HERMES INTELLIGENCE  /  VEXTRUM
READ-ONLY DEMO ROOM · GLP-1 OBESITY MARKET INTELLIGENCE
Back to Demo Index Request Pilot
PUBLIC DEMO · ILLUSTRATIVE · NOT INVESTMENT ADVICE
Hermes Intelligence · Vextrum Case Study

GLP-1 Obesity Market
Intelligence Room

A healthcare PM wants to know whether GLP-1 obesity leaders are still structurally underappreciated - or whether payer friction, net price pressure, compounding dynamics, supply limits, and next-generation competitors will compress the upside. Vextrum turns the debate into a continuously updating intelligence universe: sources, entities, hypotheses, evidence, deltas, and analyst actions.

Illustrative case study. Public-source examples used for demonstration. Not investment advice.

Vextrum does not issue buy/sell recommendations. It structures public-source evidence, thesis movement, confidence, contradictions, and analyst actions for PM review.

41 source families mapped
FDA · CMS · PBM · payer · employer · clinical · company · KOL · pipeline
126 entities resolved
companies, drugs, indications, payers, regulators, trials, policies, competitors
5 investment debate pillars
Demand · Access · Pricing · Competition · Execution
18 active thesis deltas
evidence-weighted updates to bull / bear / base case
Morning Delta Evidence Pack Red Team Challenge Analyst Task PM Brief
Designed for a discretionary healthcare PM evaluating long/short exposure, catalyst risk, and thesis confidence - not a generic news feed.
How to Read This Case Study in 3 Minutes
1. THE DEBATE
A healthcare PM tests whether GLP-1 leaders remain structurally underappreciated or whether access, pricing, compounding, and competition are compressing the upside.
2. THE SYSTEM
Vextrum builds a PM-defined intelligence universe: sources, entities, products, hypotheses, evidence rules, alert materiality, and output cadence.
3. THE KILLER EXAMPLE
A single FDA compounding event is not treated as automatically bullish or bearish. Vextrum decomposes it: regulatory overhang down, substitute risk down, branded conversion uncertain, affordability unchanged.
4. THE PILOT
A 2–4 week pilot proves whether Vextrum surfaces useful deltas faster than manual monitoring, reduces analyst time, and produces auditable PM-ready evidence packs.

Bull Case

Volume expansion, broader cardiometabolic indications, employer adoption, supply normalization, oral formulations

Bear Case

Coverage restrictions, price compression, compounded substitutes, supply bottlenecks, next-gen competitors

Variant Perception

Market may be overweighting headline demand and underweighting access / pricing / persistence deltas

Problem
Thesis
Questions
Config
Discovery
Sources
Ontology
Launch
Command
Sit. Room
Delta
Evidence
Arch
Pilot

3-Minute Talk Track

Want to see how the output looks like? Click here →
OPEN
We are not showing you a GLP-1 dashboard. We are showing how Vextrum turns one live PM debate into an evidence-weighted intelligence room.
PROBLEM
The PM question is not 'what happened in GLP-1 news?' It is whether the market is still underestimating long-duration demand or overestimating revenue quality given access, pricing, compounding, and competition.
SYSTEM
Vextrum converts that debate into a structured intelligence universe: sources, entities, products, hypotheses, evidence rules, confidence movement, Red Team objections, analyst tasks, and PM briefs.
KILLER
The FDA compounding example is the key. Most systems would mark it simply as bullish for incumbents. Vextrum decomposes it: regulatory overhang down, substitute risk down, but branded demand conversion unresolved and affordability unchanged.
VALUE
That creates a Red Team objection, analyst task, and PM brief update. The system does not recommend buy or sell; it shows which part of the thesis moved, what remains unknown, and what evidence the analyst should collect next.
ASK
Give us one live healthcare research controversy. In 2-4 weeks, we will show whether Vextrum can build a source-linked, evidence-weighted intelligence room that surfaces useful deltas faster than manual monitoring and reduces analyst time.

Durable GLP-1 Supercycle or Consensus Peak?

The platform starts not from a dataset, but from an investment controversy. The PM defines the debate, the securities universe, the critical assumptions, and the signals that would change their view.

Investment Debate

Primary Question
Are GLP-1 obesity leaders still mispriced because the market underestimates long-duration volume expansion, or are consensus estimates too aggressive given payer, price, competition, and regulatory pressure?
What evidence supports continued volume expansion beyond current consensus?
Which access restrictions or reimbursement changes could slow adoption?
Are net price assumptions vulnerable to Medicare, commercial payer, PBM, or employer pressure?
Is compounding/channel risk declining or merely shifting?
Which competitors could erode incumbent share or pricing power?
Which clinical, safety, persistence, or tolerability signals change the durability of demand?
Are international access/pricing dynamics supportive or margin-dilutive?

Decision Use Case

LLY · NVO · ZEAL · AMGN · PFE · VKTX · ALT
Securities Watched
Zepbound · Mounjaro · Wegovy · Ozempic · Rybelsus
Products Watched
Daily · Weekly · Event
Decision Cadence
Evidence map · PM alerts · Red Team · Packs
Output Required
Book Context
Position context: illustrative long exposure to GLP-1 incumbents, with active monitoring of competitor/payer risks.
Objective: detect evidence that changes conviction, sizing, hedging, or catalyst preparation.
Risk types: thesis risk, crowding risk, access risk, pricing risk, pipeline risk, execution risk.
No confidential holdings. No investment advice. No prediction of stock direction.
Variant Perception
Consensus sees demand. The PM wants to know whether access/pricing/persistence/competition deltas are moving faster than the market is updating estimates.
Known Blind Spots
Employer coverage pullback not captured in headline Rx growth
PBM/formulary access changes buried in PDFs
Compounder/telehealth channel shifts after FDA/regulatory actions
Competitor trial readouts reframing differentiation
International reimbursement decisions becoming read-throughs for US/EU payer behavior

AI-Guided Thesis Capture

Vextrum converts a PM's ambiguous research concern into a machine-readable intelligence program: watchlist, hypotheses, source priorities, event taxonomy, confidence scoring, and output cadence.

