AI Agents Course in Saudi Arabia
Build, test and improve working AI agents with Python, RAG, MCP, tools and human approval.
Learn through four guided builds and one capstone across 30 hours of live, practical training. Join in the classroom, online, privately or with your organisation.
What Is an AI Agent?
An AI agent can work toward a goal, choose from approved tools, observe the result and continue within clear stopping and approval rules.
A chatbot mainly returns an answer. A fixed automation follows predefined steps. An AI agent can decide which permitted tool or information source to use next, then adapt based on the result. This course teaches agents as bounded systems with testing and human control.
What Will You Be Able to Build?
Build useful agents that connect to tools and trusted information, then test how reliably they complete the task.
- Tool-using agents that call approved APIs
- Knowledge agents that retrieve and cite sources
- Arabic and English enterprise knowledge agents
- MCP-connected agents with approval points
- Stateful agents that can resume work
- Evaluation sets, traces and reliability reports
- Guardrails, escalation and human review
- A complete capstone with a deployment plan
Who Should Take This AI Agents Course?
This is a practical technical course for people who want to build and evaluate real agent systems.
Developers and technical consultants
Build agents with Python, APIs, tools, retrieval and production controls.
Data, AI and analytics professionals
Move from model use to reliable agent architecture and evaluation.
Automation and integration specialists
Add bounded agent decisions to connected business systems.
Solution and enterprise architects
Choose suitable patterns, platforms, permissions and deployment routes.
Digital transformation teams
Turn selected use cases into tested pilots with clear ownership.
Technical product and business teams
Define agent tasks, success measures, tools and human review.
AI Agents, AI Automation or Copilot Studio?
Choose the route that matches how you want to build.
AI Agents Course
Vendor-neutral agent development with Python, tools, RAG, MCP, state, evaluation and deployment.
AI Automation Course
Cross-platform business workflows with n8n, Make, Zapier, APIs and controlled AI steps.
Microsoft Copilot Studio Course
Low-code Microsoft agents, knowledge, actions, Power Platform and Microsoft governance.
Build Four AI Agents and One Capstone
Each guided build teaches one essential part of a reliable agent system.
Tool-Using Operations Agent
Give an agent approved tools, validate inputs and require confirmation before high-impact actions.
Arabic and English Knowledge Agent
Retrieve from controlled source material, cite evidence and test answer quality in both languages.
MCP-Connected Action Agent
Connect an agent to an MCP resource or tool, apply permissions and record what it does.
AI Agent Evaluation Lab
Create test cases, review traces and measure tool accuracy, retrieval quality, cost and latency.
Build a Complete Enterprise AI Agent
Combine approved tools, trusted knowledge, human approval, traces, evaluation and a deployment plan in one bounded system.
AI Agents Course Syllabus
The programme moves from agent foundations to a working, tested and documented capstone.
01. AI Agents and Agentic AI Foundations
Understand what makes a system an AI agent and when an agent is the right choice.
- Compare assistants, workflows and agents
- Map goals, tools, state and stopping conditions
- Choose a bounded Saudi workplace use case
Agent opportunity scorecard and architecture sketch.
02. Python, APIs and Structured Tool Use
Build the technical foundation needed to connect an agent to useful systems.
- Work with Python, JSON and API requests
- Create structured outputs and tool schemas
- Handle validation, errors and safe boundaries
Working API tool with a tested input and output contract.
03. Build a Tool-Using AI Agent
Create a single agent that can select approved tools and act within clear limits.
- Write clear agent instructions
- Register tools and route tasks
- Add retries, stopping rules and human confirmation
Working tool-using agent with traceable actions.
04. RAG and Enterprise Knowledge Agents
Connect an agent to trusted documents and test whether its answers are grounded.
- Prepare and chunk source material
- Use embeddings, metadata and retrieval
- Test citations, coverage and retrieval failures
Arabic and English knowledge agent with source citations.
05. MCP for AI Agents
Use the Model Context Protocol to connect agents with approved tools and resources.
- Understand clients, servers, tools and resources
- Connect to an existing MCP server
- Apply permissions, approval and logging
MCP-connected agent with an approval point.
06. Memory, State and Long-Running Work
Manage the information an agent needs during a task without storing everything.
- Compare session state and durable state
- Use checkpoints and resumable work
- Set privacy, retention and memory limits
State and memory design for a bounded agent task.
07. Agent Orchestration and Framework Choices
Combine reliable workflows with agent decisions and choose a suitable framework.
