Why Your Agent Pilot Stalled — And What Changed for Teams That Shipped
A field view of where agent and copilot programs break down between prototype and production, and the patterns behind the ones that made it.
Speaker announcing soonDec 2–3, 2026 · New York City · In person
ODSC AI NYC— AI for Work Summit —
AI Has Changed. Workflows Haven’t.
Join AI builders, business practitioners, and leaders in Midtown NYC for a two-day, workshop-driven summit on moving AI from pilots to real workflow change.
Our training summits have helped launch thousands of AI careers across a decade of ODSC AI East and West. The people we trained on data science and ML in 2016 are the AI leads of 2026.
joined by practitioners and leaders from
What past ODSC AI attendees say about us
“A great learning experience diving into MSFT Copilot Foundry, MCP, and building agents with Claude Code. The sessions on scaling AI adoption in the enterprise provided practical, real-world perspectives that go far beyond the hype. I'm walking away with actionable ideas to put into practice immediately.”
“It's a thrill to learn from top AI and data science experts and immediately turn those insights into working prototypes.”
“ODSC AI East 2026 provided an incredible deep dive into cutting-edge AI—from the math behind LLMs and Agentic AI to production-grade MLOps with PyTorch and Kubernetes. The hands-on focus on GraphRAG, fine-tuning, and context engineering delivered immediate value and practical tools.”
“The conference centered on trust and performance – specifically how we verify AI and build scalable infrastructure. The emphasis on strong data foundations highlighted just how fast this space is evolving.”
“A great learning experience diving into MSFT Copilot Foundry, MCP, and building agents with Claude Code. The sessions on scaling AI adoption in the enterprise provided practical, real-world perspectives that go far beyond the hype. I'm walking away with actionable ideas to put into practice immediately.”
“It's a thrill to learn from top AI and data science experts and immediately turn those insights into working prototypes.”
“ODSC AI East 2026 provided an incredible deep dive into cutting-edge AI—from the math behind LLMs and Agentic AI to production-grade MLOps with PyTorch and Kubernetes. The hands-on focus on GraphRAG, fine-tuning, and context engineering delivered immediate value and practical tools.”
“The conference centered on trust and performance – specifically how we verify AI and build scalable infrastructure. The emphasis on strong data foundations highlighted just how fast this space is evolving.”
“The conference centered on trust and performance – specifically how we verify AI and build scalable infrastructure. The emphasis on strong data foundations highlighted just how fast this space is evolving.”
“ODSC AI East 2026 provided an incredible deep dive into cutting-edge AI—from the math behind LLMs and Agentic AI to production-grade MLOps with PyTorch and Kubernetes. The hands-on focus on GraphRAG, fine-tuning, and context engineering delivered immediate value and practical tools.”
“It's a thrill to learn from top AI and data science experts and immediately turn those insights into working prototypes.”
“A great learning experience diving into MSFT Copilot Foundry, MCP, and building agents with Claude Code. The sessions on scaling AI adoption in the enterprise provided practical, real-world perspectives that go far beyond the hype. I'm walking away with actionable ideas to put into practice immediately.”
“The conference centered on trust and performance – specifically how we verify AI and build scalable infrastructure. The emphasis on strong data foundations highlighted just how fast this space is evolving.”
“ODSC AI East 2026 provided an incredible deep dive into cutting-edge AI—from the math behind LLMs and Agentic AI to production-grade MLOps with PyTorch and Kubernetes. The hands-on focus on GraphRAG, fine-tuning, and context engineering delivered immediate value and practical tools.”
“It's a thrill to learn from top AI and data science experts and immediately turn those insights into working prototypes.”
“A great learning experience diving into MSFT Copilot Foundry, MCP, and building agents with Claude Code. The sessions on scaling AI adoption in the enterprise provided practical, real-world perspectives that go far beyond the hype. I'm walking away with actionable ideas to put into practice immediately.”
Why Attend ODSC AI NYC?
ODSC AI NYC is built for that next stage: two workshop-driven days for anyone responsible for making AI useful inside teams and organizations.
From pilots to production
Work through what it takes to move AI agents, copilots, and workflow automation beyond experiments and into real deployment.