Topics (12)
Investment Debate
Securities Universe
Product Universe
Critical Assumptions
Bull Case
Bear Case
Variant Perception
Source Priorities
Event Taxonomy
Materiality Rules
Evidence Standard
Delivery Cadence
Vextrum Onboarding Agent
Let's define the investment controversy. Is the main question demand durability, reimbursement risk, competitive entry, pricing pressure, or all of the above?
All of the above. We are not looking for a GLP-1 news feed. We need a living debate map that tells us whether the long LLY/NVO thesis is gaining or losing evidence.
Understood. I'll structure the case into five debate pillars: Demand Expansion, Access & Reimbursement, Pricing & Margin, Competitive Pipeline, Execution & Supply. Which names and products are in scope?
Core: Novo and Lilly. Watch: Zealand/Boehringer, Amgen, Viking, Altimmune, Pfizer, and relevant PBM/payer/employer signals. Products include Wegovy, Ozempic, Rybelsus, Zepbound, Mounjaro, oral GLP-1s, dual/triple agonists, and MASH-related obesity assets.
What evidence would change your view?
Commercial coverage changes, CMS/FDA actions, employer benefit changes, PBM formulary shifts, trial readouts that change differentiation, management commentary on supply/net price, and signs that compounding or telehealth channels are structurally changing.
I'm generating an intelligence configuration: securities watchlist, product ontology, source registry, event taxonomy, scoring rubric, and PM alert templates. I'll also create Red Team challenge rules so the system does not simply confirm the dominant thesis.
OPERATING PRINCIPLE
I will not treat any single event as bullish or bearish by default. Each signal will be decomposed into: event fact, affected entities, source authority, hypothesis impact, confidence, materiality, unknowns, Red Team objection, and required analyst action. The PM decides interpretation - the system provides structured evidence.
Type your answer...
Thesis Capture Output

The onboarding agent generates a machine-readable intelligence program from the PM conversation — covering the full debate structure, source priorities, and output requirements.

Investment Debate Map

5 pillars: Demand, Access, Pricing, Competition, Execution. Bull/bear/variant branches defined.

Securities & Product Watchlist

LLY, NVO, ZEAL, AMGN, VKTX, ALT, PFE + Zepbound, Wegovy, Ozempic, Mounjaro, Rybelsus, pipeline assets

Hypothesis Register

7 hypotheses (H1-H7) with bull/bear classification, initial confidence, evidence requirements

Event Taxonomy

25+ event types: regulatory, payer/access, pricing, clinical, competitor, company, channel, international

Source Priority Matrix

4-tier authority: FDA/CMS/filings (T1), PBM/ICER/journals (T2), trade press (T3), KOL/social (T4)

Evidence & Confidence Rules

Source-linked, confidence-scored, hypothesis-tagged. Weak signals cannot auto-upgrade conclusions.

Alert Materiality Rules

FDA/CMS action → immediate. PBM restriction → analyst review. Competitor readout → thread update.

PM Brief Template

Weekly: exec summary, what changed, confidence movement, bull/bear evidence, Red Team, watchlist

Confirm the GLP-1 Investment Debate Questions

Vextrum presents the investment debate questions derived from the thesis capture. Each question is mapped to a pillar, evidence requirement, and materiality threshold. The PM reviews, edits, and approves before they drive the intelligence universe.

Generated Strategic Questions 8 QUESTIONS · AWAITING APPROVAL
Is demand expansion still accelerating faster than access friction?
Category: Demand · Pillars: Demand, Access
Evidence: Rx proxies, employer adoption, persistence data, coverage breadth · Trigger: source-confirmed demand or access change · Output: Demand thread + access comparison task
Edit Remove
Are net price assumptions for Novo/Lilly vulnerable to Medicare, commercial payer, PBM, or employer pressure?
Category: Pricing · Pillars: Pricing, Access
Evidence: CMS pricing docs, PBM policy updates, management commentary, employer benefit changes · Trigger: source-confirmed pricing/access change touching named product · Output: Pricing thread + model sensitivity task
Edit Remove
Is compounding/channel risk being reduced by regulatory action or simply migrating into harder-to-measure channels?
Category: Channel · Pillars: Execution, Access
Evidence: FDA 503B notices, telehealth commentary, pharmacy enforcement, 503A workarounds · Trigger: regulatory action or channel shift signal · Output: Compounding thread + channel migration task
Edit Remove
Which competitor assets can credibly challenge incumbents on efficacy, tolerability, route, muscle preservation, MASH benefit, or price?
Category: Competition · Pillars: Competition, Demand
Evidence: Phase II/III readouts, conference abstracts, mechanism differentiation data · Trigger: efficacy/tolerability data exceeding incumbent benchmarks · Output: Competitor matrix + terminal share task
Edit Remove
Are oral GLP-1s and next-gen agonists expanding the market or cannibalizing premium injectable economics?
Category: Pipeline · Pillars: Competition, Pricing
Evidence: oral GLP-1 readouts, pricing language, payer commentary, adherence/persistence signals · Trigger: oral formulation data with pricing/access implications · Output: Oral/next-gen thread + mix dilution analysis
Edit Remove
What evidence would force a revision to 2027/2028 consensus sales assumptions?
Category: Materiality · Pillars: All
Evidence: T1 pricing/access/clinical events combined with management commentary · Trigger: multiple converging signals across pillars · Output: Thesis review + cross-pillar evidence pack
Edit Remove
Which signals are likely visible in primary/public sources before they appear in sell-side consensus?
Category: Edge · Pillars: All
Evidence: regulatory filings, payer policy changes, employer benefit shifts, conference data before sell-side coverage · Trigger: T1/T2 source update preceding consensus revision · Output: PM alert + information advantage assessment
Edit Remove
What Red Team evidence could invalidate the current bull case?
Category: Red Team · Pillars: All
Evidence: PA tightening, persistence failure, competitor superiority, management net-price admission · Trigger: any T1/T2 evidence supporting bear case · Output: Red Team challenge + falsification review
Edit Remove
+ Add Custom Question Approve All & Generate Config → These questions will drive source discovery, ontology generation, hypothesis scoring, and the five debate pillars.

Generated Configuration Review

Vextrum generates a first-pass intelligence configuration that can be reviewed, edited, and approved. Every parameter traces back to the onboarding conversation.

Approve Configuration → All 8 sheets validated. Configuration will drive source discovery, ontology generation, and intelligence pipeline.

Autonomous Source Discovery for the GLP-1 Debate

Vextrum does not start from a fixed feed. It discovers, classifies, scores, and proposes a source universe based on the PM's actual investment debate.

