Developer docs
API playgroundTry for free, no card

Search company profiles

Conjecture

Full company profile

uuid0000p5q

Namestring
Conjecture
Legal namestring
Conjecture, Ltd.
Websiteurl
conjecture.dev
Company typeenum
Private
Founded yearint
2022
Descriptiontext

Conjecture is a London-based AI safety company founded in March 2022 by Connor Leahy, Sid Black, and Gabriel Alfour, all co-founders of the open-source LLM community EleutherAI. The company operates Conjecture, Ltd. as its UK entity alongside Conjecture, Inc. in the US, and has grown from seven founders to approximately 20 employees by early 2024.

The company's core product is Cognitive Emulation (CoEm), an AI architecture that builds controllable, bounded systems by emulating human reasoning processes. CoEm combines LLMs used sparingly with software scaffolding, purpose-specific finetuning, and synthetic data to produce modular systems whose behavior can be inspected and edited by users. The approach is positioned as an alternative to frontier-scale model scaling, prioritizing reliability and interpretability over capability breadth, and supports text and multimodal workflows.

Conjecture generates revenue through enterprise subscriptions for CoEm-based solutions, targeting business buyers that require auditable and controllable AI deployments. Go-to-market is a direct enterprise field-sales motion anchored by a "Request a Demo" CTA on conjecture.dev. The company also pursues policy and thought-leadership positioning, including international conference presentations and advising politicians, while maintaining a research presence on LessWrong and the Alignment Forum. Capital was raised in 2022 from a syndicate of high-profile individual investors including Nat Friedman, Daniel Gross, the Collisons, Arthur Breitman, Andrej Karpathy, and Sam Bankman-Fried; specific round size and post-money valuation were not disclosed.

Short descriptiontext

Conjecture is a London-based AI safety company that builds Cognitive Emulation systems, combining LLMs with software scaffolding to deliver controllable, auditable enterprise AI, serving business customers requiring reliable and bounded AI deployments.

Operating statusenum
Operating
Ownership categoryenum
Headcount rangeband
1–10
akta.pro rankint
HeadquartersLondon, United Kingdom
HQ citystring
London
HQ countrystring
United Kingdom
HQ regionstring
Europe
Markets served

Serves global market

Offices1 record

Each record includes

City, Country, Type, Description, Source

Keyword5 values
AI safety software, AI alignment platforms, enterprise AI solutions, cognitive emulation systems, controllable AI architecture
Industry3 codes
PRIMARYBPAEAHAG
Generative AI & LLM Solutions Services (RAG, Agents, Copilots)
SECONDARYHDAAAMAE
Regulatory Readiness & Audit Automation (e.g., EU AI Act, NIST AI RMF, ISO/IEC 42001)
SECONDARYHDAAACAD
Prompt Engineering, Orchestration & LLMOps Tooling
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
NAICS code1 code
  • Computer Systems Design and Related Services5415
SIC code1 code
  • Services-Prepackaged Software7372
Product category
AI Safety and Alignment Software
Social media profiles2 records
GTM motion1 record

Each record includes

Type, Description, Source

Revenue model1 record
1Enterprise AI Solutions
TypeSubscription Recurring
Description

Conjecture generates revenue by providing enterprise customers with Cognitive Emulation-based AI systems that offer high reliability and controllability guarantees. They are already working with enterprise customers and seeking additional enterprise partners.

conjecture.dev
Marketing channels5 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels1 record

Each record includes

Title, Type, Scope, Target buyer, Description, Source

Cost components5 values
Personnel, Technology or R&D, Operations, Marketing or Sales, Infrastructure
GTM typeB2B
B2B
Offering typeSoftware
Software
Brand1 record
1Cognitive Emulation (CoEm)
Description

An AI architecture approach that builds systems to follow trusted human reasoning processes, designed to be powerful yet safe and controllable.

conjecture.dev
Core offering1 text field

Conjecture builds and sells Cognitive Emulation (CoEm), an AI architecture and set of AI systems that emulate human-like reasoning by decomposing cognitive tasks into primitive building blocks and combining fine-tuned LLMs with synthetic data and scaffolding software. The resulting systems are designed to be bounded, controllable, safe, and interpretable, enabling enterprise customers to automate real workflows reliably. Conjecture delivers these capabilities through direct enterprise engagements, supplementing product work with conceptual and applied AI alignment research.

