Work in Progress First public draft — March 2026. Scores are preliminary. We invite reader feedback to refine classifications. Share your thoughts ↓
Closelooknet
Closelooknet
Closelooknet Frameworks · WORK IN PROGRESS

ABR — Agent Beneficiary Ratio

Three strategic archetypes classifying how autonomous AI agents structurally impact public companies. A first attempt at the most important structural question in tech investing today. We publish this as work in progress and invite reader input to sharpen the classifications.

Universe
101
Nasdaq 100 tickers
Dimensions
5
L · A · M↑ · M↓ · C
Archetypes
3+1
Natural · Cannibalize · Terminal · Neutral
Base Date
Q1 '26
Quarterly update cadence

The agentic economy doesn't create a simple winners-and-losers list. It creates three fundamentally different structural situations for public companies. Understanding which situation a company is in matters more than any single score — because the situation determines the range of possible outcomes.

The SaaSpocalypse of February 3, 2026 — when $285 billion in SaaS market value was wiped out in 24 hours — confirmed what we've been building toward: the per-seat software model is dying. But the market still hasn't figured out who survives, who transforms, and who was dead before the crash happened.

ABR is our answer. Not a static score, but a structural classification that tells you the shape of a company's future — and whether the market has figured it out yet.

"Terminal describes the business model, not the corporate entity. A company with a terminal business model and a strong balance sheet has an escape hatch. One without doesn't."
This distinction matters. Booking Holdings generates $7B+ in free cash flow — enough to acquire agent-native capabilities. Paychex, focused on SMB payroll, has far less room to maneuver. Same archetype, different investment thesis.

The Three Archetypes

Every company falls into one of three structural situations. The archetype determines the range of outcomes — and therefore the right investment approach.

🟢
Natural Position
Right product, just execute
The existing business model directly benefits from agent proliferation. More agents = more revenue through existing infrastructure. But "Natural Position" does not mean "easy long" — customer base erosion, competitive dynamics, and what the market has already priced in all matter. The structural tailwind is real; the investment case requires separate analysis.
31
companies
Top: NVDA CRWD DDOG PANW CDNS
🟡
Cannibalize or Die
CEO execution is everything
Current SaaS 2.0 model is dying, but the company has assets — data, relationships, workflow authority — that could power a SaaS 3.0 business. If management executes the transition. Even Microsoft belongs here: Azure is pure Natural Position, but Office 365, Dynamics, and LinkedIn are all per-seat businesses facing the same structural pressure.
19
companies
Key: MSFT CRM GOOGL ADBE NFLX
🔴
Terminal
Can prolong but not reverse
The core business model is structurally incompatible with the agentic economy. Terminal describes the business model, not the corporate entity — a company with strong cash flows could acquire its way into a new model. But the current revenue engine faces irreversible structural decline regardless of management quality.
8
companies
Top: PAYX ADP CTSH BKNG
⚪
Agent-Neutral — 43 companies
Physical goods, regulated utilities, consumer staples, healthcare/biotech. The agentic economy has minimal structural impact. ABR adds no informational edge. Useful as portfolio ballast and hedges against "AI doesn't pan out" scenarios. Traditional analysis applies.
"Even companies with the perfect AI product face legacy risk — if your current customers are the ones being disrupted."
The Customer Base Erosion Problem: Datadog sells observability to SaaS companies. Those SaaS companies are shrinking. CrowdStrike protects endpoints — but agent-driven workforce reduction means fewer endpoints. Palo Alto has significant legacy firewall revenue alongside its cloud security growth. Even Cadence faces chip-design customer consolidation. This is why almost no company in the Nasdaq 100 scores L=0 in this framework. The agentic economy creates indirect exposure even for structural beneficiaries. A company can have the right product AND a shrinking customer base simultaneously — the question is whether new agent-native demand grows faster than legacy demand decays.

SaaS 1.0 → 2.0 → 3.0

The transition from seat-based software to agentic software is the mechanism that determines which archetype a company falls into.

Obsolete
SaaS 1.0
The License Era
On-premise, perpetual licenses, maintenance fees. User: human, trained on the software. Already dead for most categories.
Dying
SaaS 2.0
The Seat Era
Cloud subscriptions, per-seat pricing. Revenue scales with human headcount. The model the SaaSpocalypse of February 3, 2026 declared terminal. $285B wiped in one day.
Emerging
SaaS 3.0
The Agentic Era
Software IS the worker. Per-outcome, per-task, per-agent pricing. Revenue scales with automation volume, not headcount. The model that replaces 2.0 — for those who make the jump.
The Escape Matrix
Can a company transition from SaaS 2.0 to 3.0? It depends on two assets: proprietary data gravity and workflow authority (system-of-record status).
No Data Gravity
Has Data Gravity
Has Workflow Authority
Cannibalize — Path B (M&A)
Workflow position funds acquisition of agent-native capabilities. The old business pays for the new one.
WDAY ADSK TEAM
Cannibalize — Strongest Position
Data moat protects during transition. Workflow authority means agents MUST interact with the system. Best chance at SaaS 3.0.
CRM MSFT GOOGL META TRI
No Workflow Authority
Terminal — No Product Escape
No data moat, no workflow lock-in. Only corporate reinvention (M&A) or graceful decline.
PAYX CTSH MAR
Cannibalize — Data Transfers
Data transfers to 3.0 but workflow position is weaker. Agents could use the data through a different tool.
INTU ADBE APP

Nasdaq 100 — ABR Scorecard

All 101 tickers classified and scored. Filter by archetype. Sort by any column. Scores are preliminary — this is our first public draft.