INTELLIGENCE NETWORK BUILDER SESSION: VX-HC-0847 LIVE

Approved Source Portfolio

The PM's custom intelligence infrastructure — sources discovered, classified, scored, and mapped to the five GLP-1 debate pillars. Not a generic feed; a purpose-built research network.

SOURCE INTELLIGENCE NETWORK 41 SOURCES · 5 DEBATE PILLARS · 4 GEOGRAPHIES Export CSV Export PDF
Coverage: 100% PASS Blind Spots: PASS Geographic: PASS
Strategic Questions Sources
CORE 10 — Highest Authority Sources These sources form the foundation of your intelligence universe
FDA GLP-1 / Obesity Notices
RegulatoryUS
Primary source for GLP-1 approvals, compounding actions, label changes. Tier 1. Authority: 97%.
CMS Medicare / GLP-1 Bridge
AccessUS
Medicare drug price negotiation, GLP-1 coverage determinations. Tier 1. Authority: 95%.
ClinicalTrials.gov
ClinicalGLOBAL
Pipeline monitoring for GLP-1 competitor trials, endpoint changes, enrollment. Tier 1.
Eli Lilly IR / SEC EDGAR
CompanyUS
10-K/Q, earnings transcripts, supply/pricing commentary. Tier 1. Authority: 93%.
Novo Nordisk IR / SEC EDGAR
CompanyUS/DK
Annual report, transcripts, Wegovy/Ozempic supply/access commentary. Tier 1.
PBM Formulary Policies
AccessUS
PBM coverage, PA requirements, step-edit policies for GLP-1 products. Tier 2. Authority: 88%.
ICER / KFF Evidence Reports
Access/HTAUS
Value assessments, affordability analysis, payer impact. Tier 2. Authority: 91%.
ADA / EASD / ObesityWeek
ClinicalGLOBAL
Conference abstracts: efficacy, tolerability, differentiation data. Tier 1/2.
STAT News / Endpoints News
Trade PressUS
Premium GLP-1 coverage: regulatory, clinical, competitive. Tier 3. Authority: 89%.
FDA 503B Compounding Notices
ChannelUS
Compounding exclusion proposals, enforcement actions. Tier 1 for channel risk.
EXTENDED 30 95% coverage
20 Additional GLP-1 Intelligence Sources
The Extended 30 includes additional GLP-1 sources: EMA/NICE obesity decisions, employer benefit surveys, Zealand/Amgen/Viking/Pfizer company pages, additional PBM policy databases, trade press (Fierce Pharma, BioSpace, AJMC), and telehealth channel commentary.
EMA · 84% Fierce Pharma · 78% BioSpace · 75% AJMC · 73% Employer · 72% +15 more
COMPLETE 50 100% coverage
10 Deep-Coverage GLP-1 Sources
The Complete 50 adds specialized GLP-1 sources: state Medicaid formularies, employer benefit databases, international reimbursement authorities (PMDA, PBAC), additional clinical registries, and pharmacy/provider channel intelligence.
BEYOND 80 114% coverage
5 Weak Signal / Peripheral Sources
The Beyond 80 tier includes KOL commentary, physician community signals, patient access discussions, telehealth channel analytics, and social signals. Tier 4 only — cannot auto-upgrade high-confidence conclusions.
Gap Analysis
CRITICAL
Employer coverage pullback — only 2 sources detected. Employer benefit surveys and large employer coverage databases needed.
MODERATE
Commercial payer policies — moderate coverage but fragmented across hundreds of PDFs. Change detection partially covered.
MINOR
Compounding/telehealth channel — regulatory sources strong, but telehealth substitution behavior weakly covered.
Provenance Rules
Every claim must be traceable to a source document with URL/citation slot
Every event stores source tier, timestamp, extraction method, entity mapping, and audit record
Weak signals cannot alter high-confidence thesis states without corroboration
Source updates without old/new comparison are baseline updates, not material deltas
Material alerts require T1/T2 source authority, cross-source corroboration, or human approval
Contradictory evidence creates a contradiction object, not a forced conclusion
9
Demand
10
Access
7
Pricing
8
Competition
7
Execution

Ontology: Turning GLP-1 Research Into Structured Knowledge

Every document is transformed into entities, relationships, evidence snippets, hypothesis links, and confidence updates. The ontology maps the full GLP-1 investment debate.

Entity Types (21)
COMPANY
Novo, Lilly, Zealand, Amgen, Viking, Altimmune, Pfizer
DRUG_PRODUCT
Wegovy, Zepbound, Ozempic, Mounjaro, Rybelsus, pipeline
MECHANISM
GLP-1, GIP, glucagon agonist, dual/triple agonist
INDICATION
Obesity, T2D, MASH, cardiovascular, CKD
CLINICAL_TRIAL
Phase I-IV trials, endpoints, readouts
REGULATORY_ACTION
FDA, CMS, EMA actions on GLP-1 products
COVERAGE_POLICY
PBM, payer, Medicare, employer coverage
PRICING_ACTION
Net price, gross-to-net, negotiated pricing
PAYER_PBM
CVS Caremark, Express Scripts, Optum Rx, CMS
COMPETITOR_ASSET
survodutide, MariTide, VK2735, pemvidutide
SAFETY_SIGNAL
Adverse events, tolerability, persistence
Predicates / Relationships
DRUG → owned_by → COMPANY
DRUG → targets → MECHANISM
TRIAL → evaluates → DRUG
REGULATORY_ACTION → affects → DRUG
COVERAGE_POLICY → restricts → DRUG
PRICING_ACTION → pressures → NET_REVENUE
COMPETITOR_ASSET → threatens → INCUMBENT_DRUG
SIGNAL → supports/refutes → HYPOTHESIS
Provenance
Ontology Version: 3.0 (Catalog-First)
Generated from: Approved universe config
Catalog-backed predicates: 135 relationship types
Config hash: sha256:e4f7a2...

Mission Launch: GLP-1 Intelligence Room Activated

With the GLP-1 intelligence universe approved, ontology generated, and sources mapped — the pipeline activates. The system begins continuous monitoring of all approved sources.

V
Intelligence Universe Active
Pipeline heartbeat established — continuous monitoring across 41 approved source families
Triage Engine
Active

Strategic Judge screening incoming GLP-1 signals against debate pillars. Multi-stage hybrid semantic + reranker scoring.

