Learn Agentic AI: From Low-Code to Code

Build production-grade agents with OpenAI AgentKit

This Free Course is Offered by: Panaversity

Register in this Free Panaversity Course: Click to Register

Live YouTube Classes @panaversity channel on Wednesday and Thursday and 8:00 pm - 10:00 pm: Class starting from Oct 15, 2025. Recording Also Available on @Panaversity YouTube Channel

YouTube Recording Playlist

All-Pakistan and All-Star Faculty: Qasim (Karachi), Ameen (Saudi Arabia), Wania (Wah Cantt.), Junaid (Lahore), Aneeq (Karachi), Hammad (Peshawar), Rehan (Islamabad) and Zia (Karachi).

Here’s a practical, non-coding syllabus which we are running as a 5-week cohort (8 x 2–3h sessions) or we will also offer it as a 2-day intensive in Companies and Universities. It stays hands-on with AgentKit’s visual tools—no programming required.

No-Code Agents with OpenAI AgentKit — Course Syllabus

Course overview

Build, test, and ship a production-ready AI agent using OpenAI’s new AgentKit—focusing on the visual Agent Builder, governance & guardrails, built-in evals, and one-click chat UI embedding with ChatKit. By the end, learners publish a working agent and a simple plan to measure impact and iterate. (OpenAI)

Who it’s for

Non-programmers (general managers, accountants, engineers, doctors, financial managers, product managers, ops, Customer Experience (CX), Human Resources / Learning & Development (HR/L&D), consultants, educators, everyone) who can define workflows and business outcomes, but don’t want to write code.

What you’ll build

A capstone agent of your choice—for example:

• Customer support triage & answer bot
• Research & brief generator for sales/BD
• Internal knowledge assistant (policies/SOPs)
• Classroom co-teacher / onboarding guide

Learning outcomes

By the end, participants can:

1. Understand the concepts of Agentic Web and Agentic Organizations
2. Describe core agent concepts (agents, tools, handoffs, sessions, guardrails) in plain language and map them to business workflows. (OpenAI GitHub)
3. Use Agent Builder to design multi-step workflows on a visual canvas, version them, and add approvals/guardrails without code. (OpenAI)
4. Connect knowledge safely (files/connectors) and set up the Connector Registry with admins. (OpenAI)
5. Deploy a branded chat experience with ChatKit and share it with test users. (OpenAI)
6. Measure and improve quality with Evals (datasets, trace grading, prompt optimization), and understand when to consider reinforcement fine-tuning. (OpenAI)
7. Apply Guardrails for safety (PII masking, jailbreak detection, hallucination checks) using the wizard and presets. (OpenAI GitHub)

Format & pacing

• 6 weeks · 2 sessions/week · 2–3 hours/session (+ extra Question/Answer Sessions)
• Alt: 2-day bootcamp covering the same modules with condensed labs to be offered in Companies and Universities
• Delivery: live workshop or flipped classroom (short videos + guided labs)

Detailed Schedule

Week 1 — Agentic Web and Organizations

Session 1: Agentic Web (2h)

Session 2: Agentic Organizations (2h)

Week 2 — AI and Human Readiness & No-Code Agents first workflow

Session 3: AI and Human Readiness (2h)

• Assignment 1 Real AI Agents Reserach Discussion
• Introduction to AI Agents

Session 4: What is an “agent” (no code)? (2h)

• The AgentKit stack at a glance: Agent Builder (visual), Connector Registry (admin), ChatKit (UI), Evals (quality), Guardrails (safety).

• Concepts in human terms: tasks, tools, multi-step flows, agent handoffs, memory/sessions.

• Demo tour of Agent Builder: canvases, nodes, versions, templates; publish/preview lifecycle.

  Lab: Clone a template, customize instructions, add an approval step, run test conversations. (OpenAI)

Week 3 — Knowledge Connectivity & Design Patterns

Session 5: Connecting knowledge safely (2–3h)

• What to put in vs. link to; file search basics; connector options (e.g., Drive, SharePoint, Teams) and MCP servers.

• Data handling patterns: least-privilege, redaction, auditability.

  Lab: Attach a small policy pack (PDFs/Docs), configure retrieval, and test relevance safely. (OpenAI)

Session 6: Agentic design patterns (2–3h)

• Drag-and-drop nodes: tools, file search, guardrails, decision/branching, human-in-the-loop.

• Multi-agent patterns via handoffs—when and why to split responsibilities.

• Versioning & change logs; rollback and safe launches.

