Home / About
About Lyzr
Lyzr AI is a full-stack agent infrastructure platform that builds, orchestrates and governs production AI agents for enterprises and governments.
Founded in 2023 and headquartered in New York. The platform is sovereign: it runs in a customer's own cloud, on bare metal, fully air-gapped, or as SaaS. More than 3,000 customers, including Fortune 500 companies, and 50+ implementation partners including Accenture, Deloitte, KPMG, TCS and Firstsource.
01 / What Lyzr does
Each with the outcome a customer can measure.
01
Lyzr's reimagination consultants, solution architects, agent engineers and product managers redesign one enterprise workflow at a time and ship it as a production agentic workbench.
WTW's billing automation, live across five enterprise systems, cut end-to-end processing time by more than 60%.
02
A low-code platform for developers: 22 agent services in one place, including memory, knowledge bases, guardrails, tool and MCP integration, simulation, agent CI/CD, deployment and observability, with SuperFlow as the durable orchestrator.
Verifone took a 15-year-old customer support process to its first production agent in eight weeks.
03
Registers every agent in the enterprise, whether built on Lyzr, AWS Bedrock AgentCore, Google Vertex AI, Salesforce Agentforce, ServiceNow or Databricks, and traces every model call, tool call, token and data access through one gateway.
A US government agency runs it fully air-gapped as its central agent platform.
04
A vibe coding agent builder for business users: turns a plain-language requirement into a working agent, its orchestration and its interface, with central IT approval before anything reaches production.
Every Architect agent registers in Agent Studio: one registry, one security model.
05
Lyzr's open-source fine-tuning SDK, formerly ShadowLM: twelve post-training methods behind one config, moving a task from a rented frontier model to a small model the customer owns via shadow mode.
Designed to cut agent LLM cost by more than half without changing the agent.
06
Gives every employee LLM chat, enterprise search and agent building in one sovereign workspace that runs inside the customer's environment.
About 90% of inference runs on locally hosted open-weight models.
07
An on-premises appliance preloaded with open-weight models, fine-tuning tooling and the full Lyzr platform, so an organisation can run agents with zero data egress.
Built for regulated industries and government. First edition ships Q4 2026.
02 / Why Lyzr
01
Self-hosted in the customer's AWS, Azure or GCP account, on bare metal, fully air-gapped, or as SaaS, with every layer replaceable. Foundry, Gemini Enterprise and Bedrock AgentCore each run only inside their own cloud.
02
Opencontroller registers, traces and applies policy to agents from Agentforce, ServiceNow, Databricks, Bedrock AgentCore and Vertex AI, and to coding agents on developer laptops. A $1B+ software company chose Lyzr over ServiceNow on this point.
03
A LangChain or CrewAI build needs six or seven separate products before production. Lyzr ships 22 agent services in one platform with no external dependencies.
04
Lyzr ships the distillation loop inside the platform as Openfinetuner, with twelve post-training methods, and passes inference through at cost on the customer's own keys.
05
SuperFlow pauses a run for months on a human approval, resumes from the exact step and never repeats a side effect. Microsoft retires Foundry Workflows on 1 December 2026.
03 / Who uses Lyzr
40+
Enterprise customers
3,000+
Organisations building
50+
Implementation partners
1M+
Agents deployed
CIOs, CTOs and enterprise architects at Fortune 500 and 2000 companies who need one control plane for agents across clouds and vendors
Government agencies and regulated enterprises in banking, insurance, defense and telecom that require in-VPC, on-premises or air-gapped deployment
Function leaders in finance, marketing, sales, HR, procurement and customer operations automating one complex workflow end to end
Global systems integrators, including Accenture, Deloitte, KPMG, TCS and Firstsource, and emerging SIs delivering agentic transformation
Software companies that white-label Lyzr's core SDKs through Agent Blocks
Developers on Agent Studio, business users on Architect, teams on Open Cowork, and startups on the SaaS edition
Notable clients
04 / Team
Founder and CEO
Founded PowerUp Cloud, an AWS Premier Consulting Partner, and led its AWS business group to $400M in three years before its acquisition by the L&T Group in 2019. Previously a big data engineer at Tesco, Singapore Airlines, Allianz Life and Google.
LinkedIn ↗Cofounder and CTO
Known as Jimmy, leads platform and engineering, including Agent Studio, SuperFlow and the open-source projects. A Penn State graduate, he worked at Cisco and Oracle and ran his own startup.
LinkedIn ↗Cofounder and CGO
Leads growth and go-to-market. A Columbia graduate, he handled growth at Rocket Internet and built lifestyle startups of his own before cofounding Lyzr.