Extraction Engine
Active

Entity resolution mapping drug names, brands, generics, mechanisms, tickers. Multi-stage pipeline: normalize → exact → alias → fuzzy → LLM → create.

Synthesis Engine
Active

Bayesian Nexus with confidence-weighted hypothesis scoring, causal chain detection, and Red Team falsification against the GLP-1 thesis.

First 72 Hours After Launch
Hour 0
Universe approved, pipeline activated
Hour 4
Source crawlers initialized across 41 families
Hour 8
Entity resolution baseline created (126 entities)
Hour 12
First evidence map generated
Hour 24
Initial Situation Room threads created
Hour 48
Red Team objections generated
Hour 72
First PM-ready weekly brief produced

Illustrative timeline — actual timing depends on source availability and crawl schedule

Command Layer: The PM's GLP-1 Debate Dashboard

The daily command surface for the GLP-1 investment debate. What changed, which hypothesis gained or lost evidence, and what requires analyst action — organized around the five debate pillars.

SITUATION ROOM 3 decision-ready threads
CRITICAL
Compounding Risk: FDA Action Reduces One Bear Overhang, Channel Substitution Unresolved
FDA 503B proposal affects compounded semaglutide/tirzepatide. Branded recapture uncertain — track 503A workarounds and patient churn.
ACTION: Update channel thread · Monitor comment period · Track Rx proxies →
FORMING
CMS/Medicare Pricing: Access Expansion and Price Compression Must Be Modeled Together
CMS pricing signals may broaden access but create net-price anchor. Volume vs margin trade-off unresolved.
WATCHING
Competitor Pipeline: Survodutide Phase III Adds Differentiation Pressure
Positive dual agonist data updates competition branch. Incumbents retain scale advantage. Monitor commercial timing.
THE DELTA 18 thesis deltas since your last session
REGULATORY 2 HRS AGO
FDA 503B proposal detected: semaglutide, tirzepatide, liraglutide compounding exclusion. Reduces 503B compounding overhang but demand conversion uncertain.
Impact: Updates Execution & Supply pillar. Compounding risk thread moved to active monitoring.
→ Open Situation Room · Review evidence pack
PRICING 4 HRS AGO
CMS GLP-1 Bridge page updated. Medicare drug price negotiation includes Ozempic/Rybelsus/Wegovy.
Impact: Updates Access & Pricing pillars. Creates model sensitivity task for volume vs net price.
→ View pricing thread · Create sensitivity analysis
COMPETITOR 6 HRS AGO
Survodutide (Zealand/Boehringer) Phase III obesity data reported positive. Next-gen dual agonist differentiation.
Impact: Updates Competition pillar. Raises competitor pressure confidence. Add to differentiation matrix.
→ View competitor matrix · Compare vs Zepbound/Wegovy
+ 15 more deltas · View all →
Demand Expansion6 active signals · 3 entities
20% WEIGHT
Latest Alert
Prescription volume proxies show continued demand acceleration. Employer adoption signals mixed — large employers adding coverage but with PA requirements.
Current Stance
Bullish evidence accumulating, but persistence and access confidence mixed. Next: validate persistence proxy, compare employer coverage commentary across sources.
Access & Reimbursement8 active signals
25% WEIGHT
CMS Medicare GLP-1 Bridge page update detected. PBM formulary PA language flagged. Mixed signals: broader access possible but restriction language and net price trade-offs matter.
Pricing & Margin5 active signals
20% WEIGHT
Medicare negotiated pricing anchor for Ozempic/Rybelsus/Wegovy. Management commentary on gross-to-net extracted. High materiality, uncertain direction.
Competitive Pipeline7 active signals
20% WEIGHT
Survodutide Phase III positive data. MariTide dosing differentiation. VK2735 oral/injectable pipeline. Monitoring as differentiation threats, not immediate displacement.
Execution & Supply4 active signals
15% WEIGHT
FDA 503B compounding proposal. Supply commentary extracted. Compounding overhang may decline but channel/persistence remains watch item.
Thesis Confidence Movement (Illustrative)
66from 62↑4
Bull Thesis
45from 48↓3
Bear Thesis
61from 57↑4
Access Risk
70from 68↑2
Pricing Pressure
49from 42↑7
Competition Risk
52from 55↓3
Execution/Channel
Illustrative confidence movement based on evidence weighting, not price target or investment recommendation.

Situation Room: Evidence-Weighted Debate Threads

The Situation Room synthesizes evidence into thesis-updating threads. Each thread carries bull/bear interpretations, confidence scores, evidence packs, Red Team triggers, and recommended analyst actions.

