GEPA Energy Leadership Academy
Log in

GEPA Energy Leadership Academy

Leadership Workshop Series

Maximising AI in Your Workflow: From Prompts to Agents

A hands-on workshop on chat tools, deep research and no-code AI agents for everyday energy work.

Beginner Leadership Workshops Certificate on completion

Taught by Mr. Emmanuel Chidi Ashinze

Watch a 75-second preview

How a no-code agent turns a morning voice note into reminders

A working session on putting AI into everyday energy work: prompting for usable reports, running deep research, and building a no-code agent in Copilot Studio.

About this course

This course is built from a live GEPA leadership workshop on using AI to maximise your workflow. The speaker, an engineer who builds AI-driven automation tools for a living, treated it as a working session rather than a lecture. He asked people to open their laptops, typed prompts in front of the group, ran deep research in two different tools at once, and built an AI agent from scratch in Microsoft Copilot Studio while the group watched.

The course follows that arc. It opens with why manual, repetitive workflows cost the energy sector time and accuracy, then sets out the two ideas the rest of the session rests on: the large language model and the AI agent. From there it moves into prompting, which is where most of the practical gain sits, and into deep research as a way of getting a referenced draft report in minutes rather than days. The build then shows how to create a no-code agent that reads your task list, puts it in your calendar and emails you reminders, including the moment the agent stalls in testing and what that teaches about writing its instructions.

Read more Show less +2 paragraphs

The session ran short of time and the agent was left only partly configured, so the closing module is the questions the group actually asked: how to reach Copilot Studio without a work subscription, whether deep research output is safe to use in academic work, which tools generate video with sound, and whether a spreadsheet can be updated and reported on automatically. Those answers are kept as they were given, caveats included.

It suits energy professionals with no coding background who write reports, chase data and carry a lot of repetitive admin. By the end you should be able to write a prompt that returns something usable, run and judge a deep research report, and build and test a simple agent of your own.

What you will be able to do

  • Explain the difference between a large language model and an AI agent, and when each is the right tool.
  • Write prompts that specify structure, evidence, visuals and referencing rather than just a topic.
  • Run a deep research report in ChatGPT or Gemini and judge what still needs your own work.
  • Build a no-code agent in Microsoft Copilot Studio that reads your tasks and acts on your calendar and email.
  • Connect agent tools, test an agent, and diagnose why it stops and asks you a question.
  • Decide which parts of your own reporting, research and admin work should be handed to AI.

Who this course is for

  • Energy professionals who write reports and carry repetitive admin
  • Engineers and analysts with no coding background
  • Consultants and project developers working to research and reporting deadlines
  • Team leads deciding where AI fits into their unit's workflow
  • Postgraduate students and researchers in the energy sector

Curriculum

Course structure

  1. Module 1

    Why this matters for energy professionals

    2 lessons

    What rigid, manual workflows cost the sector, and why the speaker argues that using AI has stopped being optional for working professionals.

    • The hidden cost of rigid, manual workflows Free preview 4 minutes
    • "I don't use AI" is no longer a badge of competence 3 minutes

  2. Module 2

    The building blocks: models, agents and what they are for

    2 lessons

    The two definitions the rest of the session rests on, the main tools available, and the range of jobs they can take on.

    • Large language models and AI agents defined 4 minutes
    • What these tools are actually for 4 minutes

  3. Module 3

    Prompting and deep research

    3 lessons

    How to tell a model what you want, the prompting techniques behind that, and how to run and compare a deep research report.

    • Equal access, unequal results 5 minutes
    • Building a prompt that returns something usable 4 minutes
    • Running deep research, and agent mode 7 minutes

  4. Module 4

    Building an agent in Copilot Studio

    3 lessons

    The live build of a no-code agent that turns a morning task list into calendar entries and email reminders, including where it broke in testing.

    • Deciding what the agent should do, and creating it 6 minutes
    • Connecting the tools the agent needs 4 minutes
    • Testing the agent and reading what went wrong 4 minutes

  5. Module 5

    Questions from the floor

    3 lessons

    The session ran short of time, so it closed on questions: getting access to Copilot Studio, using deep research in academic work, video tools, and automating spreadsheets.

    • Getting into Copilot Studio without a work subscription 3 minutes
    • Reviewing the report, and using AI in academic work 4 minutes
    • Closing recap, video tools and spreadsheet automation 4 minutes