LinkedIn ↗Founded
2023
HQ
New York
Offices
Jersey City, London, Dubai, Bangalore
People
200+
Lyzr, short for Data Analyzer, began in early 2023 as an AI data analyst platform Siva was building on open-source tools, including LangChain. Managing and debugging it through an IDE alone proved very difficult, and his enterprise background told him developers needed something closer to the AWS or Snowflake console for AI. That insight turned Lyzr into an enterprise AI platform; the company was founded in August 2023 and launched the platform in January 2025.
Leadership
Shreyas Kapale
Founding Research Architect
Khush Patel
Founding Research Architect
Shekar Murthy
Chief Delivery Officer
Sid Asokan
COO - India
Husnaa Baig
Head HR
Vidur Rajpal
Founding Product Lead
Anju Chaudhary
VP - Consulting Partnerships
Aditya Pattabi
CFO
Ram Motipally
SVP, Technology Partnerships
Jeremiah Sadow
Chief Business Officer, Lyzr Banking
Advisors
Srinivas Narayanan
Advisor
Former CTO, OpenAI
Sunit Patel
Advisor
Former CFO, T-Mobile
Tom Lounibos
Advisor
Former MD, Accenture
Backed by Rocketship.vc, BGV, Accenture Ventures, GFT, the Partnership Fund for New York City and Plug and Play.
05 / How Lyzr works
Depending on who does the building.
Path 01
Developers sign up for Agent Studio and business users for Architect, build with 22 services, and deploy to SaaS or their own cloud. Usage is metered in credits; teams graduate to an enterprise licence for VPC or on-premises.
Path 02
More than 50 implementation partners scope, build and run Lyzr engagements on their own delivery model, with engineers certified through the Lyzr academy.
Path 03
Lyzr's forward-deployed teams redesign a workflow from the ground up and take it to production inside the customer's environment. Lyzr is accountable for the outcome.
Step 1
Knowledge workshop
60 to 90 minutes with business and IT leaders on where agents fit.
Step 2
Ideation hackathon
About 100 ideas, 20 shortlisted, 10 to 15 selected for production.
Step 3
Platform deployment
Self-hosted, Lyzr-managed in your cloud, dedicated tenant, or SaaS.
Step 4
Build and test
Four tracks build in shadow mode on historical data, then UAT.
Step 5
Production and scale
IT-approved push, continuous monitoring, then the next workflow.
Every enterprise customer works with a named Client Account Lead, a dedicated customer success manager and, in year one, forward-deployed engineers embedded with the customer's team. Delivery runs under Lyzr's Chief Delivery Officer, and leadership reviews progress through an AI transformation portal that tracks every workflow, agent and outcome.
Four tracks run in parallel on each workflow: a reimagination consultant designs the agent-native workflow in BPMN and builds the business case; a solution architect maps data, integrations, security and sovereignty; agent engineers build and run simulations and evals; and a product manager owns the experience so agents live inside Teams, Slack, SharePoint or a new workspace.
06 / Key facts
As of September 2026.
07 / FAQ
Lyzr AI is a full-stack agent infrastructure platform for enterprises and governments, founded in 2023 and headquartered in New York. It covers build, orchestrate, simulate, deploy, observe and govern in one product, and it runs in the customer's own cloud, on bare metal, fully air-gapped, or as SaaS.
Enterprise pricing starts at $250,000 per year. Opencontroller is $250,000 per year with unlimited usage, Agent Studio, Architect and SuperFlow are priced per agent-month, and Openfinetuner is $250,000 for 2.5 trillion training tokens. Customers bring their own model keys and Lyzr never marks up inference.
Lyzr runs in any cloud, on bare metal or fully air-gapped, while Foundry runs inside Azure. Lyzr does not replace Foundry: agents built there register in Opencontroller for governance, and SuperFlow runs the durable workflows that call them.
Siva Surendira (Founder and CEO), Jithin George (Cofounder and CTO) and Anirudh Narayan (Cofounder and CGO), in 2023. Siva previously founded PowerUp Cloud, acquired by the L&T Group in 2019.
Agentic workflow transformation by reimagination consultants, plus Agent Studio, Opencontroller, Architect, Openfinetuner, Open Cowork and the Optimus appliance. More than 50 implementation partners also deliver on Lyzr.
Yes. Self-hosted in AWS, Azure or GCP, on bare metal, and fully air-gapped with no outbound connection. A US government agency runs Lyzr air-gapped, and Optimus packages the whole stack as an appliance.