CRITICAL · UPDATED TODAY Illustrative example
Same Event, Two Thesis Impacts: FDA Compounding Action Reduces 503B Overhang but Does Not Prove Branded Demand Conversion
Many systems would mark this simply as bullish for incumbents. Vextrum decomposes it: the regulatory event is high-confidence, the revenue impact is uncertain. The system separates what is known from what is assumed.
503B regulatory uncertainty: down
Unauthorized substitute risk: down
? Branded volume recapture: uncertain
= Affordability constraint: unchanged
? 503A/local channel migration: unresolved
? Margin impact: unresolved
RED TEAM: Do not upgrade revenue confidence until channel migration or branded prescription evidence appears. Sentiment is not revenue.
BULL INTERPRETATION
Regulatory tightening supports branded volume recapture and reduces unauthorized substitute risk. Incumbents retain manufacturing scale, brand recognition, and physician familiarity.
CONFIDENCE: 68% · 6 SOURCES · H4 UPDATED
BEAR INTERPRETATION
Cost-sensitive demand may not fully convert to branded products. Affordability remains a constraint. 503A workaround potential. Patient churn risk in cost-sensitive populations.
CONFIDENCE: 55% · 4 SOURCES · DEMAND CONVERSION UNRESOLVED
68%
6 SOURCES · RED TEAM ACTIVE · 2 PILLARS
Decision Options
Monitor Comment Period Track Channel Migration Compare Branded Rx Proxies Update Weekly PM Brief
OPEN FULL PROOF PATH → 9-step pipeline: source event → extracted claim → entity mapping → thesis decomposition → Red Team → analyst task → PM brief
Source fact: high Entity mapping: high Regulatory impact: med-high Revenue conversion: unresolved
Evidence Timeline
2 DAYS AGO FDA 503B proposal published: semaglutide, tirzepatide, liraglutide excluded from bulks list WIRE
1 WEEK AGO Company response: Novo reaffirmed patient access programs and supply commitments WIRE
2 WEEKS AGO Trade press analysis: compounding market sizing, patient count estimates EXTENDED
3 WEEKS AGO Telehealth platform commentary on channel adaptation strategies EXTENDED
FORMING Pillars: Access, Pricing · Entities: NVO, LLY, Ozempic, Rybelsus, Wegovy · H2, H3
CMS / Medicare Pricing: Access Expansion and Price Compression Must Be Modeled Together
CMS pricing/access signals may support broader patient access but create a net-price anchor. Vextrum keeps volume expansion evidence separate from margin compression evidence - the same CMS action updates both bull and bear branches.
Bull: Broader coverage increases eligible lives and strengthens long-duration adoption
Bear: Negotiated pricing and payer anchoring may compress net realized economics
Unknown: whether access expansion comes with sufficient price concession to offset volume gains
61%
5 SOURCES · T1: CMS.gov, Medicare.gov · T2: ICER, KFF
Create Sensitivity TableMap Volume vs PriceTrack CMS UpdatesAdd to PM Brief
WATCHING Pillars: Competition, Demand · Entities: ZEAL, BI, survodutide, AMGN, MariTide · H5
Competitor Pipeline: Survodutide Readout Adds Differentiation Pressure, Not Immediate Displacement
Positive Phase III data updates the next-gen competition branch. Vextrum avoids headline overreaction by mapping efficacy, tolerability, route, timing, regulatory path, and commercial partner separately.
Bull for incumbents: Scale, manufacturing, brand, prescribing familiarity, and indication breadth
Bear: Differentiated dual agonists may pressure terminal market share/pricing
Red Team: reduce immediate materiality because commercial timing is 2+ years and label path uncertain
49%
7 SOURCES · HIGH CONFIDENCE on data, uncertain on commercial timing
Build Competitor MatrixView Trial DataCompare Efficacy/Tolerability/Dosing
CRITICAL Pillars: Access, Demand · Entities: PBMs, employers, LLY, NVO, commercial payers · H2
Employer/Payer Coverage: Headline Demand Is Not the Same as Reimbursed Demand
Headline Rx demand is strong, but employer benefit design, PA requirements, BMI thresholds, and commercial payer coverage determine reimbursed demand. Vextrum tracks changes buried in payer/employer documents and links them to products and assumptions.
Bull: Coverage broadening can unlock large patient pools; employer adoption accelerating
Bear: PA, BMI thresholds, comorbidity requirements, and discontinuation rules slow paid adoption
Blind spot: employer-level coverage data is fragmented and sourcing is partial (2 T3 sources)
57%
4 SOURCES · EVIDENCE GAP: employer coverage · T2-T3
Create Payer/Employer WatchlistBuild Evidence PackRequest Source Discovery
WATCHING Pillars: Demand, Pricing, Competition · Entities: LLY, NVO, PFE, oral sema, orforglipron · H6, H7
Oral GLP-1 / Next-Generation Formulations: TAM Expansion or Mix Dilution?
Oral formulations can expand adoption (access-friendly, needle-averse patients) but may also reprice the category or cannibalize premium injectable economics. Vextrum treats this as a two-sided thesis branch.
Bull: Oral access expands TAM, reduces needle-aversion barrier, improves persistence
Bear: Lower-priced oral products may cannibalize premium injectable mix
Action: monitor oral GLP-1 readouts, pricing language, payer commentary, adherence/persistence signals
44%
3 SOURCES · EARLY STAGE · Monitoring
Monitor Oral ReadoutsTrack Pricing Language

The Delta: What Changed, Why It Matters, What To Do Next

The Delta is a live tape of thesis-updating events. Every entry explains what changed, which pillar and hypothesis are affected, how confidence moved, and what analyst action is required.

18 THESIS DELTAS · LAST 48 HOURS
All Bull Bear Regulatory Access Pricing Competition Requires Review
2 HRS AGO
REGULATORY_DELTA · COMPOUNDING
Source: T1 · Entities: FDA, semaglutide, tirzepatide, liraglutide, LLY, NVO · H4 updated · 54 → 62 · Materiality: Medium
FDA 503B proposal: semaglutide/tirzepatide/liraglutide excluded from bulks list. Reduces 503B regulatory uncertainty. Demand conversion unresolved.
→ Open evidence pack · Monitor channel migration · Update compounding thread
4 HRS AGO
PRICING_ACCESS_DELTA · CMS
Source: T1 · Entities: CMS, Ozempic, Rybelsus, Wegovy, NVO · H2, H3 updated · Access 57 → 61 / Price 68 → 70
CMS Medicare negotiation source updated. Ozempic/Rybelsus/Wegovy price applicability year 2027. Access expansion and price compression branches both updated.
→ Create model sensitivity task · Map volume vs price trade-off
6 HRS AGO
COMPETITOR_SIGNAL · PIPELINE
Source: T1/T2 · Entities: Zealand, BI, survodutide · H5 updated · 42 → 49 · Materiality: Medium (timing distant)
Survodutide Phase III positive obesity data. Efficacy/tolerability differentiation vs semaglutide/tirzepatide. Commercial timing 2+ years.
→ Build competitor matrix · Compare efficacy/tolerability/dosing · Red Team: reduce immediate materiality
8 HRS AGO
COVERAGE_SIGNAL · PAYER
Source: T2 · Entities: commercial payer, GLP-1 class · H2 updated · Confidence: pending diff review · Materiality: pending
Commercial payer policy page changed - GLP-1 PA language flagged. Old/new diff not yet verified.
→ Analyst review required · Do not classify as material until diff verified
12 HRS AGO
MANAGEMENT_COMMENTARY · EARNINGS
Source: T1 · Entities: NVO, LLY, supply, gross-to-net · Execution + Pricing updated
Management language on supply normalization and gross-to-net dynamics extracted from earnings transcript.
→ Add to weekly PM brief · Link to pricing assumptions
18 HRS AGO
RED_TEAM_OBJECTION · CREATED
Auto-generated · Targets: Bull H1 · Trigger: demand/access evidence gap
System-generated objection: "Demand ≠ reimbursed demand." Employer coverage and PA barrier evidence insufficient to validate bull case access assumptions.
→ Request employer coverage evidence pack · Review bull case access assumptions
20 HRS AGO
CLINICAL_SIGNAL · TRIAL
Source: T1 · Entities: oral semaglutide, NVO · H6, H7 updated · Materiality: Medium
Oral semaglutide trial data update on ClinicalTrials.gov. Enrollment status changed. Monitor for readout implications on oral/injectable mix.
→ Update oral GLP-1 thread · Track mix dilution signals
24 HRS AGO
SOURCE_AUTHORITY_UPDATE
New T1 source added: FDA enforcement action page
New official regulatory source approved. Improves compounding/channel evidence coverage.
→ Source approved · Added to monitoring
36 HRS AGO
EVIDENCE_PACK_READY · WEEKLY
18 deltas · 5 open analyst tasks · 3 unresolved contradictions
Weekly GLP-1 PM brief generated. Summarizes thesis confidence, pillar updates, Red Team objections, and watchlist.
→ Export PDF · Send to PM

