B
B.1: Outcome attributable to the agent without dispute?
YES → B.2 Per-outcome + quality SLA + tier floor
NO → B.3 Hybrid seat + credit + attribution rules
C. Augmented productivity → Branch C
C.1: User is a licensed/credentialed professional?
YES → C.2 Per-seat unlimited or per-case flat
NO → C.3 Hybrid seat + meter with Lite/Pro split
D. Net new capability → Branch D
D.1: Process unit-cost bounded within a known range?
YES → same as B.2
NO → D.2 Per-run + spend cap + quality SLA + tier floor
E. Code or developer output → Branch E
E.1: Product spans autocomplete and autonomous agentic features?
YES → E.2 Tiered hybrid + mandatory spend cap
NO → E.3 Per-ACU or per-task + mandatory spend cap
The 6 archetypes mapped to these branches: Support agent, Sales/marketing agent, Vertical professional-services agent, Knowledge/horizontal copilot, Autonomous workflow agent, Developer tool / code agent. Every terminal node includes a meter, a cap, a floor, or a structural pricing-power anchor (vertical labor cost). That is not an accident.
Dual-vocabulary translation table
| Value Metric | Internal Mechanism | Customer-Facing Sell |
|---|
| Hybrid seat + meter (A.2, C.3) | Two-sided cost-coverage equilibrium with decoupled layers | "Predictable seat; metered AI stays inside contracted guardrails" |
| Per-interaction + floor (A.3) | Visible price + invisible cost-coverage fee + quality SLA | "Your support cost moves only with ticket volume; the quality bar is contracted" |
| Per-outcome + quality SLA + tier floor (B.2, D.1) | Quality-gated outcome with vendor floor protection | "You pay only for completed work that meets the quality bar" |
| Hybrid seat + credit + attribution (B.3) | Decoupled architecture with pre-resolved attribution | "Rep seats stay procurement-friendly; AI credits stay metered; both forecastable" |
| Per-seat unlimited / per-case flat (C.2) | Labor-cost anchor absorbs inference variance | "Flat per-seat or per-matter, anchored to associate cost, not model cost" |
| Per-run + spend cap + quality SLA (D.2) | Quality-gated outcome with runaway-loop protection | "Every process run has a hard cost ceiling; quality is contracted and auditable" |
| Tiered hybrid + mandatory cap (E.2) | Tiered absorption with cap (mandatory for enterprise) | "Predictable seat for daily work; clear ceiling on agentic spend" |
| Per-ACU + per-task cap (E.3) | Effort-based meter with bounded exposure | "Pay per completed task; contracted ceiling per task and per month" |
Chapter 3. The Cost-Coverage Risk Model
The Cursor June 2025 case, re-read
On June 16, 2025, Cursor issued a routine-looking pricing update. The Pro plan, previously $20/month for 500 fast model requests, became $20/month for $20 of model credits at API rates. The language was anodyne. The mechanism was a transfer of the entire AI cost-coverage risk from Cursor's balance sheet to the customer's credit card.
CEO Michael Truell apologized publicly on July 4, 2025: "We have to be more careful about how we communicate pricing changes, and we have to do it in a way that is transparent. The previous model was financially unsustainable." Cursor refunded affected users. Trustpilot settled at 1.7/5 across 203 reviews. ARR grew from $100M in January 2025 to $2B by February 2026; the company entered funding talks at a $50B valuation. The pivot triggered persistent trust friction that did not stop one of the fastest B2B scaling events in history.
