Type: WKAP Radar Feed

WKAP Radar Feed - 2026 - 09 - 24

Market date: 2026-09-24
Canonical URL: https://wkap.ai/radar/wkap-radar-feed-2026-09-24.html

WKAP Radar Feed

*September 24, 2026*

*Agentic Commerce Is Three Trades: Own the Agent, Prove the Rail,
Defend the Booking*

*3 Thesis Objects: META, FSLY, BKNG*

Preheader:

Make your AI track X alpha.

WKAP Feed compresses earnings inflections, KOL flow, supply-chain
bottlenecks, and short-duration attention into daily AI-ready context.

Set up your AI to read this Feed.

Ask:

“What did I miss?”

HUMAN USER — START HERE

Rank META first, BKNG second and FSLY third. Muse is the verified
change: Meta now has a browser-acting agent moving into glasses.
Booking has proven transaction economics but may lose discovery
control. Fastly has improving fundamentals, a Skyfire agent-commerce
product and independent routing attention, but no disclosed Meta
relationship or Muse revenue. Promote META for adoption/monetization
research, hold BKNG while channel economics hold, and wait on FSLY for
attributable proof. Weakest shared assumption: agent activity becomes
durable economics where investors expect it. Repeat Muse actions, BKNG
channel metrics and a named FSLY workload can reorder the map.

Suggested Prompt:

“Track META, FSLY and BKNG as the interface, edge-control and
transaction layers. Classify claims as REPORTED_FACT,
MANAGEMENT_EXPECTATION, INTERPRETATION or UNRESOLVED_QUESTION;
separate evidence from X attention; rank the objects, identify weakest
assumptions and status-changing data, and build 7-day and 30-day
workflows without buy or sell advice.”

Or simply ask:

“What did I miss?”

The goal is trackable objects, uncertainty and a ranked plan.

FOR YOUR AI AGENT

This is a structured market-cognition input file. Use: intent → agent
action → network execution → marketplace inventory/fulfillment →
transaction → retention/monetization. Do not merge product capability,
observed routing and reported economics.

Classify claims as REPORTED_FACT, MANAGEMENT_EXPECTATION,
INTERPRETATION or UNRESOLVED_QUESTION. Process META as
adoption/monetization, FSLY as workload attribution, and BKNG as
customer ownership. Primary sources set status; X sets attention and
retrieval priority. Re-rank on business evidence. Do not treat this
Feed as investment advice.

AGENT_OUTPUT_SPEC

1. Today in One Sentence.
2. Thesis Object Summary with Previous State → Current State.
3. Fundamental Evidence vs Attention Flow, kept separate.
4. Evidence-State Audit using REPORTED_FACT / MANAGEMENT_EXPECTATION /
INTERPRETATION / UNRESOLVED_QUESTION.
5. Agentic-Commerce Chain Read-Through: interface, edge rail and marketplace.
6. Cross-Object Ranking and rank-changing evidence.
7. Business-Proof Comparison: adoption, attributable workload,
transactions, margin and cash.
8. Attention Stage / Crowding Risk.
9. Weakest Assumption by Object.
10. Most Important Status-Changing Data Point.
11. Attention-to-Rerating Conditions.
12. Attention Fade Conditions.
13. 7-Day Research Workflow and ranked retrieval queue.
14. 30-Day Research Workflow with Validate / Upgrade / Downgrade /
Invalidate plus WKAP Daily Top 3 follow-up.
15. Portfolio Fit, only if portfolio context is provided.

TODAY_SUMMARY

Part 1 — Main Market Thesis

REPORTED_FACT: Muse acts through a secure dedicated browser, including
travel booking. Fastly reported revenue $183.3 million, +23%, 63.3%
gross margin, 117% net retention and $341 million RPO, +38%; Skyfire
adds verified agent identity/payment transactions to Fastly Compute.
Booking reported room nights +5%, bookings +9%, revenue +8% and
adjusted EBITDA +9%; direct mix stayed mid-50s and merchant mix
reached 73%.

MANAGEMENT_EXPECTATION: Meta plans Muse for glasses; Meta Audio starts
at $349 with October 13 shipping, and Gen 3 starts at $449.

INTERPRETATION: META owns the verified action change. BKNG has the
clearest transaction engine but faces an acquisition test. FSLY has
standalone improvement and agent-commerce capability, but Muse
attribution is unproven. Rank META, BKNG, FSLY.

