Market MapSales Tech
Vertical Market Map · Sales Tech
Sales Tech AI, mapped across 10 layers & 50 sublayers.
AI-native SDR, prospecting, dialers, meeting intel, deal execution, RevOps AI.
ThesisSales Tech is where the L5 → L1 bifurcation is happening fastest. At L5a, the AI SDR category (11x, Artisan, Rox, Unify) is dangerously crowded and structurally exposed — each foundation-model release closes the moat further, and Salesloft/Outreach/Apollo are absorbing agent features into the incumbent engagement platforms. The durable positions sit in L1 data: Clay owns L1b/L1c composition, Gong owns L1d call-outcome data (a decade of anonymized B2B calls no rival can replicate), and Common Room / UserGems / Pocus own the L1c signal graph. The fundable open sublayers are L1d beyond meeting transcripts, L8d institutional selling memory per-account, and L3a compliance for regulated-vertical outbound (TCPA / caller ID / GDPR) — the one place where regulation makes the outbound sublayer un-commoditizable.
Filter by stage:
Solid = primary layer · faded = secondaryOpen whitespaceFeature (absorbed)Horizontal (owned outside vertical)
· A ·
· B ·
· C ·
· D ·
· E ·
L8
Memory
L8A
Session & Short-Term Memory
Horizontal frameworks (LangChain, LlamaIndex)
L8B
User & Entity Profiles
Absorbed into CRM/RevOps platforms
L8C
Aggregated Network Learning
⌁ Aggregated selling-network learning, largely unclaimed
L8D⌁ OPEN
Institutional Knowledge
L8E
Learned World Models
⌁ Learned deal-outcome world model, wide open
L7
Surface
L7A
Conversational
Consumer / buyer-side surfaces, not seller-side
L7B
Visual Interfaces & Media
Horizontal media tools
L7C
Embedded & Embodied AI
L7D
Transaction Surface
L7E
Async & Ambient Surfaces
⌁ Ambient / wearable selling surface, unbuilt
L6
Orchestration
L6A
Agent Loops
L6B
Human-in-the-Loop
Built into L5a agent platforms
L6C
Role Routing & Task Decomposition
L6D
Context & State Management
Built into L5a platforms
L6E
Runtime Assurance & Learning Loops
Built into L5a platforms
L5
Execution
L5A
Domain Execution & Tool Use
L5B
Decision Frameworks & Reasoning Scaffolds
L5C
Retrieval-Augmented Workflows
L5D
Operating Playbooks
⌁ Firm-specific selling playbooks, under-built as a product
L5E
Interaction Skills & Actuation
Feature inside L5a agent platforms
L4
Access
L4A
API & Integration Layer
Horizontal integration infra (Salesforce/HubSpot APIs)
L4B
Agent Interface Protocols
Cross-industry protocols (emerging: MCP)
L4C
Access Governance & Agent Commerce
Platform feature
L4D
Real-Time Interaction Infrastructure
Horizontal infra
L4E
Agent Identity & Provenance
⌁ AI-SDR identity / provenance (whose agent is emailing me?), open
L3
Gates
L3A⌁ OPEN
Compliance & Export Controls
L3B
Quality Gates
⌁ Outbound quality/safety gates (deliverability, hallucination in email), emerging
L3C
Safety, Security & Provenance
Platform feature
L3D
Editorial Gates
Platform feature
L3E
Distribution Gates
Platform feature
L2
Models
L2A
Foundation & Multimodal Models
Foundation labs (OpenAI / Anthropic / Google)
L2B
Specialized & Fine-Tuned Models
⌁ Sales-tuned models, essentially unbuilt as a standalone
L2C
Embedding & Retrieval
Horizontal infra
L2D
Model Routing & Composition
Horizontal / infra
L2E
Reasoning & World Models
Foundation labs
L1
Data
L1A
Public & Open Data
Public web data — feeds Clay et al.
L1B
Proprietary Data
L1C
Behavioral & Sensor Data
L1D⌁ OPEN
Outcome Data
L1E
Synthetic & Simulation Data
⌁ Synthetic sales-conversation data, emerging
L0
Infra
Vertical-agnostic, shared GPUs / silicon / data centers / cloud. Not Sales-Tech-specific.
L-1
Resources
Vertical-agnostic, energy, materials, fabrication. Not Sales-Tech-specific.
Layer scorecard, contested vs. open
| Sublayer | AI-first occupants | State |
|---|---|---|
| L1b Proprietary Data | Clay, Apollo, Ocean.io, Amplemarket | Most crowded |
| L1c Behavioral & Signal Data | Common Room, UserGems, Warmly, Pocus, People.ai | Squeezed middle |
| L1d Outcome Data | Gong, Rilla, Chorus (ZoomInfo), Attention, Aviso, Nooks | Scarce → bought |
| L3a Compliance & Export Controls | Regal (dominant) | Open & fundable |
| L5a Domain Execution | 11x, Artisan, Regie, Rox, Unify, Nooks, Orum, Regal, Amplemarket, Attention, Endgame, Momentum, Piper, Default, Relevance | Most crowded |
| L5b Reasoning Scaffolds | Clari, Aviso, (Gong secondary) | Squeezed middle |
| L6a Agent Loops | 11x, Artisan, Unify, Regie, Salesloft (Rhythm) | Most crowded |
| L6c Multi-Agent Coordination | Rox, Regie, Relevance | Open & fundable |
| L7c Embedded Copilot | Fathom, Fireflies, Attention, Attio, Day.ai, Twenty, Scratchpad | Most crowded |
| L8d Institutional Knowledge | Gong, Rox, Attio, Day.ai (all aspiring) | Open & fundable |
| L8e Learned World Models | None | Open & fundable |
Where the next winner comes from
1 · L1d Deal-outcome data beyond call transcripts.
Gong owns meeting audio; nobody owns the full deal graph (email + CRM + product usage + outcome). Whoever wires that corpus first wins the L8e learned-world-model race.
2 · L8d Institutional selling memory per-account.
Rox, Attio, Day.ai aspire; nobody yet stores 'what worked / what didn't with THIS buyer over 3 years' in a queryable, portable, rep-agnostic form.
3 · L3a Regulated outbound compliance.
Regal is nearly alone in taking TCPA / branded caller-ID / consent orchestration seriously. Insurance, mortgage, healthcare outbound cannot be commoditized by a horizontal L5a agent — the regulation is the moat.
4 · L2b Sales-tuned specialist models.
Empty. Every Sales Tech company sits on foundation models. First real domain-specialist model (deal-language, negotiation, objection-handling) has an opening — but capex bar is high.
† Chili Piper acquired by Thoma Bravo (Feb 2025). Chorus acquired by ZoomInfo ($575M, Jul 2021). Funding lines drawn from primary announcements + TechCrunch, The Information, Bloomberg, Reuters, SaaStr, PitchBook coverage (2023–May 2026). Point-in-time valuations. Every unverifiable figure carries a `flag:` field — treat those cautiously. This map is deliberately narrow (SDR/AE/CRM-adjacent). Marketing automation, ABM ad platforms, CS/PLG, pricing, and deal-desk suites live in a sibling GTM/marketing map (queued).