Cost taxonomy and the eight hidden cost amplifiers
Six top-level cost categories: inference (HIGH volatility, EXTREME on reasoning), agent architecture (MEDIUM-HIGH), compute infrastructure (MEDIUM), model variability (LOW frequency, HIGH magnitude on migrations), human fallback (matters when outcome pricing is committed), infrastructure overhead (LOW). The categories tagged HIGH or EXTREME are where contract terms must be most aggressive.
| # | Amplifier | Multiplier | Detection signal |
|---|
| 1 | Planner re-run / agent loop overhead | 2x-5x (8x at P95) | Planner-to-executor token ratio > 1:1 |
| 2 | Tool-call fan-out | 3x-10x | Tool calls per user query > 15 |
| 3 | Multi-model chain | 2x-4x | Distinct model IDs per query > 1 |
| 4 | Evaluator / judge / verifier | 1.5x-2.5x | Evaluator stage in agent graph |
| 5 | Reasoning token explosion | 5x-20x | Output count > 5x visible response |
| 6 | Observability at scale | 1.05x-1.15x on stack | Datadog AI bill > 5% of inference |
| 7 | RAG recursive depth | 2x-8x | Retrieval events per query > 25 |
| 8 | Cache miss patterns | 1.4x-2.0x | Cache hit rate < 40% |
Amplifiers compound, they do not add. A fully-stacked agent (loop × multi-model × evaluator × fan-out × reasoning) can produce 100x+ cost on the same nominal task. Rule of thumb: assume 5-8x amplifier on baseline inference at P75 traffic for unbounded agentic surfaces. For bounded agentic surfaces (Perplexity-class, Notion-core-class), 2-3x is the realistic P75 assumption.
Sensitivity scenarios: archetype × pricing
Five scenarios stress-test six archetype-and-pricing combinations. S1 base case; S2 inference cost +50%; S3 volume +200%; S4 agent loop overhead 3x; S5 combined adverse "AI Winter Pivot." Effective GM (%) per archetype × scenario:
| Archetype × Pricing | S1 Base | S2 (+50% infer) | S3 (+200% vol) | S4 (3x loop) | S5 Combined |
|---|
| Support, Pure outcome ($0.99) |
21% | -19% | -8% | -44% | -118% |
| Support, Hybrid (platform + outcome) |
52% | 38% | 41% | 26% | 9% |
| Dev tool, Pure seat ($20) |
18% | -23% | -32% | -110% | -280% |
| Dev tool, Decoupled meter (Copilot AI Credits) |
45% | 34% | 38% | 22% | 11% |
| Vertical, Per-seat ($1,500) |
78% | 67% | 55% | 38% | 12% |
| Vertical, Per-case ($500 EvenUp) |
80% | 71% | 68% | 52% | 32% |
The hybrid and Writer-class architectures stay positive across all five scenarios. The pure-outcome and pure-seat-application-layer architectures go negative on a single bad quarter. That is the cost-coverage risk in one table.
Five named margin-collapse scenarios
- The Support Agent Volume Trap (S1+S3). Pure-outcome pricing loses 29 percentage points of margin during success-driven scaling as query mix shifts toward complex queries and HITL escalation rises. Remedy: outcome rate paired with a platform-fee floor OR a quality-weighted definition of "resolution."
- The Developer Tool Loop Creep (S1+S4). 18 months of agentic feature additions (planner step, evaluator step, multi-file context, test-runner integration) without repricing compress margin from 18% to -110%. Cursor June 2025 is exactly this. Remedy: every new agentic feature requires a cost-class declaration at design review; surface the tail to customers before margin goes negative; grandfather existing customers when the pivot lands.
- The Vertical Reasoning Avalanche (S2+S4). A legal AI product enables extended-thinking by default; a 50% upstream price increase on reasoning compounds with a 3x loop feature. Margin moves from 78% to 12%. Remedy: reasoning tokens as a separate meter; per-deliverable hard cap on reasoning tokens; pre-staged model-version pass-through clause.
- The Combined Adverse "AI Winter Pivot" (S5). Provider repricing coincides with success-driven volume and accumulated complexity. Pure-seat application-layer developer tool reaches -280% margin. Remedy: pre-positioned model-version hedges (2+ providers, BYOK at enterprise, contractual pass-through trigger above 25% upstream cost movement); pre-staged customer caps; quarterly margin review at product line level.
