VizuaraJoin the next cohort Assist.
Check a document.
Prepare a decision.
A HANDS-ON BOOTCAMP FOR PROFESSIONALS & TEAMS
AI Agents for Enterprises
Build AI agents
for the work
you actually do.
Learn to connect AI to your company’s documents, tools, and business processes. Start with guided workflows, test realistic problems, and build a pilot your team can evaluate. No prior coding required.
Starts 31 October 2026Sat & Sun, 7–9 AM IST 8 lectures
Early-bird discount until 10 October
Led by Dr. Rajat DandekarCo-founder, Vizuara · PhD, Purdue UniversityTHE ENTERPRISE, REIMAGINED.FIG. 01
PEOPLE, POLICIES, AND ERP SERVICES—WORKING TOGETHER.Original AI-generated illustration
WHY WE’RE TEACHING THIS NOW
From supplier onboarding to purchase requests, the shift is already underway.
See what’s changing ↓THE NEXT TWO TO THREE YEARS2026–2029
More work will happen
between systems.
Agents connect the steps.
People set the direction.
Coordinate.
Connect teams and tools.
Keep approvals in the loop.
Delegate.
Let tested workflows run.
Bring exceptions to people.
Our outlook, not an adoption forecast.See it in a real business workflow ↓
Research behind the outlook +
Changing skills. WEF’s 2025 survey found employers expected 39% of core skills to change by 2030, across multiple economic trends. Read the report ↗
Value still matters. Gartner forecast that over 40% of agentic AI projects would be canceled by the end of 2027. This is a prediction, not an observed failure rate. Read the forecast ↗
HEADLESS ERP · AN INTERACTIVE EXPERIENCETRY THE WORKFLOW
“Add this supplier.
Raise a purchase request.”
Make a request. Review the evidence. Approve the change. Try the workflow below.
What does headless mean?
Work through a conversation. The ERP still enforces business rules and keeps the records.
01 Request02 Verify03 Approve04 Create
AN
Ananya · ProcurementChennai operations · Sample organization
v.
Operations agent
Guided simulation with synthetic records. No live AI, ERP connection, document upload, or external transaction.
WORKING CONTEXTREADY
BEHIND THE CONVERSATION
Identity & access→Policy knowledge→Supplier master→Purchasing→Audit log
Policy explains the rules.
Policies guide the plan. Current supplier status and budget come from the ERP.
A request is not a purchase order.
Create a supplier and a purchase request. PO release and payment activation remain separate approvals.
A retry must not create a duplicate.
Check the receipt after a lost response. Recover the existing record instead of creating another.
The real-world inspiration +
Salesforce describes agents guiding supplier applications, checking documentation, and coordinating approvals. Our walkthrough explores that business pattern in an independent, vendor-neutral teaching example. It is not a reproduction of the Dreamforce demo or a live enterprise integration.
Explore Salesforce’s supplier-onboarding use case ↗Would this workflow pay off? Open the calculator +
CHECK THE ECONOMICS
Would this workflow be worth automating?
Adjust the assumptions. This estimates the value of staff time released, not cash savings or guaranteed ROI.
Net capacity value = time released × hourly cost − AI usage − operating overhead. Excludes initial implementation cost, tax, rework, and demand changes.
WHAT COMPANIES ARE ALREADY DOINGINDIA & THE WORLD
Where agents are
being put to work.
Five familiar businesses. Five concrete workflows you can learn from.
A PROJECT YOU CAN BUILDSee the curriculum ↗
Independent industry examples, not Vizuara client claims. Deployment, pilot, and announcement status follow the linked sources.
THE TOOLKIT, WITH A PURPOSEMODELS · KNOWLEDGE · ORCHESTRATION · INTEGRATION
Know what to use.
And where it belongs.
A focused toolkit for decisions, knowledge, and reliable execution.
TypeSafe AI DECISION MODEL · EARLY ACCESSJev
Classify, extract, and route with typed outputs and confidence estimates.
IN THE SUPPLIER WORKFLOWRoute a document. Escalate uncertainty.
Meet Jev, from TypeSafe ↗Exploration lab; provider access may be limited. Recorded fixtures keep the exercise runnable.Compare routing accuracy against an LLM baseline. Valid types still need correct decisions.
LLM Wiki KNOWLEDGE PATTERN · VECTORLESS OPTION
Knowledge that stays
with the business.
Turn source documents into linked Markdown pages. Retrieve policy context through an index and file search—a vectorless option at modest scale.
sources/onboarding-policy.pdf↓ ingest, cite, reviewwiki/supplier-onboarding.mdwiki/index.md → answer + sources
Test citations, freshness, and permissions against hybrid RAG. Keep live ERP state in the ERP.
Read the original LLM Wiki approach ↗An approach, not a branded framework. The document symbol above is illustrative.
LangGraph
Model the workflow as state and transitions. Pause for a human decision and resume with a checkpoint.
BuildSupplier review → approval → ERP write, with explicit rejection and recovery paths.
Explore LangGraph ↗
MCP
Expose tools through a common protocol. Keep authorization and business rules enforced by the service.
BuildScoped supplier lookup and purchase-request tools with validated inputs.
Explore MCP ↗
TemporalTemporal
Carry a business workflow through outages, retries, and approvals that arrive hours or days later.
ExploreRecover a saved purchase request after a lost response, using receipts and idempotent activities.
Explore Temporal ↗One working system first. Alternatives with evidence.
Start with guided n8n workflows. Explore LangGraph, compare Wiki retrieval and Jev, and use Temporal for a recovery exercise.
Logos identify the technologies discussed; no partnership or endorsement is implied. Links go to original documentation. Tool coverage may evolve before the cohort begins.