Living Knowledge: Thesis Graph, Not Document Storage

The living memory of the GLP-1 intelligence universe. Every entity, relationship, and evidence snippet is linked to a hypothesis, a debate pillar, and an analyst action. The graph is the PM's research operating layer.

Filter by type: Company Drug/Product Clinical Trial Payer/Access Regulatory Hypothesis Show All
COMPANY
DRUG/PRODUCT
CLINICAL TRIAL
PAYER/ACCESS
REGULATORY
HYPOTHESIS
Click a node to inspect · Double-click to expand
Entity Detail
Click a node to inspect entity details, relationships, and evidence.
Example Path
FDA 503B Proposal
→ affects compounded semaglutide/tirzepatide
→ updates Compounding Risk signal
→ supports Bull Hypothesis H4
→ creates Analyst Task: monitor channel
→ updates Situation Room Thread #1
Second Path
Survodutide Phase III
→ updates Competitor Pipeline signal
→ supports Bear Hypothesis H5
→ creates Competitor Matrix task
→ updates Weekly PM Brief
126
Entities
312
Relationships
87
Evidence Snippets
18
Hypothesis Links
11
Contradictions
9
Analyst Tasks

Preventing Confirmation Bias

Vextrum enforces structured dialectic: the Blue Team builds the investment thesis, the Red Team challenges it. Every conclusion requires falsification rules.

BLUE TEAM — Investment Thesis
GLP-1 leaders remain structurally advantaged because demand, indication expansion, brand trust, manufacturing scale, and clinical breadth overwhelm pricing and access friction.
Supporting Evidence
Strong underlying demand proxies (prescription volume, patient starts)
Cardiometabolic benefit expansion (MASH, CKD, cardiovascular risk)
Supply normalization and manufacturing capacity investment
Regulatory pressure on compounders supporting branded channels
Incumbent scale and physician prescribing familiarity
What Would Strengthen It
Broader coverage with manageable rebates, persistence improvement data, competitor differentiation failing to translate commercially.
RED TEAM — Counter-Thesis
The market is underestimating the speed at which payers, employers, price negotiation, and competition can compress net revenue and terminal value.
Challenging Evidence
Coverage restrictions and PA tightening at major PBMs/payers
Medicare pricing anchors compressing net realized economics
Employer affordability pressure limiting reimbursed demand
Competitor Phase III readouts with credible differentiation
Oral/next-gen cannibalization of premium injectable economics
What Would Strengthen It
Widespread PA tightening, lower-than-expected persistence, competitor data showing clear differentiation, management commentary on net price pressure.
Falsification Rules
If access expansion requires material price concession → do not treat as pure bull signal
If competitor efficacy is strong but commercial timing is distant → reduce immediate materiality
If regulatory action limits compounding but affordability remains unresolved → split channel risk into "regulatory" and "demand conversion" sub-signals
If source is Tier 4 weak signal → cannot change high-confidence thesis without corroboration

Evidence Pack: From Signal to PM-Ready Research Artifact

When a material signal fires, Vextrum assembles a structured evidence pack: affected entities, pillar mapping, confidence scoring, source-linked claims, bull/bear interpretation, and analyst next steps.

EVIDENCE PACK EXPORTABLE · PDF / API
Signal
FDA compounding action affecting GLP-1 channel risk
Classification: Regulatory event / compounding-channel risk / thesis impact requires validation
Affected Entities
Novo Nordisk Eli Lilly FDA semaglutide tirzepatide liraglutide 503B facilities telehealth/cash-pay
Affected Products
Wegovy Ozempic Rybelsus Zepbound Mounjaro compounded semaglutide compounded tirzepatide
Confidence
Source event: High
Entity mapping: High
Investment materiality: Medium
Revenue conversion: Low/unknown
Pillars: Execution & Supply, Pricing & Margin, Access & Reimbursement
Evidence Sources
Source Date Extracted Claim Entities Hypothesis Tier Conf. Limitation
FDA.gov 503B Notice[citation slot]Semaglutide/tirzepatide/liraglutide proposed for 503B exclusionNVO, LLY, 503BSupports H4T1HighDoes not address demand conversion
STAT News analysis[citation slot]Compounding demand sizing, patient count estimates, telehealth adoptiontelehealth, compoundersContext for H4, H2T3MedEstimates vary widely; no claims data
Novo IR statement[citation slot]Reaffirmed patient access programs and supply expansion commitmentsNVO, Wegovy, OzempicSupports H1, H4T1HighCompany self-reporting; verify with claims
Telehealth commentary[citation slot]Channel adaptation strategies; potential 503A workarounds discussedtelehealth platformsBear risk for H4T4LowWeak signal; single source, unverified
PBM / payer policy[watch]PA requirements, step therapy barriers, employer coverage decisions pendingPBMs, employersBear for H2, H3T2GapNo data yet; explains why conversion is unresolved
[citation slot] = URL/DOI populated at source ingestion. Not fabricated for illustration.
PER-SOURCE AUDIT FIELDS
FDA:source_url: [source URL slot]retrieved_at: [retrieval timestamp slot] doc_hash: [SHA-256 hash slot]parser_ver: [parser version slot] · human_review: [status] CMS:source_url: [CMS source URL slot]retrieved_at: [retrieval timestamp slot] doc_hash: [SHA-256 hash slot]parser_ver: [parser version slot] · human_review: [status]
BULL INTERPRETATION
Protects branded channels and reduces unauthorized substitute risk. Incumbents retain manufacturing, brand, and physician familiarity advantages.
BEAR INTERPRETATION
Affordability-driven demand may not fully convert to branded products. 503A workaround potential and patient churn behavior are unknown.
Analyst Next Steps
Monitor Comment Period Track Channel Substitution Check Rx/Claims Proxy Add to Weekly PM Brief Update Compounding Thread

Illustrative evidence pack. Public-source examples used for demonstration.