Differentiator
Functional benefit
Problem solved
Quantifiable outcome1 of 3 values shown
  • Systems that do exactly what they are told, nothing more, nothing less
+2 more records
Product overview1 text field

Conjecture is an AI Safety company offering Cognitive Emulation (CoEm), an alternative AI architecture approach that enables powerful, safe, and controllable AI systems. The company builds state-of-the-art LLMs as components within the Cognitive Emulation framework, combining fine-tuning, synthetic data, and scaffolding methods to create modular systems. CoEm is designed to solve the AI alignment and control problem by emulating trusted human reasoning processes rather than scaling unbounded models.

Product and service1 record
1Cognitive Emulation (CoEm)
CategoryAI Safety and Alignment Software
Description

Conjecture's core AI architecture approach that builds powerful, controllable, and safe AI systems for enterprises by emulating human-like reasoning. CoEm breaks cognitive tasks into primitive building blocks combined with LLMs and software, producing modular systems that perform reliably without hallucinations and that surface their reasoning in human-readable form for editing and auditing.

Scale indicator3 records

Each record includes

Type, Value, Description, Source

Recent move6 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight4 records

Each record includes

Type, Description

Peers10 records
TypeDirect peer
Description

Arthur AI provides AI observability, monitoring, and governance for enterprise LLM deployments. Directly comparable as an enterprise-facing AI safety/reliability layer.

TypeBroad incumbent
Description

OpenAI is the dominant frontier AI lab and the primary indirect competitor Conjecture must differentiate against on safety, controllability, and enterprise reliability.

TypeBroad incumbent
Description

AI21 Labs builds enterprise LLMs with emphasis on reliability and controllability features. Comparable enterprise reliability narrative and target buyer, though AI21 also pursues scale.

TypeDirect peer
Description

Anthropic is the most direct competitor: an AI safety-focused lab building controllable, interpretable LLMs (Claude) for enterprise use. Both companies sell reliability and alignment as the primary differentiator against frontier-scale competitors.

TypeBroad incumbent
Description

Cohere is an enterprise-focused LLM provider serving regulated industries. Comparable go-to-market and buyer segment, though differentiated by general-purpose rather than safety-first positioning.

TypeEmerging player
Description

Adept builds AI agents that automate enterprise workflows. Comparable in workflow-automation thesis, though Adept pursues autonomous agent architecture rather than Conjecture's bounded CoEm.

TypeDirect peer
Description

Holistic AI provides an AI governance, risk, and compliance platform for enterprises deploying AI systems. Both Conjecture and Holistic AI target the same buyer pain point — auditable, controllable AI deployment under emerging regulation.

TypeBroad incumbent
Description

DeepMind is a frontier AI research organization with a substantial alignment/safety team. Conjecture's narrative of solving alignment competes directly with DeepMind's published safety agenda.

TypeEmerging player
Description

Mistral AI is a European-headquartered LLM startup emphasizing efficient, controllable models. Comparable in geography and philosophy around bounded capability, though focused on open-weight release rather than enterprise services.

TypeDirect peer
Description

Credo AI builds an AI governance and compliance platform aligned with EU AI Act and NIST AI RMF. Comparable to Conjecture's regulatory-readiness positioning for enterprise AI buyers.

