Earns Sophyx Certified AI Agent Builder – Professional
A 12-lesson course in 4 modules for people who already run AI workflows and want to build an AI agent that plans, uses tools, remembers and stays safe, with low code: what makes an agent (the loop of plan, act, check and decide, when a fixed workflow is better, a job card with a finish line, a never list, stop rules and a budget, and written plans with curated memory); tools and data (tool descriptions, narrow read, draft and act tools, MCP servers connected with trust and least access, and a short source list with a source beside every fact and an access map); making it reliable (evals of whole runs with rule, rubric and human graders and several runs per case, guardrails the system enforces against prompt injection and excessive agency, and approvals on the exact action for everything that can’t be undone); and shipping and improving it (traces, health signals, alerts and audit logs, cost per run and hard caps, and change control with an improvement record, rollback and a one-page operating guide). The third course of the AX line (AX-100 and AX-200, then AX-300). Worked example: Ines, the one-person marketing team at Larchmere Tea Co., a fictional online loose-leaf tea shop with a monthly subscription box, who builds its marketing agent.
12 lessons
38 min of video
40-question exam
Certificate valid 12 months
Regular price $49.99, free for a limited time.$49.99FreeLimited-time offer
Every lesson, the exam and the certificate, free while the offer runs.
Telling an agent from a chat and a workflow, the agent loop (plan, act, check, decide) and when a fixed workflow is the better choice; a standing job card with a goal and a finish line, a scope, a never list, stop-and-ask rules and a budget; written plans, working and long-term memory, and deciding what an agent may remember.
Tell an agent from a chat and a workflow
Write a goal with a clear finish line
Ask the agent for a plan before it acts
Module 2
Tools and Data
Tools as described actions the agent asks for and the system runs: clear descriptions, narrow tools, and read, draft and act tools kept apart; MCP servers and their tools, resources and prompts, connected only from trusted makers, with only the needed tools switched on at the narrowest access; retrieval from a short list of current, owned sources with a source beside every fact, outside text treated as information, and an access map with the least access for every tool and source.
Write a tool description an agent can follow
Explain what an MCP server offers an agent
Choose the sources an agent may search
Module 3
Make It Reliable
Evals of whole runs on real cases with the right result written first, including cases where the agent should not act, scored by rules, a rubric and a person, several runs per case and a pass only if every run passes, re-run after every change; guardrails the system enforces against prompt injection and excessive agency, with outside text labelled and the risky mix of outside text, private data and a way out broken; approvals on the exact action for everything that can’t be undone, tiered by what can be undone, with decline and silence meaning nothing happens.
Write eval cases with the right result first
Spot text that tries to give an agent orders
Gate every action that can’t be undone
Module 4
Ship and Improve
Traces of every run, a few health signals each week, alerts for tool failures, limits and never-list attempts, and audit logs without personal data; cost per run and per useful result, with hard caps on steps, time and spend in the settings; one change at a time, an improvement record, evals before going live, rollback to the last good version, re-testing when the model changes, and a one-page operating guide.
Read a trace to find why a run went wrong
Work out what one run costs
Change one thing at a time and re-test
Syllabus
Every lesson
Lessons run 2 to 3 minutes. Each has captions, a transcript, key terms, a practice quiz and a short exercise for your own business.
Four questions after each lesson. Try as often as you like: every wrong answer explains why it is wrong. One attempt finishes the lesson.
Final exam
40 questions about situations at different businesses, 60 minutes, pass at 75%. Everyone gets their own mix of questions. 24 hours between attempts, up to 3 in 30 days.
The certificate
Sophyx Certified AI Agent Builder – Professional
It carries your name, a unique credential ID and a public page where anyone can check it. Add it to LinkedIn in one click, or download it as a PDF. Valid for 12 months, because AI tools change fast.
A Sophyx credential, not an accredited qualification.
AX-200 (AI Workflow Specialist) is recommended, not required; no programming: learners read and edit plain structured text such as tool descriptions, settings and test cases in a generic agent builder. All brands and people in examples are fictional; all numbers are illustrative. Assistants, agent builders and servers are always generic; open standards (MCP, the OWASP Top 10 lists, the EU AI Act) are named with their version or date. Platform details (features, limits, prices) are as of October 2026 and change. Nothing in the course is legal advice, and completing it does not make anyone compliant with any law.