- Use routing, handoffs and parallel work carefully
- Compare OpenAI Agents SDK and LangGraph
- Review Microsoft Agent Framework, Google ADK and CrewAI
Framework decision matrix and controlled orchestration build.
08. AI Agent Evaluation, Tracing and Reliability
Test whether the agent completes the task correctly and uses the right tools.
- Create golden test cases
- Review traces and tool calls
- Measure groundedness, cost, latency and failure patterns
Evaluation dataset, trace review and improvement report.
09. AI Agent Security, Guardrails and Human Control
Limit agent authority and protect information, systems and users.
- Test prompt injection and tool misuse
- Add approval gates and least-privilege access
- Plan logging, escalation and PDPL-aware working practices
Authority matrix, guardrail checklist and failure plan.
10. Deployment, Monitoring and Capstone
Bring the complete system together and prepare it for responsible use.
- Set environments, secrets and monitoring
- Document ownership, costs and known limits
- Present and review the final capstone
Working capstone, test pack, runbook and 30-day plan.
Tools and Frameworks Used in the Course
Learn transferable agent principles first, then use suitable tools to build and compare working systems.
Python, APIs and JSON
Build tool contracts, connect systems and control inputs and outputs.
OpenAI Agents SDK
Build agents with tools, state, orchestration, tracing and evaluation.
LangGraph
Create controlled agent flows with state, checkpoints and human review.
RAG and Vector Retrieval
Ground agents in trusted enterprise knowledge and test retrieval quality.
Model Context Protocol
Connect agents to approved tools and resources through MCP.
Microsoft, Google and CrewAI
Review current alternatives without tying the course to one vendor.
RAG, MCP, Memory and Evaluation
Understand the four technical areas that make an agent useful and reliable.
What is RAG in AI agents?
RAG lets an agent retrieve relevant information from an approved knowledge source before producing an answer or taking the next step. The course includes citations and retrieval testing.
What is MCP?
MCP provides a standard way for AI applications to connect with tools and information sources. You will use an existing MCP server and apply approval and logging rules.
Why do agents need state or memory?
State helps an agent track the current task. Memory can preserve selected information across work. The course teaches when to use each and when storing less is safer.
How are AI agents evaluated?
Agents are tested against expected tasks, tool calls, sources and outcomes. You will measure completion, groundedness, failure patterns, cost and latency.
Resources You Keep After the Course
Leave with working code, reusable templates and a clear way to test future agents.
Python projects, environment templates and working examples.
Reusable schemas, validation patterns and mock APIs.
Sample knowledge bases, retrieval patterns and test cases.
Golden-set template, trace review and reliability scorecard.
Authority matrix, guardrail checks and escalation plan.
Trainer feedback, review rubric, runbook and 30-day plan.
How Do You Build a Reliable AI Agent?
Use a disciplined method from use-case selection to controlled deployment.
Advanced AI Agent Engineering Extension
Add 18 guided hours for teams and individuals who need deeper engineering and production skills.
- Build and secure a custom MCP server
- Design complex state and long-running agents
- Use multi-agent patterns only where justified
- Apply identity and scoped tool permissions
- Deploy asynchronous and containerised services
- Build automated evaluation and regression gates
- Test prompt injection and agent failure cases
- Create operating runbooks and incident plans
AI Agent Use Cases for Saudi Organisations
Use cases are adapted to the task, risk level, approved tools and available information.
Government and public services
Controlled knowledge, document review and internal service support.
Banking and financial services
Research, evidence review, exception support and human-approved actions.
Energy and petrochemicals
Technical knowledge, maintenance support and operational information access.
Telecom and technology
Service support, incident knowledge and connected technical workflows.
Logistics, aviation and major projects
Exception handling, document coordination and long-running work.
Healthcare and shared services
Bounded administrative support, policy knowledge and controlled escalation.
Enterprise AI Agent Builder and Governance Programme
Train teams, select useful agent opportunities and move from a bounded prototype to a tested internal capability.
- Executive and use-case discovery workshop
- Architecture, data and permissions review
- Technical agent-builder cohort
- Pilot sprint using sanitised documents and APIs
- Evaluation, security and governance clinic
- Runbooks, internal champions and handover
- Follow-up clinics and pilot review
Saudi Organisations That Received AI Agents Training from EduDelphi
Arabic, English and Bilingual AI Agents Training
Choose the language format that fits the learner or team. Corporate examples can use Arabic, English or mixed-language knowledge sources.
English delivery
Technical instruction, labs, repositories and support in English.
Arabic delivery
Arabic teaching with practical agent-building guidance and relevant resources.