Complete Classification DRAFT
Ticker Company Archetype L A Net M↑ M↓ C Signal

Structural Extremes

The widest structural tailwinds and headwinds in the Nasdaq 100. These are ABR positions, not investment recommendations — what the market has already priced in (M scores) requires separate, ongoing analysis.

🟢 Widest Structural Tailwind (Net ABR)
1
NVDA
NVIDIA
Compute monopoly — but market fully recognizes it
0:10
Net +10
2
ASML
ASML Holding
EUV monopoly — also largely priced
0:9
Net +9
3
PLTR
Palantir
AIP = agent orchestration platform
0:9
Net +9
4
AVGO
Broadcom
Custom silicon + VMware agent infrastructure
0:8
Net +8
5
MU
Micron
HBM monopoly — memory for agent compute
0:8
Net +8
🔴 Widest Structural Headwind (Net ABR)
1
PAYX
Paychex
Revenue scales with client headcount — agents reduce headcount
6:1
Net −5
2
CTSH
Cognizant
IT services body shop — agents replace the product
7:2
Net −5
3
ADP
ADP
Same per-employee math as PAYX — larger, slower
6:2
Net −4
4
BKNG
Booking Holdings
OTA intermediary — agents book direct
6:2
Net −4
5
WDAY
Workday
HCM per-seat death — Cannibalize with low execution odds
6:3
Net −3

Methodology

How we score each dimension. The criteria matrix is designed to be reproducible and semi-automatable.

L — Legacy Exposure (0–10)

▼

Five sub-dimensions, each scored 0–2. Total range 0–10. Measures how severely autonomous agents erode existing revenue.

SubFactor012
L1Revenue ModelAgent-immune pricingMixed exposurePer-seat/employee/headcount
L2Human Labor in DeliveryAutomated/hardwareSome servicesSignificant human-delivered revenue
L3Core Offering CommoditizationInfra agents need more ofDefensible moat, partial substitutionDirectly replicable by AI
L4Switching Cost ErosionPhysical/regulatory lock-inAI partially erodesBased on human expertise/habits
L5Channel DisintermediationProducer, not intermediaryPlatform with value-addPure intermediary/broker

A — Agent Benefit (0–10)

▼

Five sub-dimensions, each scored 0–2. Total range 0–10. Measures how directly the company captures new agentic revenue.

SubFactor012
A1Infrastructure PositioningNot required by agentsIndirect AI benefitCompute/network/memory/power agents require
A2Usage-Based RevenueFixed pricingMixed consumptionPer-API/token/query/compute-hour
A3Agent-Native ProductsAI as marketing labelMeaningful AI features, earlyAgent-native with measurable revenue
A4Data MoatNo proprietary dataValuable but replicableIrreplaceable data agents must access
A5Platform EcosystemStandalone productSome ecosystem benefitsStrong network effects growing with agents

M↑ / M↓ — Mispricing (0–10)

▼

M↑ (Upside Unpriced) — primarily for Natural Position. Weighted: Narrative Gap (40%), Consensus Blind Spot (30%), Multiple Gap (30%). Scale: 0 = fully priced → 10 = zero recognition.

M↓ (Downside Unpriced) — primarily for Terminal. Weighted: Narrative Protection (40%), Consensus Assumption (30%), Multiple Support (30%). Scale: 0 = fully punished → 10 = zero disruption discount.

For Cannibalize companies, BOTH M↑ and M↓ can be elevated simultaneously — the market has priced neither the upside potential nor the downside risk correctly.

C — Confidence & Execution Score

▼

Composite conviction: H (High), M (Medium), L (Low). Derived from data quality, catalyst visibility, disruption analogs, and management signaling.

For Cannibalize companies, C incorporates the Execution Score (5–25 scale) measuring: willingness to self-disrupt, speed of transition, competitive moat in transition, balance sheet capacity, and shareholder alignment.

Disclaimer

▼

This is analytical research, not investment advice. The ABR Framework is a structural classification tool — it does not account for all factors relevant to investment decisions including macroeconomic conditions, management quality beyond execution scoring, valuation relative to peers, or short-term catalysts. Closelooknet does not manage money or make investment recommendations. All scores are preliminary and subject to revision.