From tools to workflows
Map how AI changes the way product, engineering, operations, analytics, and business teams actually get work done.
From adoption to trust
Build practical approaches to evaluation, governance, security, and oversight so AI systems are reliable enough to use.
From deployment to transformation
Identify what has to change across workflows, teams, governance, and measurement for AI adoption to produce real business value.
Program Themes
Six themes shape the two-day program — spanning the technical, operational, and leadership work of applied AI.
AI Agents & Copilots at Work
How agents, copilots, and assistants are being used across business, product, engineering, data, and operations workflows.
AI Engineering, Evals & Trust
Practical approaches to testing, evaluation, monitoring, reliability, human oversight, and production readiness.
Enterprise AI Deployment
Case studies and patterns for moving from isolated pilots to governed, scalable AI systems.
AI-Native Teams & Workflows
How teams are redesigning processes, roles, and collaboration around AI-first work.
Governance, Security & Risk
How organizations are managing AI risk, compliance, oversight, data access, and responsible deployment.
Hands-On Workshops & Demos
Practical sessions for builders and teams who want to see how AI systems are designed, tested, and deployed.
Who It's For — and What You'll Leave With
You build
AI engineers and practitioners making agents, copilots, and RAG systems production-grade.
- An agent workflow instrumented for observability — traced, logged, debuggable
- An eval suite pattern wired into CI, with golden sets your team will actually maintain
- A model-routing approach for cost, latency, and quality trade-offs
You apply
Sales, marketing, finance, ops, and HR practitioners building AI into their own function — no code required.
- One workflow from your own job automated in the room
- A delegation checklist: what to hand to AI, what to keep, how to QC without redoing the work
- Your function's workflow map — what to automate next and in what order
You lead
Executives, data and AI leaders, and founders funding, governing, and measuring AI adoption.
- A 90-day build/test/retire roadmap, built in the closing session
- A measurement baseline your CFO will accept
- A governance review process that lets the second team ship without starting over
Bring your team. Each of you trains in your own track — everyone leaves aligned on the same 90-day roadmap.
Full Apply-track and workshop schedule announcing soon.
Two Days Built for Focus and Depth
One hour of shared context. Hands-on by 10:30.
Day 1
9:00–10:00 · Main Stage Plenary
Everyone together- The state of AI at work: case studies from teams that moved past pilots
- The frameworks for workflow redesign, governance, and ROI used across both days
10:00–10:30 · Coffee & Networking
10:30 onward · Four Tracks
Workshops begin- Build — agents, evals, and AI engineering in production
- AI in Your Workflow — hands-on automation for sales, marketing, finance, ops, and HR
- Lead — governance and measurement working sessions
- AI Engineering Fundamentals — five sessions in sequence, foundations to a running agent
Evening
- Networking reception for all attendees
- VIP dinner with speakers and instructors
Day 2
All Day · Four Tracks
Deeper implementation- Function labs: build a working automation for your own role
- Agent, evals, and context-engineering workshops
- The Team AI Operating Agreement — drafted live in-session
- Fundamentals: your agent runs live, certificate awarded
- Closing session, everyone together: your 90-day build/test/retire roadmap
A Focused Summit, Not a Generic Expo
No trade-show floor, no vendor keynotes. Partner sessions are curated and clearly labeled, and every track is programmed for what attendees build and take home — not what sponsors want to present.
Choose Your Pass
Select the pass that fits how you want to experience ODSC AI NYC.
Preliminary Program
A working view of the two-day agenda. Filter by track or search by topic. Sessions are added and speakers named as they are confirmed.
Sessions marked announcing soon are confirmed in format and topic; speakers are added as they're booked.
27 sessions shown
From 40 Pilots to 4 Products: Consolidating an Enterprise AI Portfolio
How a large organization killed most of its pilots, picked the survivors, and built the platform underneath them.
Speaker announcing soonProving ROI When the Baseline Was Never Measured
Operators and finance partners on what counts as evidence of value, and how to instrument work before you automate it.
Speaker announcing soonMaking Brittle Agent Workflows Boring and Reliable
Retries, idempotency, tool contracts, state, and failure containment for multi-step agents running against real systems.
