After completing the course, participants can:

• Write structured, reusable prompts for their own business tasks using the six-block template;
• Diagnose weak AI outputs, identify the missing building block and correct it through a fast iteration loop;
• Apply data-safety rules in daily use, keeping personal data, trade secrets and credentials out of unapproved tools;
• Select automation candidates with the frequency-structure matrix, set up a team prompt library, and run three workflow patterns (meeting minutes, inbox and document triage, periodic reporting) with human oversight, measuring the time and quality gains.

Module 1: Strategic prompt architecture (the brilliant-intern model, how LLMs read your words, the six building blocks: role, context, task, format, constraints, examples; a reusable template)
Module 2: Diagnosing bad outputs (four common failures and their fixes, the iteration loop, quality boosters, data-safety red lines)
Module 3:From one good prompt to a working system (what to automate, the team prompt library, three ready workflow patterns, human review gates, measuring the gain)

Dr. Doğu SIRT

Dr. Doğu Sırt has been a lecturer at the Faculty of Computer and Informatics at Istanbul Technical University since 2015. He holds degrees in computer engineering and business administration, a master’s degree from Gebze Technical University, and a PhD in computer engineering from the National Defence University (2023). Alongside his academic work, he is a corporate trainer, delivering courses on Python programming, artificial intelligence, data science, and data engineering for universities, public institutions, and the private sector. He also serves as a project referee and consultant for KOSGEB, ITU ARI Teknokent, and 1773 Teknokent, supporting the entrepreneurship ecosystem. His work focuses on integrating AI into business processes and turning AI into a strategic element that creates sustainable business value.

Most disappointing AI outputs are caused by missing instructions, not missing ability. Starting from this idea, the course treats prompting as written delegation and builds a strategic prompt architecture out of six building blocks: role, context, task, format, constraints and examples, ending with a fill-in-the-blanks template participants can reuse for any business task. The second module is a diagnosis-and-treatment kit: the four most common failures (vague ask, overloaded prompt, missing context, blind trust) with a fix for each, the iteration loop, three quality boosters (step-by-step reasoning, self-checks, rubrics) and the non-negotiable data-safety rules on personal data, trade secrets and credentials.

The final module moves from individual skill to organisational gain: choosing what to automate with a frequency-structure matrix, building a team prompt library, three ready workflow patterns (meeting-to-actions, inbox and document triage, the report pipeline), human-in-the-loop review gates, and two simple measures for proving the gain. The course is part of the BSC AI Factory training series. The BSC AI Factory is a EuroHPC JU AI Factory run by a Spain, Portugal, Türkiye and Romania consortium, with TÜBİTAK as the Turkish partner; the closing section points to the compute and expert-support opportunities offered through this ecosystem and TRUBA.

Basic AI literacy: familiarity with concepts such as large language models, hallucination and bias, for example through the introductory course of this training series or equivalent. No coding or technical background is required.

None required

An internet-connected computer or tablet, a current web browser and audio-playback capability. No additional software or programming environment is required.