  Lab: Build a 5–7 node workflow from a blank canvas; add a human approval and a fallback path. (OpenAI)

Week 4 — Evaluations, Branding the Chat UI & Deployment

Session 7: Deploy with ChatKit (2h)

• Shipping a usable interface without front-end work: embed options and theme/brand tweaks.

• Sharing with pilot users; capturing transcripts and feedback for iteration.

  Lab: Deploy your agent’s chat UI, set a custom name/avatar, and invite 3 pilot testers. (OpenAI)

Session 8: Evals you’ll actually use (2–3h)

• Designing a simple eval dataset from real tickets/prompts.

• Trace grading: grading whole runs to spot brittle steps.

• Prompt optimization: generate improved prompts from grader + human annotations.

• Third-party model comparisons (what, why, when).

  Lab: Create a 20-case eval, run it, and apply one optimization round. Re-run and compare. (OpenAI)

Week 5 — Quality: Guardrails, Observe, Governance and Iterate

Session 9: Observability & cost control (2h)

• Reading traces; identifying tool-call loops and dead ends.

• Lightweight A/Bs: instruction tweaks and guardrail thresholds.

• When RFT helps (conceptual only) and how to scope an RFT request with your tech team.

  Lab: Use traces to remove one redundant step and reduce tokens/time per task. (OpenAI)

Session 10: Guardrails & governance (2–3h)

• The Guardrails Wizard: select checks (moderation, jailbreak, PII, hallucinations) and set policies—no code.

• Human approvals & audit trails; rollout controls and change management.

  Lab: Add PII masking + jailbreak detection to your agent, document your policy, and re-test. (OpenAI GitHub)

Week 6 — Capstone

Session 11: Capstone build & publish (2–3h)

• Finalize workflow, run evals, harden guardrails, and polish the ChatKit UI.

• Write a one-page “launch note”: scope, metrics, SLAs, rollback plan.

  Capstone demo: 5-minute live run + Q&A; submit your Launch Note and a 14-day iteration plan.

Assessments & deliverables

• Checkpoints (30%): Session labs (Builder, connectors, evals, guardrails).
• Capstone (50%): Working agent, embedded chat UI, eval results, safety config.
• Launch Note (20%): Clear goals, success metrics, and iteration plan.

Tools & accounts (no coding needed)

• Agent Builder (visual canvas, versioning, preview/publish)
• Connector Registry (admin-managed app & data connections across orgs)
• ChatKit (embedded chat UI with branding controls)
• Evals (datasets, trace grading, prompt optimization, optional third-party models)
• Guardrails (wizard + presets for PII/jailbreak/hallucination checks) Availability: as of Oct 6–8, 2025, Agent Builder is in beta; ChatKit and new Evals features are generally available; Connector Registry is rolling out in beta via the Global Admin Console. All are included under standard API model pricing. (OpenAI)

Instructor prep & room setup

• Learners need: laptop, browser, sample documents (FAQs, policies, SOPs), and access granted to AgentKit features in your org.
• Create an org sandbox + test connectors; prepare three Builder templates per track (support, research, internal knowledge).

Optional 2-day bootcamp agenda (condensed) to be offered in Universities and Companies

Day 1 AM: Foundations + first workflow (Sessions 1–3) Day 1 PM: Data connections + ChatKit deploy (Session 4) Day 2 AM: Evals + optimization (Sessions 5–6) Day 2 PM: Guardrails + capstone launch (Sessions 7–8)

Four learning levels

After completing this course (Level 1) you will be at Learning Level 1, after that you will be ready for Levels 2, 3, and 4.

• Learning Level 1: No-Code Agent Development — build end-to-end in AgentKit (This Course)
• Learning Level 2: Code-First Chat GPT Apps SDK — Learn Python and start in Apps SDK, then import visual flows where helpful (This course will be offered by Panaversity later).
• Learning Level 3: Full-Code Agents SDK - Learn Agentic AI using OpenAI Agents SDK and MCP
• Learning Level 4: Full-Code AI Assisted: Spec-Driven Vibe-Coding - Spec-Kit Plus

References

• OpenAI Introducing AgentKit (features, availability, pricing). (OpenAI)
• OpenAI Agent platform overview (Builder, ChatKit, Evals). (OpenAI)
• Agents SDK docs (background on agents/handoffs/guardrails—used conceptually here). (OpenAI GitHub)
• Guardrails docs (wizard & presets). (OpenAI GitHub)
• Launch coverage & context. (TechCrunch)

If you want, I can tailor this to a specific audience (e.g., support orgs, schools, or consulting firms) and pre-fill the Builder templates for that track.