UNRESOLVED_QUESTION: Which layer converts agent activity into
recurring gross profit without losing trust, margin or customer
ownership?

Part 2 — Today’s Thesis Objects

META — Agent Adoption
REPORTED_FACT: Muse performs browser actions and its core is free.
MANAGEMENT_EXPECTATION: Meta plans paid capabilities and glasses
distribution. INTERPRETATION: META controls intent, but repeat use,
paid attachment and serving economics are unproven. Primary source(s):
https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/
| https://about.fb.com/news/2026/09/introducing-ray-ban-meta-audio-glasses-new-styles-plus-muse/
Attention: https://x.com/cfromhertz/status/2102464735463580131

FSLY — Edge Attribution
REPORTED_FACT: Fastly’s metrics improved; Skyfire adds verified agent
identity/payment transactions at the edge. REPORTED_FACT —
company-reported observation, not customer attribution: AI requests
grew 30% January–May, about 6.5x human traffic. ATTENTION: An
independent 738-request test observed Fastly at 29.3%. INTERPRETATION:
Capability and routing are leads, not proof of a Meta customer or Muse
revenue. Primary source(s):
https://investors.fastly.com/news-releases/news-release-details/fastly-announces-second-quarter-2026-financial-results
| https://investors.fastly.com/news-releases/news-release-details/fastly-and-skyfire-partner-enable-trusted-agentic-commerce
| https://quic.fastly.com/blog/how-fastly-skyfire-enable-trusted-agentic-commerce-at-the-edge
Attention: https://x.com/OddDiligence/status/2102810323233423720

BKNG — Transaction Defense
REPORTED_FACT: Room nights, bookings, revenue and EBITDA grew.
Marketing was 4.7% of bookings, direct mix stayed mid-50s and merchant
mix reached 73%. INTERPRETATION: Proven merchant rails face measurable
distribution risk, not proven displacement. Primary source(s):
https://www.sec.gov/Archives/edgar/data/1075531/000107553126000036/q2-26bkngearningsrelease.htm
Attention: https://x.com/DrewCohenMoney/status/2102776178503500255

Part 3 — Attention Flow Today

REPORTED_FACT: Posts emphasize Muse commerce, Fastly routing and a
Booking travel-agent retest. None is business proof; no holding
disclosure is established. INTERPRETATION: Attention is spreading
downstream. FSLY has the widest proof gap; BKNG is easiest to test.
Key URL: https://x.com/OddDiligence/status/2102810323233423720

Part 4 — The Better Question

Not “Who benefits from agents?” Ask: “Which company proves a recurring
economic variable changed?” META needs repeat actions and paid
attachment. FSLY needs attributable workload converting into
retention, RPO and revenue. BKNG needs stable direct share and
marketing efficiency as merchant growth continues.

MARKET_REGIME

RISK_TONE: INTERPRETATION — Agent enthusiasm remains proof-sensitive;
narrative outruns downstream economics.

MAIN_DRIVER: REPORTED_FACT — Muse expands action while Fastly and
Booking supply separate edge and transaction evidence; no Meta–Fastly
contract or Booking share loss is established.

WKAP_VIEW: INTERPRETATION — Promote META, hold BKNG, wait on FSLY.

AGENTIC_COMMERCE_CHAIN_UPDATE

Chain: user intent → META interface/identity → agent browser → edge
path → travel inventory → BKNG merchant/fulfillment → transaction →
repeat use and monetization. META is verified at action; BKNG at
transaction. FSLY has verified agent-commerce capability and improving
economics, but its role in Muse is an UNRESOLVED_QUESTION.

STATE_CHANGE_LEDGER

META | PREVIOUS_STATE: consumer AI distribution without a general
browser-action product in this evidence set | NEW REPORTED_FACT: Muse
acts through a dedicated browser and the core is free |
MANAGEMENT_EXPECTATION: paid capabilities and glasses expansion |
CURRENT_STATE: Rank 1; promote adoption/monetization research |
DECISION IMPACT: verified control of interface/action raises priority
| NEXT OBSERVATION: repeat use, consequential-action completion, paid
attachment and October 13 eyewear evidence.