Open Question
Does regulatory tightening convert demand back to branded channels, or does affordability-driven demand churn, migrate to 503A alternatives, or leave the market entirely?
This is the question the evidence pack cannot yet answer. It requires: channel migration data, branded Rx/claims proxy, 503A enforcement visibility, and patient persistence tracking.

From Public Source to PM Action

The value of Vextrum is not that it detects a headline. The value is that it converts a source event into an auditable investment object: source → claim → entity mapping → hypothesis impact → Red Team objection → analyst task → PM brief.

1
SOURCE EVENT
Source type: FDA / official regulatory
Source tier: T1
Event: FDA action touching GLP-1 compounding for semaglutide, tirzepatide, liraglutide
Status: Official source detected, trade press corroboration available
SOURCE & AUDIT METADATA
source_url: [source URL slot]
retrieved_at: [retrieval timestamp slot]
document_hash: [SHA-256 hash slot]
parser_version: [parser version slot]
extraction_method: official-source parser + human-review queue
audit_event_id: [audit event ID slot]
2
EXTRACTED CLAIM
Regulatory treatment of certain compounded GLP-1 products changed / was proposed to change. FDA action targets 503B compounding of semaglutide, tirzepatide, and liraglutide.
Paraphrased claim from public regulatory source. Not verbatim quote.
3
ENTITY MAPPING
semaglutidetirzepatideliraglutide Wegovy · NVOOzempic · NVORybelsus · NVO Zepbound · LLYMounjaro · LLY 503B facilitiescompounding pharmaciestelehealth/cash-pay
4
EVENT CLASSIFICATION
Event type:
REGULATORY_DELTA
Sub-event:
COMPOUNDING_CHANNEL_RISK
Affected pillars:
Execution, Pricing, Access
5
THESIS DECOMPOSITION
Bull-side: Regulatory pressure may reduce unauthorized/compounded substitute overhang and support branded channel protection.
Bear/unknown: Branded demand conversion is not proven. Affordability constraints, payer restrictions, 503A workaround risk, and patient churn remain unresolved.
This decomposition is what makes Vextrum different from a news feed or generic RAG system.
6
CONFIDENCE MOVEMENT
H4 “Compounding overhang fades”: 54 → 62
Revenue conversion confidence: unchanged / low-medium
Source confidence: high
Investment interpretation confidence: medium
7
RED TEAM OBJECTION
“Do not upgrade branded revenue assumptions until there is evidence that cash-pay/compounding users migrate to branded prescriptions rather than churn, seek 503A alternatives, or remain outside covered channels.”
8
ANALYST TASKS
Monitor FDA comment period / follow-up regulatory actions
Monitor 503A workaround language
Track telehealth/compounder channel messaging
Compare branded Rx/claims proxy if available
Check payer PA / employer affordability constraints
9
PM BRIEF UPDATE
“Regulatory overhang improved, but revenue conversion remains unproven. Treat as bull evidence for channel protection, not yet bull evidence for volume/revenue assumptions.”
Decision use: No automatic buy/sell signal. PM uses evidence to update conviction, sizing, hedge, or catalyst preparation.

How Confidence Moves

Confidence is not a stock call. It is evidence confidence and thesis-confidence movement. The numbers have methodology, not vanity.

Confidence Scoring Inputs
Source authorityT1 official > T2 payer/research > T3 trade press > T4 weak signal
Entity resolutionproduct/company/mechanism mapping quality
Event clarityexplicit update vs ambiguous language
Cross-source corroborationindependent confirmation from 2+ sources
Materiality relevancetouches key PM-defined assumption?
Contradiction penaltyevent supports both bull & bear branches?
Recency / cadencenew information vs stale background
Human review statusapproved / needs analyst review / rejected
Worked Example: FDA Compounding Event
Source authorityHIGH (T1 FDA)
Entity resolutionHIGH
Event clarityHIGH
Cross-source corroborationMEDIUM
Materiality relevanceMEDIUM-HIGH
Contradiction penaltyAPPLIED
RESULT
H4 compounding-overhang confidence rises (54→62). Revenue-conversion confidence does not rise - contradiction penalty applied because the event supports branded protection but does not prove demand conversion.
PM Action Trail: Event → Task → Brief
EVENT FDA compounding action detected. Source tier: T1. SYSTEM REGULATORY_DELTA created. Affected entities mapped. H4 confidence updated. Contradiction object created: branded-channel protection vs conversion uncertainty. TASK Validate whether demand migrates to branded prescriptions. Check channel migration, payer access, employer affordability, Rx/claims proxy. BRIEF “Treat as bull evidence for channel protection, not yet bull evidence for branded revenue conversion.” USE No automatic buy/sell signal. PM uses evidence to update conviction, sizing, hedge, or catalyst preparation.

Command Layer Assistant

Hermes is a premium intelligence agent that operates across the entire platform — querying the knowledge graph, searching evidence, proposing configuration changes, and dispatching follow-up intelligence missions.

What evidence supports the bear case on GLP-1 access friction? Which sources are strongest?
The access-friction bear case is supported by 11 sources across 3 tiers:

Tier 1: CMS Medicare GLP-1 Bridge page, PBM formulary PA documents (2)
Tier 2: ICER value assessment, KFF affordability analysis, employer benefit survey
Tier 3: STAT News, AJMC, Endpoints coverage (3), trade press employer commentary

Strongest: CMS page (authority 95%), ICER report (authority 91%). Employer pullback sources are weakest (2 sources, both Tier 3).

Based on: KG query (11 facts) + semantic search · Max confidence: 0.91
Run a follow-up source discovery focused on employer coverage blind spots.
I've staged a follow-up source discovery targeting employer GLP-1 coverage gaps. The mission will focus on:

• Large employer benefit plan filings and updates
• Benefits consulting firm publications (Mercer, WTW, AON)
• Employer health coalition reports on obesity coverage

PROPOSAL STAGED · REQUIRES APPROVAL
Agent Capabilities
READ ONLY Search sources, entities, and evidence across the knowledge graph
READ ONLY Query document intelligence and extraction results
READ ONLY Semantic search across facts, insights, and strategic threads
PROPOSE Propose configuration changes with dry-run simulation
APPROVAL Dispatch follow-up source discovery missions (requires user approval)
Evidence-First Policy: Every Hermes response carries a structured evidence section citing source IDs, confidence scores, and tool invocations. Responses without tool-backed evidence are flagged as general knowledge.