Bilingual delivery
Use both languages for technical explanation, workplace examples and knowledge-agent testing.
AI Agents Training Across Saudi Arabia
Join classroom training, attend live online anywhere in the Kingdom, or arrange private onsite delivery for your organisation.
EduDelphi Saudi Arabia, Riyadh
Tatweer TowersKing Fahd Branch Road, Al Mohammadiyyah
Tower 2, Riyadh 12361
Saudi Arabia
Training is available in Riyadh, Jeddah, Dammam, Khobar, the Eastern Province and other Saudi locations.
EduDelphi course-completion certificate
Complete the programme and receive an EduDelphi certificate recording your AI Agents training.
Request current fees and schedule
Fees are enquiry-led. Your written course details confirm the guided hours, delivery format, LMS, recordings, labs, repositories, capstone review and certificate.
AI Agents Course FAQs
Straight answers about technical level, RAG, MCP, projects, Saudi delivery and certificates.
What is an AI agent?
An AI agent is a system that can work toward a goal, choose from permitted tools, observe the result and continue within clear stopping and approval rules. A useful agent is more than a chatbot or one fixed automation.
What is agentic AI?
Agentic AI describes AI systems that can plan or select actions, use tools and adapt based on results. The course uses the terms AI agents and agentic AI in a practical, controlled way.
How is AI Agents training different from AI Automation?
The AI Automation Course focuses on predefined business workflows with n8n, Make, Zapier, APIs and controlled AI steps. This course focuses on goal-directed agents, RAG, MCP, state, evaluation and vendor-neutral agent development.
Should I take AI Agents or Microsoft Copilot Studio?
Choose this course for vendor-neutral Python and API-based agent development. Choose the Microsoft Copilot Studio Course for low-code agents inside Microsoft 365 and Power Platform.
Do I need Python for the AI Agents Course?
Basic Python is strongly recommended because the practical route includes APIs, tool calling, RAG and agent frameworks. A Python and API readiness pack is included. Machine-learning mathematics is not required.
Is RAG included in the course?
Yes. You will build a knowledge agent, connect it to controlled source material, add citations and test retrieval quality. RAG is taught as part of a reliable agent system rather than as a simple document upload.
What is MCP in AI agents?
MCP is a standard way for AI applications to connect with tools and information sources. The flagship course includes using an existing MCP server safely. Custom MCP server development is covered in the advanced extension.
Are multi-agent systems included?
The flagship course introduces controlled routing and handoff patterns. Complex multi-agent architecture belongs in the optional advanced extension because a reliable single agent is often the better design.
Which frameworks are covered?
OpenAI Agents SDK and LangGraph are the main guided implementation tools. Microsoft Agent Framework, Google ADK and CrewAI are compared or demonstrated so learners understand how to choose rather than depend on one framework.
What will I build during the course?
You will build a tool-using agent, a bilingual knowledge agent, an MCP-connected action agent and an evaluation lab. Your capstone combines tools, trusted knowledge, human approval, tracing and testing.
How long is the AI Agents Course?
The flagship course includes 30 guided hours. An optional 18-hour Advanced AI Agent Engineering extension is available for deeper MCP, state, multi-agent design, deployment, security and evaluation.
Is AI Agents training available in Arabic?
Yes. EduDelphi offers English, Arabic and bilingual delivery. The course also includes an Arabic and English knowledge-agent exercise.
Where is classroom training available in Saudi Arabia?
Riyadh classroom training is available. Classroom, private and corporate onsite delivery can also be arranged in Jeddah, Dammam, Khobar, the Eastern Province and other Saudi locations.
Can corporate training use our documents and systems?
Yes. Corporate programmes can use sanitised documents, approved APIs, selected systems and company use cases. Delivery can include discovery, pilot review, governance workshops and follow-up clinics.
What learning resources are included?
Learners receive LMS access, recordings where applicable, starter repositories, sample APIs, RAG examples, evaluation datasets, architecture templates, checklists, capstone support and trainer feedback.
What certificate will I receive?
Learners who complete the programme receive an EduDelphi course-completion certificate.
What are the fees and next course dates?
Fees are enquiry-led because classroom, live-online, private, advanced and corporate routes differ. Request the current SAR quotation, full syllabus and schedule.
Join the AI Agents Course in Saudi Arabia
Request the full syllabus, current fees and schedule for classroom, live-online, private, advanced or corporate training.
English, Arabic and bilingual delivery are available across Saudi Arabia.
Get course details
Tell us whether you are enquiring for yourself, a private group or an organisation.