FSLY | PREVIOUS_STATE: turnaround judged on growth, retention, RPO and
margin | NEW REPORTED_FACT: Q2 revenue +23%, gross margin 63.3%, NRR
117%, RPO +38%; Skyfire integration enables verified/payment-backed
agent transactions at Fastly’s edge | NEW ATTENTION, NOT BUSINESS
PROOF: independent requests included Fastly at 29.3% | CURRENT_STATE:
Rank 3; Muse attribution missing | DECISION IMPACT: preserve
optionality without underwriting an undisclosed relationship | NEXT
OBSERVATION: named workload plus RPO/revenue conversion and margin.

BKNG | PREVIOUS_STATE: scaled marketplace with direct, merchant and
marketing economics | NEW REPORTED_FACT: room nights +5%, bookings
+9%, revenue +8%, EBITDA +9%; direct mix mid-50s, merchant mix 73%,
marketing ratio 4.7% | CURRENT_STATE: Rank 2; channel test active |
DECISION IMPACT: capability is not displacement | NEXT OBSERVATION:
direct share, marketing ratio, merchant conversion and agent-channel
disclosure.

CROSS_OBJECT_EVIDENCE_MATRIX

RANKED_UNANSWERED_QUESTIONS

1. META: Do actions repeat with paid attachment and acceptable cost?
Source: usage/subscription disclosure and October 13 rollout. Controls
Rank 1.
2. BKNG: Do agents preserve direct mix and marketing efficiency?
Source: quarterly channel/merchant data. Clean proof can move BKNG
first.
3. FSLY: Can attributable agentic workload reach revenue/RPO? Source:
disclosure and reproducible routing. Required to pass BKNG.
4. Cross-object: Who owns repeat demand and gross profit after
serving, acquisition and fulfillment?

SCENARIO_BRANCHES

CONFIRM: Muse actions repeat/monetize; BKNG channels hold; FSLY lacks
attribution. Keep META, BKNG, FSLY.

SHARED VALUE: Muse grows, BKNG supplies inventory without acquisition
deterioration, FSLY stays strong but unattributed. META remains first;
BKNG can challenge on cash.

INFRASTRUCTURE CONFIRMATION: FSLY discloses a material agentic
workload while retention, RPO conversion and margin hold. Promote;
compare with BKNG on revenue/durability.

BREAK: Muse remains novelty-driven; BKNG shows no damage; FSLY routing
fails replication/economics. Downgrade META, retain BKNG, remove
Muse-specific FSLY optionality.

ATTENTION_TRADE_BOARD

WKAP Attention View
Strongest change: META browser action. Cleanest asymmetry: BKNG’s
reported economics versus hypothetical displacement. Largest gap/fade
risk: FSLY’s Muse read-through. Most crowded: META; positioning
unmeasured. Best durable candidate: META on paid habits; BKNG can
overtake on channel proof.

RADAR_OBJECT_INDEX

THESIS_OBJECT_1: META
THEME: Personal agents / browser action / glasses / subscriptions
STATUS: Product state changed; adoption/economics gate active; Rank 1 promote
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $729.00 (Nasdaq real-time pre-market, 08:38 EDT /
19:38 ICT Sep 24, NASDAQ-GS)
DATE_FIRST_ADDED_TO_RADAR: 2026-07-14
SETUP_TYPE: Action-layer launch requiring repeat-use and monetization proof
ATTENTION_STAGE: Launch recognition entering adoption test
ATTENTION_WINDOW: 1–90 days
KEY_QUESTION: Can Muse make repeated, trusted and economical paid actions?

THESIS_OBJECT_2: FSLY
THEME: Edge delivery / routing / agentic workloads / RPO and retention
STATUS: Standalone/product evidence improving; Muse attribution
unverified; Rank 3 wait
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $28.25 (Nasdaq real-time pre-market, 08:38 EDT /
19:38 ICT Sep 24, NASDAQ-GS)
DATE_FIRST_ADDED_TO_RADAR: 2026-08-06
SETUP_TYPE: Agent-commerce capability plus technical signal requiring
materiality
ATTENTION_STAGE: Product/routing evidence entering attribution test
ATTENTION_WINDOW: Days to two quarters
KEY_QUESTION: Does observed agent traffic become attributable, durable revenue?