- The Quality Erosion Mirror (Klarna Deployment-Scope Failure). Vendor maintains pricing by deploying AI to use cases beyond its quality envelope; outputs degrade; customer rehires humans. Remedy: quality-weighted outcome metrics contracted (resolution-quality floor, NPS floor, escalation-rate ceiling); customer-side telemetry; quality-engineering investment alongside the pricing clause; renegotiation triggers tied to quality breaches.
What this looks like for a $100M ARR portco (worked example)
The framework's $1.8M-$3.7M program investment and the multiple-protection claim apply to a specific representative profile: $100M ARR, 22% YoY growth, 55% GM baseline, 8x revenue multiple at exit, 24-month transition window, application-layer agentic surface.
| Cost line | One-time | Annual ongoing |
|---|
| Model commitment capex (multi-provider) | n/a | $500K-$2M |
| Observability instrumentation (Datadog AI + Langfuse / Braintrust) | n/a | $200K-$500K |
| Eval infrastructure (model migration regression testing) | n/a | $100K-$300K |
| Contract retrofit (legal + ops on existing book) | $250K-$1M | $50K-$100K (residual) |
| Engineering build-out (P9 + P15 in product) | $400K-$1M | $150K-$300K |
| Advisory engagement | n/a | $750K-$1.25M |
| Total | $0.65M-$2M | $1.75M-$4.45M (peak year) |
Two-year program total: $1.8M-$3.7M. Protection band on the representative profile: 0.5-1.5x of the 8x baseline multiple, or $50M-$150M of enterprise value protected. A portco running its own version of the calculation with its own GM baseline, multiple band, and ACV mix will arrive at a different number; the work is to do the calculation.
The three levers in one slide
| Lever | Investment | Protects against | Diligence signal preserved |
|---|
| Caps | Engineering + contract retrofit ($500K-$1M) | Runaway P95 cost (loop creep) | Cost forecast variance held under 15% |
| Floors | Sales motion change + commercial training ($300K-$700K) | Margin compression during volume scaling | GM trajectory holds above 50% under stress |
| Model-version hedges | Provider commitments + multi-routing ($1M-$2M) | Upstream cost shocks | Pre-staged pass-through clauses in customer book |
Chapter 4. Customer Risk Mitigation Patterns
Three levers, top three patterns
The Predictability Architecture rests on three levers that together prevent every named margin-collapse scenario in the unbounded-agent application-layer regime: caps (per-task, per-customer, per-tenant), floors (annual minimums, platform fees, Lite/Pro split), and model-version hedges (multi-provider routing, BYOK at enterprise, contractual pass-through trigger above 25%).
Every name in the Wave 1 inventory that has survived an agentic-pricing crisis without trust damage runs all three levers. Every name that has experienced a Cursor-class trust event was missing at least one.
The top three patterns
P4, Predictable Seat + Metered Power-User Usage (Writer Architecture). This is the broadly applicable pattern for application-layer agentic products. The Writer-class architecture is the only one in the Chapter 3 sensitivity matrix that stays positive across all five margin-collapse scenarios. Year 1 priority: retrofit every existing seat-priced contract that lacks a metered escape valve on unbounded agentic surfaces.
P9 + P15, Universal Riders for Enterprise Procurement (Budget Hard Stop + Model Version Stability). These two riders are required on every multi-year metered contract in enterprise procurement-driven sales motions at $250K+ ACV. P9 is the enterprise-procurement Cursor-prevention clause; P15 is the universal "AI Winter Pivot" defense. Together they implement Levers 1 and 3 at the per-customer level.
P12, Outcome Guarantee with Quality Assurance (Klarna Remedy), scoped to $1M+ ACV. Highest aggregate moat score in the library. Builds Counter-Positioning, Switching Costs, Brand, and Process Power simultaneously. P12 is scoped to $1M+ ACV deals; below $1M ACV, use the P12-Light variant (outcome metric + quality SLO + graduated credit remedies without full audit infrastructure).
