Deep-tech AI since 2012

Making enterprises intelligent.

With AI Co-Workers® that comprehend, reason and act

E42.ai is  enterprise AI Co-Worker® platform built on more than a decade of R&D in machine comprehension, contextual intelligence, multimodal AI and cognitive automation.

Evolution of Agentic AI & Cognitive Automation

Before “agentic AI” had a name, the work had already begun

Light Information Systems was founded in Pune in 2012 around a persistent technical question: can machines move beyond keyword matching and programmed workflows to genuinely comprehend information, context and meaning?

2012

Poochh

Machine comprehension in the real world

An open-domain AI answering system over SMS. The machine attempted an answer; unresolved questions moved to human Buddies; validated answers returned to users and expanded the learning corpus. 

 

Machine answer→ Human escalation→ Validation→ Knowledge added 

 

Alongside natural-language research, early R&D also explored image and video understanding using CNNs and related neural-network architectures. 

 

Manthan / DEF ↗ 

2017-2019

NLPBots

Enterprise cognition

Years of machine-comprehension R&D became an enterprise platform for multifunctional cognitive agents, combining structured and unstructured data, contextual reasoning, APIs, workflows and machine learning across business functions.

 

NLPBOTS

 

Your A.I. Workforce

 

Dynamic intent | Auto-disambiguation | Contextual retrieval | Multilingual| Workflow orchestration | Limited-data learning

 

YourStory, 2018 ↗ | Oracle case study ↗

 

2021

E42.ai

From cognition to agency

The same foundation developed into the E42.ai platform: specialized models, multimodal intelligence, collaborating agents, orchestration and enterprise controls, designed to create AI Co-Workers® that take responsibility for business outcomes.

 

Specialized agents + LLMs / SLMs+Multimodal AI + Action models + Orchestration

 

Financial Express, 2021 ↗ | YourStory, 2023 ↗

Deep tech & intellectual property

 

Intent. Context. Evidence. Action. 

Two granted Indian patent families filed in May 2018 formalized technical problems that remain central to enterprise AI: interpreting intent dynamically and selecting the right context before reasoning over a query. 

 

Indian Patent 565295

 

A Bot Engine for Automatic Dynamic Intent Computation

Dynamic interpretation of intent using language, context, embeddings and neural-network techniques rather than a rigid predefined intent hierarchy.

 

Priority: 28 May 2018

 

Inventors: Sanjeev Thottapilly · Animesh Samuel

 

PCT WO2019229768A1 ↗ | Grant record ↗

 

Indian Patent 580011

 

Auto-Disambiguation & Dynamic Corpus Selection per Query

Query-specific corpus selection, document and passage ranking, contextual QA, attention/self-attention and evidence scoring before an answer is produced.

 

Priority: 28 May 2018

 

Inventors: Sanjeev Thottapilly · Animesh Samuel

 

PCT WO2019229769A1 ↗ | Citation trail ↗

 

What does the user mean? → What information is relevant? → What does it imply? → What happens next?

 

Trademark lineage

 

NLPBOTS — Your A.I. Workforce

 

Claimed use from 6 Aug 2014

 

Public record ↗

 

Registered trademark

 

AI Co-Worker®

 

Historical use dating to 2019

 

Public filing record ↗

Enterprise AI Deployments & Real-World Results

 

Cognitive agents, deployed in production 

E42 has been deployed across large global enterprises spanning India, North America, Europe and the Middle East, across finance, HR, operations, shared services and marketing intelligence.

01 Marketing intelligence

From scattered information to decision-ready intelligence

A marketing-intelligence deployment assembled strategic information on companies, executives and markets by identifying relevant sources, connecting context and synthesizing business intelligence.

 

View historical deployment ↗

02 HR & employee operations

From policy comprehension to enterprise employee AI

An enterprise deployment that began with understanding policy documents expanded into a broad employee-facing AI spanning payslips, leave, bookings, GRC self-help and dozens of business processes.

 

ET CIO case evidence ↗

03 Banking operations

Natural-language interaction connected to process execution

Oracle documented NLPBots combining free-flowing natural-language interaction with unstructured-text understanding and process automation inside a banking environment.

 

Oracle case study ↗

Independent & ecosystem recognition 

 

Tracked across deep tech, analyst research and major technology ecosystems. 

Recognition is presented as evidence of the journey — not as a substitute for the technology itself.

 

NASSCOM AI Game Changer

 

IDC Innovator

 

Gartner research ↗

 

Microsoft AIforALL

 

NVIDIA GTC

 

SAP ACE

 

Accenture Tech Next Challenge 2026

 

The platform today

 

Intelligence as a system, not a single model 

E42 combines the intelligence appropriate to each part of the process — specialized agents, LLMs and SLMs, multimodal understanding, deterministic controls, enterprise context, action models and orchestration.

 

Business outcome

 

AI Co-Worker®: Supervisory intelligence responsible for the end-to-end process

 

  • Understand: Documents · language · images · enterprise data
  • Reason: Context · policies · evidence · exceptions
  • Collaborate: Specialized agents · people · systems
  • Act: Workflows · APIs · ERP · communications

 

Specialized models LLMs / SLMs Multimodal AI Memory & context Rules Action models Orchestration

 

From the NLPBots years onward, learning effectively from limited, domain-specific enterprise datasets was a design requirement. The same principle now extends into specialized models and SLMs.

 

Mint — enterprise limits of ChatGPT-scale models ↗ Financial Express — SLMs ↗

 

Enterprise AI Security, Governance & Guardrails

 

Agency requires boundaries

Security, AI safety and responsible use are part of the E42 architecture itself. The objective is controlled agency: AI that can reason and act within explicit enterprise authority, policy and oversight.