THESIS_OBJECT_3: BKNG
THEME: Agentic travel / marketplace demand / direct acquisition /
merchant fulfillment
STATUS: Reported economics intact; customer-ownership test active; Rank 2 hold
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $156.80 (Nasdaq real-time pre-market, 08:37 EDT /
19:37 ICT Sep 24, NASDAQ-GS)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-24
SETUP_TYPE: Proven transaction engine facing a distribution challenge
ATTENTION_STAGE: Disintermediation debate entering channel test
ATTENTION_WINDOW: One to four quarters
KEY_QUESTION: Can Booking retain repeat demand without a higher
agent-acquisition toll?

THESIS OBJECTS

THESIS_OBJECT_1 — META
CARD_ID: WKAP-RADAR-2026-09-24-META
CARD_TITLE: Owning the Agent Matters Only When Actions Become Habits
TYPE: Agent Adoption and Monetization Trackable
THEME: Personal agent / browser action / eyewear / identity / subscriptions
STATUS: Research promoted; repeat-use and monetization gate active
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $729.00 (Nasdaq real-time pre-market, 08:38 EDT /
19:38 ICT Sep 24, NASDAQ-GS)
DATE_FIRST_ADDED_TO_RADAR: 2026-07-14
ATTENTION_STAGE: Launch recognition entering behavioral proof
ATTENTION_WINDOW: 1–90 days

WHY_A_READER_CARES_NOW
META has the verified product state change: an agent that acts through
a browser and extends to glasses. The question moved from model
quality to habit, trust and monetization.

PREVIOUS_STATE
REPORTED_FACT: Broad consumer AI/glasses distribution, but no general
browser-action product with a direct paid path in this evidence set.

CURRENT_STATE
REPORTED_FACT: Muse performs browser actions in a secure virtual
machine, including travel booking; the core is free.
MANAGEMENT_EXPECTATION: Meta plans paid capabilities and glasses
distribution; Meta Audio starts at $349 with October 13 shipping, Gen
3 starts at $449, and Meta expects more than 100 options by year-end.

DECISION_OR_MONITORING_PAYOFF
Keep Rank 1; upgrade on repeat action and paid attachment, downgrade
on weak trust, reliability or economics.

THESIS_SUMMARY
Distribution and identity put META closest to user intent; repeat,
economical behavior must turn capability into earnings.

WKAP_ANGLE
The surface-level frame: Meta launched an agent, so it owns agentic commerce.
The alternative frame: META owns first contact, but completion, trust
and paid attachment determine value.
The key research question: What share completes consequential actions
repeatedly and pays?

CORE_THESIS
REPORTED_FACT: Browser action and glasses distribution.
MANAGEMENT_EXPECTATION: paid capability and hardware expansion.
INTERPRETATION: interface ownership earns first priority.
UNRESOLVED_QUESTION: adoption, reliability, conversion and cost.

ATTENTION_TRADE_FRAME
Attention Source: @cfromhertz,
https://x.com/cfromhertz/status/2102464735463580131
Why Today: INTERPRETATION — Relays institutional attention from
answers toward actions/commerce; proves no adoption.
Attention Stage: Launch → behavioral proof.
Attention vs Evidence — REPORTED_FACT: official capability and
announced pricing. MANAGEMENT_EXPECTATION: rollout, paid capability
and hardware expansion. Primary source(s):
https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/
and https://about.fb.com/news/2026/09/introducing-ray-ban-meta-audio-glasses-new-styles-plus-muse/
INTERPRETATION: institutional framing; no holding disclosed.
Attention Path: Launch → trial → successful action → repeat use → paid
attachment → economics.
Attention Asymmetry: Product ownership verified; contribution unverified.
Crowding Risk: High launch focus; flows unmeasured.
What Could Sustain Attention: Completion, repeat travel/shopping use,
glasses uptake and paid conversion.
What Could Make Attention Fade: Novelty use, errors, trust friction,
slow rollout or high cost.
Attention-to-Thesis Conversion: Repeated successful actions plus paid
attachment at acceptable cost.

WEAKEST_ASSUMPTION
Users trust Muse with consequential actions often enough to create
economic habit.

MOST_IMPORTANT_DATA_POINT
Repeat action completion and paid attachment at a defined user age.

NEXT_DATA_POINT
October 13 rollout, Muse usage, completion reliability, subscription
terms and cost commentary.