Five specimen contract clauses (excerpts)
Clause 1, P4 Hybrid Seat + Metered API
"Customer's monthly fees consist of two components: (a) a Subscription Fee for named user seats at the Lite, Pro, or Enterprise tier ('Seat Fees'), invoiced quarterly in advance; and (b) Metered API Fees for transactional API usage above the included allotment ('API Overages'), invoiced monthly in arrears. The Per-Call Rate may not increase by more than five percent (5%) in any twelve (12)-month period without Customer's prior written consent. Lite Seats are provided at no cost up to the lesser of (i) [N] Lite Seats per paid Pro Seat or (ii) the number of Customer's active employees."
Clause 2, P3 Per-Resolution + Platform Floor + Quality SLA
"A 'Qualified Resolution' means a customer support interaction that (i) is closed by the Service without escalation to a human agent within seven (7) business days; (ii) is associated with a post-interaction CSAT of 3.5 or higher on a 5-point scale; and (iii) does not result in escalation to Tier 2 or above. Service Quality Bar: Qualified Resolution rate not less than 70%; CSAT average not less than 4.0; escalation rate not more than 25%. Quality Remedy: failure in any single calendar month reduces Resolution Fees by fifteen percent (15%); failure in two consecutive months permits Customer termination on thirty (30) days' notice."
Clause 3, P9 Budget Hard Stop
"When Customer's accumulated Metered Fees in a calendar month reach eighty percent (80%) of the Monthly Spend Cap, the Service shall send an automated notification to each Approved Signatory. When accumulated fees reach one hundred percent (100%), the Service shall suspend all metered API execution until the earlier of (i) the start of the next calendar month, or (ii) an Overage Authorization from an Approved Signatory. Lite-tier and Pro-tier seat-included functionality shall continue without interruption during any suspension."
Clause 4, P15 Model Version Stability
"Vendor reserves the right to migrate the Underlying Model from time to time, provided that Service quality shall not degrade and Customer pricing shall remain unchanged as a direct result. Upstream Cost Trigger: if Vendor's documented per-unit cost of Underlying Model inference increases by more than twenty-five percent (25%) over any rolling twelve (12)-month period, Vendor may pass through to Customer pricing not more than fifty percent (50%) of the documented increase, on ninety (90) days' notice. Customer shall have a one-time right to terminate for convenience if any pass-through exceeds ten percent (10%) of the aggregate annual fees."
Clause 5, P12 Outcome + Quality SLO + Gainshare (v2, graduated remedy)
"Vendor shall invoice Customer the Outcome Fee only for Verified Outcomes meeting Acceptance Criteria. Quality Floor: Verified Outcome rate not less than ninety percent (90%); audit pass rate not less than ninety-five percent (95%). Graduated Quality Remedy: for any calendar month in which Verified Outcome rate falls below Quality Floor, Customer shall receive a credit against that month's Outcome Fees as follows: (i) 5% credit for misses up to five (5) percentage points; (ii) 15% credit for misses greater than five (5) and up to ten (10) percentage points; (iii) 30% credit for misses greater than ten (10) and up to twenty (20) percentage points. Termination Right: Customer may terminate for cause if Verified Outcome rate falls more than twenty (20) percentage points below Quality Floor in two consecutive months. Performance Bonus: for each calendar quarter in which documented Customer savings exceed the Savings Target and the Quality Floor is met, Vendor receives a Performance Bonus equal to twenty percent (20%) of documented savings in excess of the Savings Target."