Market position
Strengths4 records

Each record includes

Headline, Details, Source

Weaknesses4 records

Each record includes

Headline, Details, Source

Competitive moat3 records

Each record includes

Type, Details

Key risks5 records

Each record includes

Headline, Details, Source

Key highlights5 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Named customers1 record

Each record includes

Name, Industry, Type, Use case, Source, UUID

Segment1 record

Each record includes

Title, Type, Primary, Description, Pain point addressed, Use case, Source

Ideal customer profile1 record

Each record includes

Profile, Firmographic size, Sales motion, Sales cycle length, Buying structure, Purchase trigger, Buyer persona, Geography, Industry vertical, Primary use case, Description, Pain points, Evidence proof points, Target buyer

Technology focused
Yes
API detail
Has APIbool
No

Docs URL, Description

AI capability5 records

Each record includes

Type, Description, Source

AI maturity
App detail

Has app

Feature5 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles4 records

Each record includes

Name, Designation, Designation category, Overview, Profile commentary, Source

No data
No data
Funding overview

Funding stage, Last funding date, Total funding USD

Funding rounds4 records

Each record includes

Round, Amount USD, Date, Pre money valuation, Total investors, Investors, News

Investors3 records

Each record includes

Name, Type, Date of entry, Rounds participated, Website

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Each record includes

Name, Acquisition type, Announced date, Completed date, Status, Website, News

Investment

Each record includes

Name, Round, Announced date, Lead investor, Website, News

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Conjecture

AI Safety and Alignment Softwareconjecture.dev

Conjecture is a London-based AI safety company that builds Cognitive Emulation systems, combining LLMs with software scaffolding to deliver controllable, auditable enterprise AI, serving business customers requiring reliable and bounded AI deployments.

What Conjecture does

Conjecture is a London-based AI safety company founded in March 2022 by Connor Leahy, Sid Black, and Gabriel Alfour, all co-founders of the open-source LLM community EleutherAI. The company operates Conjecture, Ltd. as its UK entity alongside Conjecture, Inc. in the US, and has grown from seven founders to approximately 20 employees by early 2024.

The company's core product is Cognitive Emulation (CoEm), an AI architecture that builds controllable, bounded systems by emulating human reasoning processes. CoEm combines LLMs used sparingly with software scaffolding, purpose-specific finetuning, and synthetic data to produce modular systems whose behavior can be inspected and edited by users. The approach is positioned as an alternative to frontier-scale model scaling, prioritizing reliability and interpretability over capability breadth, and supports text and multimodal workflows.

Conjecture generates revenue through enterprise subscriptions for CoEm-based solutions, targeting business buyers that require auditable and controllable AI deployments. Go-to-market is a direct enterprise field-sales motion anchored by a "Request a Demo" CTA on conjecture.dev. The company also pursues policy and thought-leadership positioning, including international conference presentations and advising politicians, while maintaining a research presence on LessWrong and the Alignment Forum. Capital was raised in 2022 from a syndicate of high-profile individual investors including Nat Friedman, Daniel Gross, the Collisons, Arthur Breitman, Andrej Karpathy, and Sam Bankman-Fried; specific round size and post-money valuation were not disclosed.

Conjecture firmographics

Firmographics
Name
Conjecture
Legal name
Conjecture, Ltd.
Website
https://conjecture.dev
Company type
Private
Founded year
2022
Operating status
Operating
Headcount range
1–10 employees
Short description
Conjecture is a London-based AI safety company that builds Cognitive Emulation systems, combining LLMs with software scaffolding to deliver controllable, auditable enterprise AI, serving business customers requiring reliable and bounded AI deployments.
Ownership category
akta.pro rank

Conjecture industry classification

Industry
Product category
AI Safety and Alignment Software
NAICS
Computer Systems Design and Related Services (5415)
SIC
Services-Prepackaged Software (7372)
akta.pro industry
PRIMARYBPAEAHAG
Generative AI & LLM Solutions Services (RAG, Agents, Copilots)
SECONDARYHDAAAMAE
Regulatory Readiness & Audit Automation (e.g., EU AI Act, NIST AI RMF, ISO/IEC 42001)
SECONDARYHDAAACAD
Prompt Engineering, Orchestration & LLMOps Tooling
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)

Keywords

  • AI safety software
  • AI alignment platforms
  • Enterprise AI solutions
  • Cognitive emulation systems
  • Controllable AI architecture