Speaker announcing soonBuilding an Eval Suite Your Team Will Actually Run
Offline and online evaluation, golden sets, LLM-as-judge pitfalls, and wiring evals into CI so regressions surface before users do.
Speaker announcing soonGovernance That Scales Instead of Blocking
Access control, audit trails, human-in-the-loop design, and the review process that lets a second team ship without starting over.
Speaker announcing soonInstrumenting an Agent for Observability
Bring a laptop. Trace a multi-tool agent end to end, add structured logging, and find where latency and errors actually come from.
Speaker announcing soonRetrieval That Holds Up Under Real Enterprise Data
Chunking, permissions-aware retrieval, freshness, and evaluating grounding quality on messy internal corpora.
Speaker announcing soonCopilots in Regulated Workflows: What Legal and Risk Asked For
The controls, disclosures, and fallbacks a regulated deployment needed before it was allowed near customers.
Speaker announcing soonCost, Latency, and Quality: Choosing Models Without Guessing
A practical routing and benchmarking approach for deciding which model does which job, and when to change your mind.
Speaker announcing soonWhat to Build, Test, or Retire in the Next 90 Days
The closing working session, every track together: build your team's 90-day roadmap and the measurement baseline that proves in January what changed. Certificates awarded at close.
Speaker announcing soonFundamentals 1: LLM & Agent Foundations
You use LLMs every day but haven't built with them. From a bare API call to structured outputs, function calling, and your first tool-calling loop — with a checkpoint every 30 minutes. You leave with a running repo: a CLI assistant that calls two real tools.
Speaker announcing soonFundamentals 2: Context Engineering & RAG
Your model doesn't know your data. Build retrieval into this morning's project: chunking, embeddings, retrieval quality, and context assembly — over documents you bring. You leave with a working RAG pipeline over your own documents.
Speaker announcing soonFundamentals 3: Building Your First Agent
Single prompts don't survive multi-step work. Add planning, memory, and failure handling — and learn when an agent is the wrong answer. You leave with a working agent that completes a three-step task with retries.
Speaker announcing soonFundamentals 4: Evals & Observability
It works in the demo — now prove it works anywhere else. Instrument yesterday's agent with golden sets, LLM-as-judge scoring, tracing, and regression gates. You leave with an eval suite wired to your agent, runnable in CI.
Speaker announcing soonFundamentals 5: Production — Cost, Routing & Guardrails
It works, and now it's expensive, slow, or risky. Model routing, caching, latency budgets, and guardrails on your own agent. You leave with a routing config and a one-page cost/latency budget.
Speaker announcing soonWorkflow Lab 1: Your First Delegation
You have AI tools and still do everything yourself. Bring a real task from your backlog — inbox triage, a draft, a research pull — and hand it off properly for the first time. You leave with one completed delegated task and a delegation checklist.
Speaker announcing soonWorkflow Lab 2: Files, Docs & Deliverables
Most of your job is turning source material into deliverables. Work against your own files: from raw docs to a finished report, deck, or spreadsheet — reviewed, not redone. You leave with a finished deliverable built from your own materials.
Speaker announcing soonOperator Case Study: The Workflow That Got Its Time Back
A practitioner walks through one real automation end to end: the workflow before, the workflow after, actual hours saved, and what broke on the way. Named operators, real numbers — announced as confirmed.
Speaker announcing soonWorkflow Lab 3: Make It Repeatable
One good result isn't a workflow. Chain the steps — research, draft, review, send — connect the tools you already use, and write it down so it runs the same way next Monday. You leave with one documented, repeatable workflow — running, not planned.
Speaker announcing soonFunction Lab: Sales & Marketing
Pipeline prep, account research, campaign drafts, call follow-ups. Map your funnel's manual steps, automate the two heaviest, and keep your voice in everything that ships. You leave with a working automation for one revenue workflow, plus your function's automation map.
Speaker announcing soonFunction Lab: Finance & Ops
Reporting packs, reconciliations, vendor reviews, process documentation. Build an automation that survives your controls — checkable, loggable, explainable to an auditor. You leave with a working automation for one reporting workflow, plus a control checklist.