Moat impact summary (Helmer 7 Powers)
| Rank | Pattern | Primary Moats |
|---|
| 1 | P12 (Outcome + Quality SLO + Gainshare, $1M+ ACV) | Counter-Positioning, Switching Costs, Brand, Process Power (all strong) |
| 2 | P4 (Hybrid Seat + Meter Lite/Pro) | Switching Costs, Process Power (both strong) |
| 3 | P15 (Model Version Stability) | Switching Costs, Brand (both strong) |
| 4 | P1 (Vertical Per-Seat Unlimited) | Counter-Positioning, Cornered Resource |
| 5 | P3 (Per-Resolution + Platform Floor + Quality SLA) | Process Power |
Chapter 5. Transition Playbook
Five phases over 24-30 months
A pricing redesign for an agentic SaaS portco is a 24-30 month operating program, not a quarter-long initiative. Reversibility drops at each phase boundary. The cliff is between Phase 2 (80% reversible) and Phase 3 (50% reversible). Cross it deliberately.
| Phase | Duration | Reversibility | Key Activity | Exit Gate |
|---|
| 1: Insight | 4-6 weeks | 100% | Cost taxonomy audit, contract risk audit, archetype selection | Target terminal node ratified |
| 2: Pilot | 8-12 weeks | 80% | Deploy P4/P3/P12 to 3-5 design partners with universal P9/P15 | >90% pilot retention; NRR meets target |
| 3: New Sales Default | 12-16 weeks | 50% | All new logos on new pricing; renewal cohort offered with grandfathering | 3 quarters new-logo retention >85% |
| 4: Cohort Migration | 12-24 months | <30% | Segmented migration; P15 retrofits on multi-year deals | >70% ARR on new pricing |
| 5: Full Cutover | 18-30 months total | ~0% | Legacy deprecation; final grandfather sunset | 97-99% ARR on new pricing |
The Phase 3 sales-comp precondition gate (new in v2)
The Phase 3 sales-comp redesign is the single most reversibility-cliff-adjacent action in the playbook. Top performers will run the math on a new comp structure and exit for competitors with simpler comp. The reversibility cliff is partially driven by human capital risk, not just contract risk.
Before Phase 3 launches, the portco must complete five preconditions:
- AE comp impact modeling. Run the new plan against actual W-2 history of the top 20% of reps. If any top-quintile rep loses more than 15% of cash comp in the model, the plan needs adjustment before launch.
- Retention conversations. 1:1s with the top 3 quota carriers (CRO-led) explicitly framing the comp change, the rationale, and the AE's place in the post-transition org.
- Hire-back / replacement runway. Recruiting pipeline filled to 1.5x of expected top-decile attrition (15-25% range). Signing bonuses pre-approved at the CEO level.
- Competitor comp benchmarking. External benchmarking against the 2 most likely competitive recruiters. If the new comp is more than 10% below competitor comp at the top-quintile AE level, the plan creates an active flight incentive.
- Executive sponsor. A single CRO or CEO accountable for AE retention through Phase 3 by name, with explicit P&L for top-rep churn during the transition.
If any of the five preconditions is not completed, Phase 3 does not launch. The phase is paused until the gate clears.
Stakeholder communications: six templates, one vocabulary rule
| # | Audience | Vocabulary | Use |
|---|
| 1 | Board | Operator-grade | Program approval; phase advancement |
| 2 | CFO | Operator-grade, quantitative | Unit economics defense; cohort modeling |
| 3 | CS playbook | Mixed | Renewal motion, objection handling |
| 4 | Sales script | Mixed | Net-new deal positioning, deal sequence |
| 5 | Customer announcement | Predictability Architecture only | Customer pricing announcement |
| 6 | Internal FAQ | Bridges both | Cross-functional alignment |
The vocabulary rule: "shared risk," "transition," "cost-coverage," "vendor margin" are operator-grade vocabulary used internally only. The customer hears predictable cost, hard ceilings, value floors, quality assurance. Mixing destroys the framing. The Cursor execution friction was, at root, partly a vocabulary failure.