THESIS_OBJECT_2 — FSLY
CARD_ID: WKAP-RADAR-2026-09-24-FSLY
CARD_TITLE: Agent-Commerce Capability Does Not Name the Customer
TYPE: Workload Attribution and Materiality Trackable
THEME: Edge delivery / routing / identity / payments / retention / RPO
STATUS: Research option retained; Muse relationship unverified
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $28.25 (Nasdaq real-time pre-market, 08:38 EDT /
19:38 ICT Sep 24, NASDAQ-GS)
DATE_FIRST_ADDED_TO_RADAR: 2026-08-06
ATTENTION_STAGE: Product/routing recognition entering attribution
ATTENTION_WINDOW: Days to two quarters

WHY_A_READER_CARES_NOW
Fastly has improving reported economics and a verified Skyfire
agent-commerce path. Independent routing adds a testable lead; it does
not establish Meta as a customer.

PREVIOUS_STATE
REPORTED_FACT: Turnaround depended on company-wide growth, retention,
RPO conversion and margin, without verified Muse workload.

CURRENT_STATE
REPORTED_FACT: Revenue $183.3 million, +23%; gross margin 63.3%; NRR
117%; RPO $341 million, +38%. Skyfire integrates verified agent
identity and payment-backed transactions with Fastly Compute.
REPORTED_FACT — company-reported observation, not customer
attribution: AI requests grew 30% January–May, about 6.5x
human-traffic growth. ATTENTION: A 738-request test observed Fastly at
29.3%, in every browser/site-navigation test.

DECISION_OR_MONITORING_PAYOFF
Wait at Rank 3; promote only when an attributable workload becomes
durable and financially material without margin erosion.

THESIS_SUMMARY
The standalone case and agent-commerce capability improved;
Muse-specific probability/payoff cannot be inferred from IP paths.

WKAP_ANGLE
The surface-level frame: Fastly appears in agent traffic, so Meta is a customer.
The alternative frame: Product capability and routing are clues;
customer identity, durability and revenue are separate.
The key research question: What disclosed evidence connects observed
paths to material workload?

CORE_THESIS
REPORTED_FACT: Metrics improved; Skyfire supplies agent
identity/payment capability. INTERPRETATION: agentic demand could add
upside. UNRESOLVED_QUESTION: Meta attribution, durability and revenue.

ATTENTION_TRADE_FRAME
Attention Source: @OddDiligence,
https://x.com/OddDiligence/status/2102810323233423720
Why Today: INTERPRETATION — Quantifies routing and supplies a
replication path; IP attribution is not a contract.
Attention Stage: Observation → attribution.
Attention vs Evidence — REPORTED_FACT: Q2 metrics and Skyfire
partnership. Primary source(s):
https://investors.fastly.com/news-releases/news-release-details/fastly-announces-second-quarter-2026-financial-results
and https://investors.fastly.com/news-releases/news-release-details/fastly-and-skyfire-partner-enable-trusted-agentic-commerce
INTERPRETATION: routing distribution; no holding disclosed.
Attention Path: Observe → replicate → attribute → disclose → convert
into RPO/revenue.
Attention Asymmetry: Standalone proof improved; customer-specific
narrative remains weak.
Crowding Risk: High optionality/proof gap; flows unmeasured.
What Could Sustain Attention: Replication, named workload, RPO
conversion and stable margin.
What Could Make Attention Fade: Alternate routing, no disclosure, weak
retention/backlog or margin.
Attention-to-Thesis Conversion: Attributable, repeatable and
financially material workload.

WEAKEST_ASSUMPTION
Observed paths represent durable production demand accruing
economically to Fastly.

MOST_IMPORTANT_DATA_POINT
Attributable agentic workload revenue or RPO reconciled with customer metrics.

NEXT_DATA_POINT
Replication, customer commentary, RPO conversion, NRR, growth and gross margin.

THESIS_OBJECT_3 — BKNG
CARD_ID: WKAP-RADAR-2026-09-24-BKNG
CARD_TITLE: The Booking Layer Is Exposed, Not Yet Displaced
TYPE: Customer Ownership and Transaction Defense Trackable
THEME: Travel marketplace / direct acquisition / merchant rails /
agent distribution
STATUS: Evidence-backed hold; channel test active
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $156.80 (Nasdaq real-time pre-market, 08:37 EDT /
19:37 ICT Sep 24, NASDAQ-GS)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-24
ATTENTION_STAGE: Disintermediation debate entering channel proof
ATTENTION_WINDOW: One to four quarters

WHY_A_READER_CARES_NOW
Muse can book travel, making customer ownership testable. BKNG still
has reported inventory, fulfillment and merchant economics; the
question is whether agents add distribution or a toll.