Where Conjecture is headquartered

Location

Headquarters

HQ city
London
HQ country
United Kingdom
HQ region
Europe

Offices1 record

Markets served

Conjecture business model

Business model
GTM type
B2B
Offering type
Software
Cost components
Personnel, Technology or R&D, Operations, Marketing or Sales, Infrastructure

Revenue model

  1. Enterprise AI Solutions: Conjecture generates revenue by providing enterprise customers with Cognitive Emulation-based AI systems that offer high reliability and controllability guarantees. They are already working with enterprise customers and seeking additional enterprise partners.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels5 records

Conjecture product offering

Product offering

Core offering

Conjecture builds and sells Cognitive Emulation (CoEm), an AI architecture and set of AI systems that emulate human-like reasoning by decomposing cognitive tasks into primitive building blocks and combining fine-tuned LLMs with synthetic data and scaffolding software. The resulting systems are designed to be bounded, controllable, safe, and interpretable, enabling enterprise customers to automate real workflows reliably. Conjecture delivers these capabilities through direct enterprise engagements, supplementing product work with conceptual and applied AI alignment research.

Product overview

Conjecture is an AI Safety company offering Cognitive Emulation (CoEm), an alternative AI architecture approach that enables powerful, safe, and controllable AI systems. The company builds state-of-the-art LLMs as components within the Cognitive Emulation framework, combining fine-tuning, synthetic data, and scaffolding methods to create modular systems. CoEm is designed to solve the AI alignment and control problem by emulating trusted human reasoning processes rather than scaling unbounded models.

Differentiator

Problem solved

Functional benefit

Brands

  • Cognitive Emulation (CoEm): An AI architecture approach that builds systems to follow trusted human reasoning processes, designed to be powerful yet safe and controllable.

Products and services

  • Cognitive Emulation (CoEm) Conjecture's core AI architecture approach that builds powerful, controllable, and safe AI systems for enterprises by emulating human-like reasoning. CoEm breaks cognitive tasks into primitive building blocks combined with LLMs and software, producing modular systems that perform reliably without hallucinations and that surface their reasoning in human-readable form for editing and auditing.

Quantifiable outcome

  • Systems that do exactly what they are told, nothing more, nothing less
  • +2 more outcomes

Companies that use Conjecture

Customer profile

Named customers1 record

Segments1 record

Ideal customer profiles1 record

Conjecture technology and API

Technology

Technology focussed Yes

API detail

Has API
No
API docs
API detail

Core technology

AI maturity

App detail

AI capability5 records

Feature5 records

Conjecture partnerships and signals

Strategic signal

Scale indicators3 records

Recent moves6 records

Expansion highlights4 records

Conjecture competitors and assessment

Company assessment

Direct peers

  • Arthur AI: Arthur AI provides AI observability, monitoring, and governance for enterprise LLM deployments. Directly comparable as an enterprise-facing AI safety/reliability layer.
  • Anthropic: Anthropic is the most direct competitor: an AI safety-focused lab building controllable, interpretable LLMs (Claude) for enterprise use. Both companies sell reliability and alignment as the primary differentiator against frontier-scale competitors.
  • Holistic AI: Holistic AI provides an AI governance, risk, and compliance platform for enterprises deploying AI systems. Both Conjecture and Holistic AI target the same buyer pain point — auditable, controllable AI deployment under emerging regulation.
  • Credo AI: Credo AI builds an AI governance and compliance platform aligned with EU AI Act and NIST AI RMF. Comparable to Conjecture's regulatory-readiness positioning for enterprise AI buyers.

Broad incumbents

  • OpenAI: OpenAI is the dominant frontier AI lab and the primary indirect competitor Conjecture must differentiate against on safety, controllability, and enterprise reliability.
  • AI21 Labs: AI21 Labs builds enterprise LLMs with emphasis on reliability and controllability features. Comparable enterprise reliability narrative and target buyer, though AI21 also pursues scale.
  • Cohere: Cohere is an enterprise-focused LLM provider serving regulated industries. Comparable go-to-market and buyer segment, though differentiated by general-purpose rather than safety-first positioning.
  • Google DeepMind: DeepMind is a frontier AI research organization with a substantial alignment/safety team. Conjecture's narrative of solving alignment competes directly with DeepMind's published safety agenda.