Speaker announcing soonFunction Lab: HR & People
Job descriptions, screening summaries, policy answers, onboarding docs. Automate the paperwork while keeping judgment calls — and candidate experience — human. You leave with a working automation for one people workflow, plus a humans-in-the-loop map.
Speaker announcing soonTeam Lab: QC Without Redoing the Work
If you re-check everything, you saved nothing. Build a review practice that catches real errors in delegated work without doubling it back. You leave with a QC checklist calibrated to your own workflows.
Speaker announcing soonOperator Case Study: The Automation That Died — and the One That Stuck
A team lead on a first automation that quietly failed, and the second attempt that held. The failure modes that never make the launch deck. You leave with a pre-mortem checklist for your team's next rollout.
Speaker announcing soonTeam Lab: The AI Operating Agreement
Who delegates what, who reviews what, what never gets automated, and how changes get made. Draft your team's operating agreement live — one page, not a policy binder. You leave with your team's AI operating agreement, drafted.
Speaker announcing soonShip & What's Next
Every attendee's agent runs once, live. Then the road ahead: what the full AI Engineering Accelerator adds, your year of Ai+ access, and how certification works. You leave with your certificate and a personal next-90-days build list.
Speaker announcing soonPreliminary program. Session titles, formats, and speakers are subject to change as the agenda is finalized.
Speaker Lineup — Being Announced
The ODSC AI NYC program will feature practitioners, founders, executives, and technical leaders working on real AI-for-work deployments. Expect a mix of case studies, technical sessions, leadership discussions, demos, workshops, and curated networking.
Partners & Sponsors
ODSC AI NYC partners reach a room of practitioners, product leaders, and executives actively selecting and deploying AI tooling — through hosted talks, workshops, roundtables, demos, and VIP networking rather than a booth on an expo floor.
Speak to buyers, not passers-by
Hosted sessions and workshops in front of teams with live agent, copilot, and platform initiatives.
Curated introductions
Roundtables, the VIP dinner, and matchmaking with attendees matching your target profile.
Limited partner slots
A deliberately small partner cohort so each one gets real visibility across the two days.
Your logo here
Partner slot
Partner slot
Partner slot
Partner slot
Partner slot
From the Team Behind ODSC AI
ODSC has spent more than a decade bringing together the people building and applying data science, machine learning, and AI. ODSC AI NYC builds on that community with a focused summit for the next phase of applied AI: agents, copilots, workflows, evaluation, governance, and enterprise deployment.
Built for practitioners and technical leaders
Designed for both builders and decision-makers
Grounded in applied AI and real-world use cases
Connected to the broader ODSC AI event community
Venue
ODSC AI NYC will take place at Jay Conference Center Bryant Park, 109 West 39th Street, New York City.
Located in Midtown Manhattan, the venue is close to Bryant Park, Times Square, Penn Station, Grand Central, major subway lines, hotels, restaurants, and offices across Midtown.
Frequently Asked Questions
What is ODSC AI NYC?
A two-day in-person AI for Work Summit focused on agents, copilots, workflows, evaluation, governance, and enterprise AI deployment.
Who should attend?
AI practitioners, engineers, product leaders, operators, executives, founders, and teams responsible for bringing AI into real workflows.
Is this a technical event or a business event?
Both. The summit connects technical implementation with business adoption, governance, and measurable value.
Will there be hands-on workshops?
Yes. The program will include practical demos and workshop-style sessions. Details will be announced as the agenda is finalized.
Where is the event?
Jay Conference Center Bryant Park, 109 West 39th Street, New York City.
What is included in the Networking+ Pass?
Everything in the Regular Pass plus access to the speaker and VIP dinner, priority networking opportunities, and access to invited guests, speakers, and partners.
Are sponsorship opportunities available?
Yes. ODSC AI NYC offers limited sponsor opportunities focused on hosted talks, workshops, panels, roundtables, networking, VIP dinners, and curated demos.
Join us December 2–3, 2026 in New York City
Two focused days on AI for Work. Whether you are building agents, deploying copilots, redesigning workflows, evaluating AI systems, or leading adoption across your organization, ODSC AI NYC is designed to help you understand what is working now and what to build next.