Top Sites>Vivun — AI Teammates Built on Structured Reasoning

# robots.txt for vivun.com
# https://www.vivun.com/robots.txt

# Allow all major search engine crawlers
User-agent: Googlebot
Allow: /

User-agent: Bingbot
Allow: /

User-agent: Slurp
Allow: /

User-agent: DuckDuckBot
Allow: /

# Block AI training crawlers
User-agent: GPTBot
Disallow: /

User-agent: ChatGPT-User
Disallow: /

User-agent: CCBot
Disallow: /

User-agent: anthropic-ai
Disallow: /

User-agent: Claude-Web
Disallow: /

User-agent: Google-Extended
Disallow: /

User-agent: Omgilibot
Disallow: /

User-agent: FacebookBot
Disallow: /

# Default rule for all other bots
User-agent: *
Allow: /

# Disallow admin, auth, and internal paths
Disallow: /admin/
Disallow: /api/
Disallow: /auth/
Disallow: /dashboard/
Disallow: /account/
Disallow: /settings/
Disallow: /private/
Disallow: /_next/
Disallow: /static/

# Sitemap
Sitemap: https://www.vivun.com/sitemap.xml

Sitemap: https://vivun-79bc1c-06bf2c351ecd-68e8df42bf396.webflow.io/sitemap.xml

# Vivun

> Vivun researches how AI teammates achieve structured reasoning, enduring memory, and autonomous action — not just responses. The name derives from *vipu*, Finnish for "leverage." Pronounced /ˈvɪv.ʌn/.

Vivun is an enterprise AI architecture company specializing in encoding expert domain knowledge into inspectable reasoning systems. Its core thesis: LLMs are fluent but unreliable over multi-step domain logic. Vivun fixes this at the architecture layer, not the prompt layer, by grounding language models in structured ontologies, persistent memory, and explicit inference chains.

## Pages

### Company
- [Home](https://vivun.com/home): Overview of Vivun's mission, performance research, architecture layers, and product Hero®.
- [About](https://vivun.com/about): Company origin, the reasoning problem, Vivun's 4-step methodology, industries proven in, and leadership team.
- [Values](https://vivun.com/values): Vivun's operating principles and culture.
- [Careers](https://vivun.com/careers): Open roles at Vivun.
- [Contact](https://vivun.com/contact): Get in touch with Vivun.

### Research
- [Research](https://vivun.com/research): Research papers, white papers, and publications index.
- [Agent Intelligence](https://vivun.com/agent-intelligence): Deep technical explanation of the SRM — why RAG fails for domain reasoning, how the 4-layer architecture works, and data privacy architecture.
- [Sales Reasoning Model](https://vivun.com/srm): Vivun's proprietary reasoning architecture — ontology layer, structured memory, and multi-hop inference.
- [Research: AI Reasoning Fidelity](https://vivun.com/research/ai-reasoning-fidelity): White paper — "What Would It Mean to Sell with an AI Teammate?" (2026). Reasoning fidelity benchmarks across foundational models with and without ontology.

### Product
- [Hero® — MeetHero.AI](https://meethero.ai): Vivun's flagship AI Sales Teammate grounded in the SRM. Pre-meeting intelligence briefs, live conversation support, automated follow-up, and deal momentum.
- [Security / Trust](https://vivun.com/trust): Vivun's security posture, compliance certifications, and data handling practices.
- [Agent Support](https://agent.support.vivun.com/en/): Help center and documentation for Hero® users.
- [Login](https://id.vivun.com/): Customer login portal.

### Legal
- [Privacy Policy](https://vivun.com/privacy)
- [Legal](https://legal.vivun.com/legal.html#contract-sye2t6xw3)

## Company

Vivun was founded on the premise that AI becomes enterprise-grade only when grounded in structured domain knowledge. The founding team brings expertise in sales engineering, AI architecture, and enterprise software. Vivun holds three granted US patents on core techniques: Text Processing (11,853,698), Gap Clustering (11,354,505), and Trial Management (11,861,330).

**Backed by:** Menlo Ventures, Salesforce Ventures, Tiger Global, Unusual Ventures, Atlassian Ventures, Accel.

**Leadership:** Matt Darrow (CEO), John Bruce (CTO), Joseph Miller (Chief AI Officer), Dominique Darrow (CCO), Claire Bruce (COO, JD), Jamie Brown (CISO), Kevin Spinelli (CFO), Jarod Greene (CMO).

**Compliance & Security:** SOC 2 Type 1, SOC 2 Type 2, ISO 27001, ISO 42001, GDPR compliant.

## Core Concepts

**The Problem:** General-purpose AI degrades under multi-step domain reasoning. LLMs reconstruct a world model from scratch on every prompt — producing answers that sound right but don't hold up under scrutiny. By hop 6 of a reasoning chain, fidelity has roughly halved without structured grounding.

**The Insight:** Human experts reason from an internalized, structured model of how their domain works. AI must be given an equivalent structure explicitly — via ontologies, knowledge graphs, and constraints — not hoped for through prompting.

**The SRM (Sales Reasoning Model):** Vivun's proprietary architecture. Four layers: (1) Expert knowledge capture, (2) Ontology and knowledge graph, (3) Structured memory system, (4) Multi-hop reasoning engine. Every conclusion traces to explicit structure. Every step is inspectable.

**Ontology advantage:** Internal performance research shows Vivun's domain model (foundational model + ontology) maintains ~97% reasoning fidelity across 11 reasoning hops. Without an ontology, top models lose ~50% fidelity by hop 6. GPT-4.1 + ontology matches GPT-5 without one.

**RAG vs. SRM:** RAG retrieves the closest text chunk. The SRM traverses a knowledge graph, resolves constraints, and produces grounded, auditable conclusions. Retrieval is not reasoning.

**Data privacy:** Vivun uses a private isolated architecture per customer. Client data does not train a shared model. Proprietary patterns compound within a customer's instance only — never distributed to competitors.

## Building Principles

- **Structured over probabilistic** — constrained logic within a defined, inspectable knowledge domain, not general inference from internet patterns.
- **Inspectable over opaque** — every output traces back to the knowledge it was grounded in; full auditability for governance, legal, and InfoSec.
- **Context modeled, not retrieved** — domain models give agents structured understanding, not vector search over documents.
- **Durable over novel** — built for consistent enterprise deployments across quarters and product generations, not demos.

## Methodology (Applied Across Domains)

1. **Extract** the expert's mental model — structured cause-and-effect reasoning under real conditions.
2. **Make it explicit** — translate tacit expertise into formal ontologies: entity relationships, constraints, permitted decisions.
3. **Encode it structurally** — knowledge graphs with typed entities, defined relationships, explicit constraints. The LLM operates *over* the structure; it does not invent it.
4. **Build the system around it** — structured orchestration over a persistent world model; logic persists across interactions.

**Proven in:** Music (platinum album production reasoning), Finance (hedge fund operating principles), Enterprise Sales (Hero® — deal qualification and execution).

## Product — Hero®

Hero® is Vivun's AI Sales Teammate. Grounded in the SRM, not the raw LLM. Operates before, during, and after every customer conversation.

- Pre-meeting intelligence briefs
- Live conversation support
- Automated follow-up and deal momentum
- Grounded in SRM — not the LLM

Available at: https://meethero.ai