Grandfathering: eight rules, four cohorts
- Honor every existing contract through its current term
- Grandfather window of 12-24 months from new pricing launch (up to 36 months for flagship)
- Long-term contracts (3-5 year deals) honored in full; voluntary migration with credit incentive
- Migration sequencing: smaller and less strategic first; flagship accounts in Wave 4 of Phase 4
- Partial migration allowed (some BUs on new, some on legacy)
- Voluntary migration credits decay over the window (10% → 7.5% → 5% → 0%)
- P9 (Budget Hard Stop) added universally at next renewal for enterprise multi-year deals
- P15 (Model Version Stability) retrofit offered free to every multi-year customer
Success metrics by phase
- Phase 1: Per-account inference cost per ARR dollar (P50, P95, P99); contract end dates by cohort; amplifier scorecard; GM stress-test against S1-S5.
- Phase 2: Pilot NRR; pilot retention; cap utilization distribution; quality SLA compliance rate; competitive-pricing-landscape monitoring.
- Phase 3: New-logo % of new ARR on new pricing (target 100%); attach rate of P9/P15 on multi-year enterprise deals (target 100%); top-quintile AE retention (target >85%).
- Phase 4: % ARR migrated; migrated-cohort 12-month retention; GM by quarter vs the Year 2-3 trajectory.
- Phase 5: Final % ARR on new pricing (target 97-99%); total program GM trajectory; total retention.
Chapter 6. The JustKeenAI Advisory Engagement
Operating the framework internally
The framework is fully self-applicable. A portco with a capable CFO, CPO, CTO, and GC can execute the entire Predictability Architecture program from the chapters above. The pattern library is drop-in usable. The specimen contract clauses lift directly into MSAs and Order Forms. The decision tree is answerable in 30 minutes by anyone familiar with their own product. The cost-coverage diagnostic is calculable from internal telemetry and the published amplifier table.
JustKeenAI does not gate the framework. The intellectual asset is the framework itself. The engagement is the optional accelerant for the conditions where it adds value an internal team cannot replicate.
Three engagement paths
Path 1
Self-Serve
Run the entire program from the chapters above. Best fit when the portco has a capable CFO, CPO, CTO, and GC, and the executive bandwidth to facilitate cross-functional decisions internally.
- Pattern library drops in
- Specimen clauses lift into MSAs
- Decision tree answerable in 30 minutes
- Amplifier table calculable from telemetry
$0 to JustKeenAI
18-24 months end to end
Path 2
Single-Purpose Interventions
Targeted advisory on the specific decisions that carry asymmetric risk. The portco runs the program; JustKeenAI accelerates a single high-stakes call.
- P12 quality SLO drafting: 3-4 weeks, $50K-$125K
- P15 Contract Portfolio Audit: 60-90 days, $150K-$300K
- Pre-diligence Board Narrative Review: quarterly, $25K-$75K
$25K-$300K per intervention
Days to weeks to months
Path 3
Full Engagement
Five high-leverage touch points across the 24-30 month transition program. Portco internal teams own execution; outside advisory accelerates and de-risks the program at the phase boundaries.
- Phase 1 Diagnostic ($50K-$125K)
- Phase 2 Pilot Design ($75K-$150K)
- Phase 3 Sales-Comp Redesign + precondition gate
- Phase 4 High-Risk Cohort Mitigation
- Quarterly Board Narrative
$1.5M-$2.5M over 24 months
Five touch points
Three trigger conditions that map to engagement paths
- A provider repricing event looks imminent and the multi-year book is exposed. GitHub Copilot's twin reprices in twelve months are the canary. Portcos with multi-year contracts lacking P15 are exposed to a forced pass-through that triggers customer renegotiation. Maps to the P15 Contract Portfolio Audit (Path 2).
- The portco is 18-30 months from a planned exit and the diligence question is forming. PE buyers in 2027 expect to see a documented Predictability Architecture diligence answer for agentic-AI portcos. Maps to Quarterly Board Narrative Review (Path 2) or the Full Engagement (Path 3).