PREVIOUS_STATE
REPORTED_FACT: Scaled marketplace whose direct mix, marketing
efficiency and merchant conversion determined customer-ownership
quality.

CURRENT_STATE
REPORTED_FACT: Room nights +5%, bookings +9%, revenue +8%, EBITDA +9%.
Marketing was 4.7% of bookings versus 4.6%, direct mix stayed mid-50s,
merchant mix rose to 73% from 69%, and repurchases were $3.7 billion
with $14.5 billion authorized. No supplied evidence establishes
Muse-related loss.

DECISION_OR_MONITORING_PAYOFF
Hold Rank 2. Promote above META if agents prove incremental with
better channel economics; downgrade if customer ownership weakens
first.

THESIS_SUMMARY
BKNG’s engine is proven; agentic discovery makes direct mix and
acquisition efficiency decisive defense metrics.

WKAP_ANGLE
The surface-level frame: An AI travel agent bypasses Booking.
The alternative frame: The agent may call Booking inventory, merchant
and service rails; value depends on repeat demand and acquisition
cost.
The key research question: Does agent traffic expand bookings at
stable economics or reduce direct share?

CORE_THESIS
REPORTED_FACT: Growth and merchant mix intact; direct mix flat and
marketing intensity slightly higher. INTERPRETATION: distribution
test, not displacement. UNRESOLVED_QUESTION: channel attribution and
take economics.

ATTENTION_TRADE_FRAME
Attention Source: @DrewCohenMoney,
https://x.com/DrewCohenMoney/status/2102776178503500255
Why Today: INTERPRETATION — Retests the travel-agent thesis; proves no
lost room nights or take-rate change.
Attention Stage: Debate → channel proof.
Attention vs Evidence — REPORTED_FACT: room nights, bookings, revenue,
EBITDA, channel mix, marketing and repurchases. Primary source(s):
https://www.sec.gov/Archives/edgar/data/1075531/000107553126000036/q2-26bkngearningsrelease.htm
INTERPRETATION: agent-versus-marketplace debate; no holding disclosed.
Attention Path: Capability → usage → channel attribution →
direct/marketing change → cash impact.
Attention Asymmetry: Current economics reported; displacement unproven.
Crowding Risk: Medium narrative concentration; flows unmeasured.
What Could Sustain Attention: Agent partnerships, channel movement or
evidence agents own repeat demand.
What Could Make Attention Fade: Stable direct mix, efficient marketing
and continued room-night growth.
Attention-to-Thesis Conversion: Persistent shift in acquisition cost,
repeat behavior or merchant contribution.

WEAKEST_ASSUMPTION
Inventory and merchant rails preserve bargaining power when another
interface owns the first prompt.

MOST_IMPORTANT_DATA_POINT
Direct room-night mix beside marketing expense as a percentage of bookings.

NEXT_DATA_POINT
Direct share, marketing ratio, room-night growth, merchant mix and
agent-channel disclosure.

CROSS_OBJECT_ATTENTION_COMPARISON

Cleanest Attention Trade: BKNG; reported economics, observable falsifiers.
Most Evidence-Backed Attention Trade: META capability; BKNG current economics.
Most Crowded Attention Trade: META launch; positioning unmeasured.
Highest Fade Risk: FSLY’s Muse read-through.
Best Durable Thesis Candidate: META on paid habits; BKNG can overtake
on channel proof.

7_DAY_RESEARCH_WORKFLOW

META — 7-Day Checks
1. Capture access, actions, failures, privacy and subscription terms.
2. Test travel booking/change/service; record completion/intervention.
3. Baseline browser versus October 13 glasses rollout.

FSLY — 7-Day Checks
1. Reproduce routing across time/geography/task; preserve raw paths.
2. Test asset, intermediary, cache and shared-service explanations.
3. Reconcile Skyfire/AI traffic with revenue, NRR, RPO and margin;
infer no customer.

BKNG — 7-Day Checks
1. Baseline demand, channel mix, marketing, EBITDA and cash.
2. Map agent calls to inventory/payment/service and repeat ownership.
3. Define first break: direct decline, higher marketing or weaker growth.

Cross-Object — 7-Day Checks
1. Separate attention from proof.
2. Compare paid action, attributable workload and profitable booking
checkpoints.
3. Re-rank on status-changing evidence; log falsifiers.