Emerging players

  • Adept AI: Adept builds AI agents that automate enterprise workflows. Comparable in workflow-automation thesis, though Adept pursues autonomous agent architecture rather than Conjecture's bounded CoEm.
  • Mistral AI: Mistral AI is a European-headquartered LLM startup emphasizing efficient, controllable models. Comparable in geography and philosophy around bounded capability, though focused on open-weight release rather than enterprise services.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat3 records

Key risks5 records

Key highlights5 records

Customer concentration

Conjecture social profiles

Digital presence

Conjecture financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Conjecture leadership team

Management profile

Number of profiles

Profiles4 records

Conjecture funding detail

Funding detail

Funding overview

Funding rounds4 records

Investors3 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

Conjecture M&A and investment

M&A and investment

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Conjecture

What does Conjecture do?

Conjecture builds and sells Cognitive Emulation (CoEm), an AI architecture and set of AI systems that emulate human-like reasoning by decomposing cognitive tasks into primitive building blocks and combining fine-tuned LLMs with synthetic data and scaffolding software. The resulting systems are designed to be bounded, controllable, safe, and interpretable, enabling enterprise customers to automate real workflows reliably. Conjecture delivers these capabilities through direct enterprise engagements, supplementing product work with conceptual and applied AI alignment research.

Is Conjecture a public or private company?

Conjecture is a private company. It is classified as venture growth investor backed and is currently operating.

When was Conjecture founded?

Conjecture was founded in 2022. It employs 1 to 10 people.

Where is Conjecture based?

Conjecture is headquartered in London, United Kingdom, in the Europe region.

How does Conjecture make money?

One revenue line is on record: enterprise AI Solutions.

Who are Conjecture's main competitors?

Direct peers on record are Arthur AI, Anthropic, Holistic AI and Credo AI. Broad incumbents are OpenAI, AI21 Labs, Cohere and Google DeepMind. Emerging players are Adept AI and Mistral AI.

Does Conjecture have an API?

No public API is recorded for Conjecture.

What industry is Conjecture in?

Conjecture's product category is AI Safety and Alignment Software. Its primary akta.pro industry code is BPAEAHAG, Generative AI & LLM Solutions Services (RAG, Agents, Copilots), with a secondary code of HDAAAMAE, Regulatory Readiness & Audit Automation (e.g., EU AI Act, NIST AI RMF, ISO/IEC 42001). Its NAICS code is 5415 and its SIC code is 7372.

Unlock the full company data

50 free credits on sign-up, no credit card required.

Contact sales
Live signals
ConjectureConjectureThis content is an organizational introduction from Conjecture, an AI Safety company, explaining the concept of AI Alignment and the risks posed by increasingly powerful AI systems such as foundation models and potential Artificial General Intelligence (AGI). The article argues that current AI Safety techniques will not scale to handle superintelligent systems, and that capabilities are advancing faster than safety research can keep pace. Conjecture outlines its approach called Cognitive Emulation as an alternative to scaling larger models, advocating for bounded AI systems while the industry develops better alignment solutions.ConjecturaCybersecurity Knowledge Graphs with RAG: Discovering CSF v2.0 Compliance Use-Cases – ConjecturaConjectura describes a proof-of-concept combining Knowledge Graphs and Retrieval Augmented Generation to map cybersecurity standards and support NIST CSF 2.0 compliance, using publicly available resources from NIST and CIS. The POC covers CSF 2.0 and 1.1, with plans to extend to NIST 800-53, SANS Critical Security Controls and CIS Controls.The GuardianSunak’s global AI safety summit risks achieving very little, warns tech bossConnor Leahy, CEO of Conjecture, warned that Sunak's AI safety summit risks achieving little, accusing tech firms of trying to capture the meeting. He argued for a moratorium on developing artificial general intelligence, citing a 10-25% chance of catastrophic AI failure. The summit is scheduled for November at Bletchley Park.