WHAT YOU’LL LEARN8 LECTURES · 16 LIVE HOURS · 31 OCT–22 NOV 2026
From LLM basics
to your own AI agent.
Eight practical sessions. Start with the foundations, then build with the tools behind modern enterprise agents.
Your four weekendsEvery lecture: 7:00–9:00 AM IST (UTC+5:30)
WEEKEND 0131 Oct & 1 NovLectures 1–2
WEEKEND 027 & 8 NovLectures 3–4
WEEKEND 0314 & 15 NovLectures 5–6
WEEKEND 0421 & 22 NovLectures 7–8
Configure workflows, connect sample data, and test business outcomes.
Extend tools, orchestration, and deployment in the starter repository.
n8nGPT · Claude · GeminiLLM Wiki · RAGMCPLangGraphJevLangfuseTemporal
All backgrounds welcome. Use guided n8n workflows and starter notebooks. Experienced developers can take the optional Python extensions. Plan for 3–5 hours of practice each week; provider API charges may be separate.
THE FINAL PROJECT
Bring a business problem.
Build a working pilot.
Choose a workflow from your own industry. Connect knowledge and tools, add human approvals, and show that the result works using synthetic or approved data.
Build your agent with us ↗enterprise_agent / release_checklistPILOT CHECKLIST
# What “ready for review” means
deliverables = {
"agent": "runnable service + API",
"evidence": "citations + scoped retrieval",
"controls": "approval + audit trail",
"evaluation": "golden set + failure tests",
"operations": "cost, latency + runbook",
"handoff": "demo + architecture memo"
}Procurement & ERP
Supplier onboarding and purchase requests.
Finance operations
Invoice exceptions and reconciliation review.
HR & people operations
Employee onboarding, leave, and policy queries.
Sales & customer success
Account research and proposal preparation.
Travel & hospitality
Booking queries and refund-case coordination.
Retail & e-commerce
Product discovery and order support.
Healthcare operations
Appointment coordination and document routing.
Legal & compliance
Contract review support and evidence collection.
YOUR TEACHERSRESEARCHERS, ENGINEERS, AND EDUCATORS
LEAD INSTRUCTORLearn with
Learn with
Dr. Rajat Dandekar.
Co-founder, Vizuara
PhD, Purdue University · B.Tech & M.Tech, IIT Madras
Rajat teaches reinforcement learning, AI agents, and reasoning models. His work at Vizuara connects the underlying ideas with the practical decisions involved in building and evaluating AI systems.
In this bootcamp, you’ll work through the design of an enterprise agent, implement its tools and controls, and test whether it does the job you intended.
Meet the Vizuara teaching team ↗The wider Vizuara team
Our co-founders teach across Vizuara’s AI programs. Rajat leads this bootcamp; any additional teaching sessions will be listed in the final schedule.
Dr. Raj Dandekar
Co-founder · PhD, MIT
LLMs, agents, and retrieval
Dr. Sreedath Panat
Co-founder · PhD, MIT
Computer vision and scientific ML
THE VIZUARA LEARNING COMMUNITYLearn with people
Learn with people
who keep building.
14,000+
students learning with Vizuara
Vizuara course library ↗Vizuara’s published figure across its course library; not a completion count or enrollment in this new bootcamp.JOIN ON YOUR OWN OR WITH YOUR TEAMINDIVIDUAL AMBITION. ORGANIZATIONAL IMPACT.
Your team knows the work.
Help them build the tools.
Give engineers, product owners, and technical leaders a shared foundation—and a working system.
Learn the engineering
from start to finish.
For professionals, founders, students, and domain experts—with or without a coding background.
- Guided practical labs, with optional code extensions
- A capstone repository and architecture narrative
- Engineering judgment that transfers across frameworks
Learn together.
Build shared capability.
Bring a group into the cohort. Discuss your team size, business workflow, procurement needs, and data constraints before enrollment.
- One shared learning path for your technical team
- Choose a capstone relevant to your business
- 7% corporate discount for teams of 26 or more
- Request a team quote and invoicing details
Dedicated private cohorts, custom integrations, and private-data deployment are scoped separately and subject to availability.
FEES & ENROLLMENTSTARTS 31 OCTOBER 2026
Bootcamp fees
and optional support.
The prices below already include the early-bird discount, valid through 10 October 2026.
OR CHOOSE A BUNDLE
COMMON QUESTIONSA few things
A few things
to know first.
Can I join if I don’t know how to code?
Yes. Everyone can participate, including people who are not comfortable coding. We begin with LLM basics and use guided workflows, visual tools, and starter templates. You’ll learn to design, build, and evaluate an agent. Optional Python extensions give experienced developers more depth.
Will we build a real agent or a chatbot?
You’ll build a workflow that retrieves evidence, uses tools, and asks for approval before taking action. Guided templates make it accessible without starting from a blank code editor.
Can our company enroll a team?
Yes. Teams with more than 25 participants (26 or more) receive a 7% corporate discount. Use the team inquiry for a quote and invoicing. Private cohorts and custom delivery are scoped separately.
Can we use confidential company data?
Labs use synthetic or approved data. Do not bring confidential data without your organization’s permission and an approved environment. The curriculum covers access boundaries and data handling, but enrollment does not include a private deployment environment.
When does it start? Is there an early-bird offer?
The bootcamp starts Saturday, 31 October 2026. All eight lectures run on Saturdays and Sundays, 7:00–9:00 AM IST, through 22 November. The displayed prices already include the early-bird discount, valid through 10 October 2026. Choose your course or bundle in the fees section, then select Enroll now to continue on Vizuara.
Does the capstone mean our system is production-ready?
The goal is a working, measurable pilot and a clear handoff. Your organization must still complete its security, compliance, load, and integration reviews before production use.
AI AGENTS FOR ENTERPRISES · VIZUARA