Agentic AI in production
An agentic assistant for global employees, and a multi-cloud agent studio.
Anil Kumar Talla
AI Engineering & Product Leader
Head of AI | Ex-Autodesk Director of Engineering & Product | Enterprise AI Products & Platform, Corporate Engineering & Agentic AI | Hands-on builder | 20+ years
Enterprise AI automation at scale, agentic in production.
I build enterprise AI that automates real work. At Autodesk I owned the Digital Enterprise AI experience across product, engineering, Enterprise AI tools, HCM, knowledge strategy, and data strategy. An agentic AI assistant for 16K global employees reached an 80% automated self-solve rate, and search across 100M+ documents raised findability from 50% to 86%. I now lead AI at a stealth company.
- Product Leadership
- Enterprise Platforms
- Agentic AI
- AI Strategy
- Digital Experiences
- TypeScript
- Java
- Python
- Agentic RAG
- AI Gateway
- Multi-agent orchestration
- Agent memory
- MCP
- LLM routing
- Evals
- Guardrails
- Knowledge strategy
- Vector databases
- Neo4j
- Postgres
- MongoDB Atlas
- OpenSearch
- AWS serverless
- Next.js
- Org scaling
- Budget ownership
- GTM partnership
- Product strategy
- Governance
What I build
Enterprise AI that reaches production and automates the work. I own the product, the engineering, and the result.
Automation that holds
Automated self-solve for those employees. Enterprise customers deflect cases across L1 and L2.
Digital Enterprise AI
Product, engineering, Enterprise AI tools, HCM, knowledge strategy, and data strategy, across web, mobile, chat, and agentic AI.
Findability
Generative enterprise search.
The organization around it
A 60+ person organization and a $10M budget. Seven functions moved onto one platform, and a new AI use case went from six months to one sprint.
Hands-on to go-live
Hired the AI engineers and forward-deployed teams, took customers from contract to go-live in four weeks, and still build.
Outcomes
Measured in production, from the assistant used by global employees to the agent studio customers adopted.
Automated self-solveAgentic assistant for 16K global employees.
80%
FindabilityUp from 50% across 100M+ documents.
86%
AdoptionEnterprise customers on the agent studio.
98%
Case deflectionL1 and L2 within 30 days.
40%
Production accuracyTracked with evals, latency, and cost.
92%
Production agents3 million tasks a month.
30
Education
B.Tech, Computer Science
JNTU, Hyderabad2001 — 2005
Diploma in Advanced Computing
CDAC, Hyderabad2005 — 2006
Certifications
- Certified Product Leader, Product School
- Pragmatic Certified Product Manager
- Neo4j Certified Professional