- The portco's product is on the wrong terminal node of the Chapter 2 decision tree. Most common case: a knowledge copilot on pure per-seat at the application layer without Perplexity-class bounded agentic surface, a support agent on pure per-outcome without a platform fee, a vertical AI on per-document without per-case anchor. Maps to the Full Engagement (Path 3), starting with the Phase 1 Diagnostic.
Two packaged products for specific conditions
P12 First-Mover Engagement ($1M+ ACV portcos only). A focused 6-9 month engagement to design, ship, and market the first operationalized P12 contract (outcome + quality SLO + gainshare with audit) in the portco's category. Eligibility: at least one $1M+ ACV deal in active pipeline; quality-conscious buyer profile; capacity to invest in audit infrastructure ($400K-$880K Year 1 setup). Typical scope: $250K-$500K.
P15 Contract Portfolio Audit. A fast-cycle 60-90 day engagement to retrofit an existing multi-year contract book with Model Version Stability riders. Retrofit cost is low ($250K-$1M legal + ops); GM-protection band 5-15 percentage points on the representative portco profile. Typical scope: $150K-$300K.
What JustKeenAI brings that internal teams cannot easily replicate
- External adversarial review. Internal teams will not consistently surface the failure modes they are personally invested in not seeing.
- Cross-portfolio benchmarking. Public data does not show where peer portcos sit on the decision tree or what comparable deals look like.
- Vendor negotiation leverage. Multi-provider model commitments, BYOK pricing, capacity-deal terms. First-time portco negotiators leave material money on the table.
- Stakeholder facilitation across the board, CFO, CCO, CS lead, and sales VP. Each owns a different fear about the transition. Internal facilitation by one of them carries political weight that distorts every recommendation.
- Phase 3 reversibility-cliff discipline. The single highest-stakes call in the program. Internal teams under-prepare it.
Appendices
Appendix A: Full 19-pattern catalogue (P1-P19, including P12-Light). Core patterns anchored to decision tree terminals, plus operational variants for transition, attribution, and BYOK.
Appendix B: Full specimen clause library with drafter notes (P4, P3, P9, P15, P12 v2 graduated remedy, P12-Light).
Appendix C: Per-amplifier cost models and scenario derivations.
Appendix D: Source citations. Fourteen named published voices, 88 primary sources. Anchor sources include Bessemer ("The AI Pricing and Monetization Playbook," Feb 2026), a16z ("AI Is Driving A Shift Towards Outcome-Based Pricing" + Sarah Wang CIO Survey), Tomasz Tunguz, ICONIQ Capital ("2025 State of AI: The Builder's Playbook"), Sequoia Capital ("AI's $600B Question"), Gartner ("2026 Hype Cycle for Agentic AI"), Klarna case study, Menlo Ventures + Foundation Capital, TechCrunch coverage of the Cursor CEO apology, DevOps.com ("GitHub Resets Copilot Pricing as AI Compute Costs Surge"), SoftwareEquity ("The Impact of Gross Profit Margin on SaaS Company Valuations"), FinanceResolver ("The Forecast Gap That Makes Investors Reprice SaaS Valuations," May 2026), Aventis Advisors ("SaaS Valuation Multiples: 2015-2026"), and Livmo ("NRR: How It Can Double Your SaaS Exit Multiple," 2026).
Appendix E: Cursor data verification. The $100M to $2B ARR trajectory (Jan 2025 to Feb 2026), 70% Fortune 1000 adoption, $50B-valuation funding talks, and persistent customer-trust friction (Trustpilot 1.7 across 203 reviews) are documented from primary and secondary sources.
© 2026 Just Keen AI. The decision is not whether to implement the Predictability Architecture. The decision is whether to implement it on a planned cadence or after an event the portco's product moat does not absorb.