 

Capability + Authority + Guardrails + Observability + Human oversight

 

Vortex

E42 enterprise AI gateway

Every model call, governed end to end.

Unified gateway

Hosted · self-hosted · on-prem · one endpoint

Adversarial defense

Prompt injection · jailbreaks · data exfiltration · toxic content

PII control

Detect · pseudonymize · mask

Credential isolation

Scoped virtual keys · budgets · rate limits · revocation

Identity & access

SSO · RBAC · least privilege · tenant isolation

Full observability

Cost · latency · success/failure · session reconstruction

Secure Governed Observable Model-agnostic

 

ISO/IEC 27001 SOC 2 Type 2GDPR CERT-In VAPT SSDLC AES-256 / TLS 1.3 ISO/IEC 42001 — in progress

 

One platform, many AI Co-Workers

 

From a reusable AI foundation to business-role ownership

Every E42 AI Co-Worker ® is created on the same underlying platform. The common foundation provides the agentic engine, models, orchestration, memory, integrations, security, governance and Vortex trust layer. Each Co-Worker adds the domain knowledge, process configuration, business rules and specialized agents required for a specific outcome.

 

E42 AI Co-Worker ® Platform

 

Agentic engine · models · orchestration · memory · Vortex · enterprise controls

 

Flagship productized Co-Worker

 

Neil

 

Supervisory AI for finance, beginning with Accounts Payable and expanding toward connected finance processes.

 

Explore Neil ↗

 

Custom Co-Worker

 

HR Operations

 

Employee self-service and high-volume HR processes.

 

Custom Co-Worker

 

Marketing Intelligence

 

Research, synthesis and decision-ready account intelligence.

 

Custom Co-Worker

 

Import & Compliance

 

Documentation orchestration and compliance verification.

 

Custom Co-Worker

 

Operations

 

Support, predictive maintenance and workflow-intensive work.

 

Custom Co-Worker

 

Industry Platforms

 

Specialized agentic systems for education and insurance.

 

 

Neil roadmap 

From Accounts Payable to Self-Operating Finance

AP establishes the enterprise foundation: data access, integrations, controls, business context and the operating model for AI-led finance. The roadmap expands that architecture across connected finance processes. 

 

AP Accounts Payable → AR Accounts Receivable → P2P Procure-to-Pay → SOF Self-Operating Finance 

 

Poochh→NLPBots→E42.ai→AI Co-Workers® 

 

Machine Comprehension → Enterprise Cognition → Agency → AI Co-Workers ®

Individual AI Co-Workers → collaborating AI Co-Workers ® → self-operating enterprise functions.

 

People & continuity 

A founder-led deep-tech journey

Core Leadership & Continuity 

The product names changed as AI technology evolved. The underlying objective such as machine comprehension translated into useful enterprise action, remained remarkably consistent. 

Animesh - Profile

Animesh Samuel

Co-Founder & CEO
 

Technology products, enterprise solutions, product strategy and go-to-market; co-inventor on Light’s 2018 patent families. 

Sanjeev - Profile

Sanjeev Menon

Co-Founder · Tech & Product 
 

25+ years across NLP, machine learning, big-data analytics, telecom/VoIP, augmented reality and predictive algorithms; also identified as Sanjeev Thottapilly in patents and legal records. 

E42 Chief Advisor - Srikanth Tanikella

Srikanth Tanikella

Director
 

Srikant brings with him strong exposure to international solution sales, operations management. Srikant is currently responsible for Technology-led Solutions and Services for the Indian Mutual Fund Sector. Srikant also has unique experience across the outsourcing lifecycle for Knowledge Services (KPO) as well as Transactional Enterprise Services (BPO).

As the Director at E42, Srikant brings to the table expertise in consultative selling, operations management, and talent management.

Wasim - Profile

Wasim Karani

Head of R&D
 

Wasim has been connected with E42 since inception with 12+ yrs of total work experience. You could reach him for help in A.I, NLP, ML related problems. He builds and executes research and development strategies / projects to deliver business results. He constantly enhances productivity, product quality, and cost effectiveness.

Swanand - Profile

Swanand Khaladkar

Senior Data Scientist
 

Swanand has 7+ years of experience in AINLP research. He has profound interest in research methodologies in NLP with latest technologies and models to enhance language processing capabilities of machine human interactions.

Core Leadership & Continuity 

R&D, product, and growth leaders, including Wasim Karani, Swanand Khaladkar, Rupesh Poojary, Vikrant Karande, Ramakrishna K, Furkan Siddique, Ajeeth B, and Srikanth Tanikella have been instrumental in carrying the technology and enterprise expansion across successive generations of the platform.

 

 

Company facts

Light Information Systems/E42.ai

Company: Light Information Systems Pvt. Ltd.

 

Founded: 2012 · Pune, India

 

Founders: Animesh Samuel · Sanjeev Menon (Sanjeev Thottapilly) · Sanjeev Prabhakaran Nair

 

Current platform / brand: E42.ai

 

Developed by Light: Poochh · NLPBots · E42.ai

 

U.S. entity: E42.ai Inc. · Delaware, USA

 

Enterprise AI gateway: Vortex

 

Flagship productized AI Co-Worker®: Neil — Finance / Accounts Payable

 

Intellectual property: Two granted Indian patent families · AI Co-Worker® registered trademark

Refund Policy

Refunds/Cancellations

We are offering free access to the platform for a month. Once the free period is over you need to pay the license fee to retain access to the platform. License fee will not be refunded.

 

If you have any questions about our Refund Policy, please email us at interact@e42.ai

Cognitive automation platform

At E42, creating a safe and healthy working environment takes precedence above all. The company has zero tolerance for prejudice, gender bias, and sexual harassment. For a comprehensive overview of our safety policy, please feel free to contact us at interact@e42.ai