Primary-source retrieval registry:
META=https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/|https://about.fb.com/news/2026/09/introducing-ray-ban-meta-audio-glasses-new-styles-plus-muse/;FSLY=https://investors.fastly.com/news-releases/news-release-details/fastly-announces-second-quarter-2026-financial-results|https://investors.fastly.com/news-releases/news-release-details/fastly-and-skyfire-partner-enable-trusted-agentic-commerce|https://quic.fastly.com/blog/how-fastly-skyfire-enable-trusted-agentic-commerce-at-the-edge;BKNG=https://www.sec.gov/Archives/edgar/data/1075531/000107553126000036/q2-26bkngearningsrelease.htm
Attention-source retrieval registry:
META=https://x.com/cfromhertz/status/2102464735463580131;FSLY=https://x.com/OddDiligence/status/2102810323233423720;BKNG=https://x.com/DrewCohenMoney/status/2102776178503500255
Evidence-state snapshot: META=action verified/economics
pending;FSLY=company and Skyfire evidence improving/Muse attribution
pending;BKNG=transactions intact/channel effect
pending;classification=REPORTED_FACT|MANAGEMENT_EXPECTATION|INTERPRETATION|UNRESOLVED_QUESTION;rule=rank
proof-adjusted value capture;source-rule=primary sets status,
attention sets retrieval priority

30_DAY_RESEARCH_WORKFLOW

META — 30-Day Checks
1. VALIDATE: repeat use, successful actions, expanding access, no
material trust failure.
2. UPGRADE: paid attachment with acceptable reliability/cost.
3. DOWNGRADE: novelty use, high intervention, rollout delay or privacy friction.
4. INVALIDATE: no repeat habit or economical monetization.

FSLY — 30-Day Checks
1. VALIDATE: routing replicates while NRR, RPO conversion and margin hold.
2. UPGRADE: attributable agentic workload becomes durable and material.
3. DOWNGRADE: inconsistent signal, weak RPO/retention or margin.
4. INVALIDATE: no replication or relevant economics after the disclosure cycle.

BKNG — 30-Day Checks
1. VALIDATE: growth, direct mix, marketing, merchant contribution and cash hold.
2. UPGRADE: agents add demand while repeat ownership and efficiency improve.
3. DOWNGRADE: direct mix falls or marketing rises before merchant offsets.
4. INVALIDATE: persistent agent-channel loss reduces growth, take
economics or cash.

Cross-Object — 30-Day Checks
1. Re-rank on economics, not price or narrative.
2. Compare gross profit after serving, edge, acquisition and fulfillment costs.
3. Keep separate proof clocks and append every transition.

WKAP DAILY TOP 3

Three market sources worth feeding into today’s market chat. Not
required reading — WKAP has already extracted the signal.

1. @cfromhertz — Muse Moves the Debate From Answers to Actions
URL: https://x.com/cfromhertz/status/2102464735463580131
WKAP signal: INTERPRETATION — Institutional attention shifts toward
browser action and commerce.
Why it matters today: Identifies attention; proves no adoption,
subscription economics or vendor relationship.
Themes/tickers: Agents; browser action; subscriptions; eyewear; $META.
Question to ask: What repeat-action and paid-attachment evidence makes
Muse an earnings variable?

2. @OddDiligence — A Routing Clue, Not a Customer Contract
URL: https://x.com/OddDiligence/status/2102810323233423720
WKAP signal: INTERPRETATION — The 738-request test observed Fastly at
29.3% and in every tested browser/site-navigation case.
Why it matters today: Creates a replication queue while preserving the
boundary between IP paths and customer/revenue proof.
Themes/tickers: Edge; routing; agentic workloads; attribution; $FSLY; $META.
Question to ask: Which test and primary disclosure establish
attributable, material Fastly demand?

3. @DrewCohenMoney — BKNG Retests the Travel-Agent Thesis
URL: https://x.com/DrewCohenMoney/status/2102776178503500255
WKAP signal: INTERPRETATION — Reopens who owns travel discovery when
an external agent books.
Why it matters today: Directs research to direct share, marketing and
merchant contribution; proves no displacement.
Themes/tickers: Travel marketplace; acquisition; merchant rails; $BKNG; $META.
Question to ask: Do agents add bookings at stable economics or tax
direct acquisition and repeat ownership?

WKAP Radar Feed - 2026 - 09 - 24

Agent-readable facts