Why Vextrum Is Different

Vextrum is not a news feed, an expert network memo, or a generic RAG tool. It is a research operating layer built around a specific investment controversy.

Not a News Feed

News tells you what happened. Vextrum tells you which hypothesis changed, which source supports it, and what action is required.

Example: FDA compounding action → updates H4, creates channel migration task, flags demand conversion as unknown.

Same Event, Two Thesis Impacts

FDA compounding action lowers one regulatory overhang but does not prove branded revenue conversion. Vextrum splits the event into separate thesis effects.

This is what makes it smarter than news/RAG.

Not Just Alt-Data

Alt-data provides a signal. Vextrum organizes the full investment controversy: thesis, sources, entities, evidence, confidence, Red Team, and PM outputs.

GLP-1 debate requires 41 source families, not a single data feed.

Not Generic RAG

Generic RAG retrieves documents. Vextrum maps evidence into an ontology and links it to hypotheses, catalysts, and analyst tasks.

CMS source update → mapped to Ozempic/Wegovy entities → updates H2 and H3 → creates model sensitivity task.

Evidence-First & Auditable

Every claim stores 12 required fields: source, tier, date, entity, hypothesis, confidence, materiality, action, limitation. No unsupported conclusions.

Built for PM Autonomy

Each PM/team defines its own debate, universe, materiality rules, output cadence, and evidence standards. No one-size-fits-all.

Red Team Prevents Bias

Falsification rules prevent confirmation bias. "Demand ≠ reimbursed demand." Weak signals cannot move high-confidence conclusions without corroboration.

Human-in-the-Loop

High-impact changes require approval. Low-confidence T4 sources cannot auto-upgrade investment thesis. Four-tier action gate.

Reusable Platform, Specific Universe

The platform compounds across deployments, but each intelligence room is built around a specific decision problem and PM-defined debate.

How Vextrum Produces Evidence-First Intelligence

A 12-stage intelligence pipeline: from PM onboarding through source discovery, entity resolution, ontology mapping, hypothesis scoring, Red Team falsification, and PM-ready delivery.

ONBOARD
AI-guided config
SCOUT
Source discovery
APPROVE
Human gate
INGEST
300+ source types
TRIAGE
Strategic Judge
EXTRACT
Entity resolution
ONTOLOGY
Knowledge model
GRAPH
Living knowledge
SYNTHESIZE
Bayesian Nexus
RED TEAM
Falsification
THREADS
Decision surface
DELIVER
Briefs + alerts
INTELLIGENCE ENGINES
Strategic Triage · Deep Extraction
Bayesian Synthesis · Catalog-First Ontology
Embedding-First Entity Resolution
LLM STACK
Frontier LLM fallback chain
Model-agnostic orchestration
Circuit breaker + deterministic validators
INFRASTRUCTURE
AWS Lambda + ECS Fargate
PostgreSQL (4 schemas)
S3 + CloudFront delivery
Each Stage, GLP-1 Context
ONBOARDCaptures GLP-1 debate: LLY/NVO long thesis, payer/competitor risks, 5 pillars, materiality rules SCOUTDiscovers FDA/CMS/PBM/trial/company/conference sources for GLP-1 coverage EXTRACTResolves semaglutide/Wegovy/Ozempic/Rybelsus/NVO; tirzepatide/Zepbound/Mounjaro/LLY SIGNALTags regulatory, payer, clinical, pricing, competitor, execution events HYPOTHESISUpdates bull/bear confidence: volume expansion vs access friction vs pricing compression RED TEAMChallenges interpretation: "demand ≠ reimbursed demand," "sentiment ≠ revenue" DELIVERPM alert cards, weekly brief, evidence packs, Situation Room threads, Delta feed
Why Generic RAG Fails Here
Retrieves documents but does not know thesis structure - cannot separate event fact from investment interpretation
Cannot map the same event to opposing hypotheses - FDA compounding action is both bull (branded protection) and bear-unresolved (demand conversion)
Lacks materiality rules - treats all documents equally regardless of source authority or thesis relevance
Lacks confidence discipline - weak KOL signals treated the same as T1 FDA/CMS actions
No analyst task workflow - retrieves information but cannot create, route, or close tasks

Pilot the GLP-1 Intelligence Room Around Your Healthcare Research Question

A 2-4 week pilot around one defined healthcare investment controversy. Bring one live research debate. Vextrum will turn it into a source-linked, evidence-weighted intelligence room.

Pilot Deliverables
PM onboarding session and thesis capture
Generated thesis map and intelligence configuration
Approved source registry with authority tiers
Securities/product/entity watchlist and ontology graph
3 Situation Room threads with bull/bear analysis
Daily Delta feed and 1 weekly PM brief
2 evidence packs with source-linked claims
Source coverage and blind spot report
Success Criteria
Does it surface useful deltas faster than manual monitoring?
Are sources sufficiently authoritative?
Are alerts material and low-noise?
Does the evidence pack reduce analyst time?
Does the Red Team layer improve debate quality?
Can the PM customize universe and materiality rules?
Does the system separate event fact from investment interpretation?
Can it produce a living intelligence room around one real healthcare debate?
GLP-1 is the demo controversy. The same workflow can be configured around any healthcare PM debate: oncology readouts, managed care reimbursement, medtech utilization, IRA pricing, hospital capex, China biotech, or a custom watchlist defined by the investment team.
Pilot input: one live healthcare research debate chosen by the PM team.
Bring one live healthcare research debate →

Vextrum will convert it into a source-linked, evidence-weighted intelligence room within 2–4 weeks.

Metrics are illustrative case-study outputs from the configured GLP-1 intelligence room, not platform-wide totals.
Hermes
Hermes
Hermes
Intelligence Command Agent
×
I'm Hermes, your intelligence command assistant. I can search evidence, query the knowledge graph, explain confidence scores, and propose configuration changes. How can I help?
Explore pilot options ← Back to Demo Room