Developing and delivering Bahrain's first PMI-CPMAI cohort

The difficult part was not explaining AI. It was explaining how to deliver it.

Most professionals have now seen enough AI demonstrations to understand its potential. Managing an AI initiative is another matter. Before a model is selected, somebody has to establish whether there is a genuine business problem, whether AI is necessary, and whether the data exists to support a credible solution.

When BIBF asked me to help introduce PMI-CPMAI in Bahrain, I saw an opportunity to bridge a gap I had encountered in delivery work: capable project managers who know how to manage a programme but have not necessarily had to question a dataset, an evaluation metric or a model's behaviour once it is live.

Develop the material, then build a course people could actually follow.

The work began months before the classroom sessions. I developed preparatory material, completed PMI's trainer requirements and worked with BIBF to organise a programme around the official certification pathway. The audience ranged across financial services, telecommunications and public-sector organisations, with different levels of technical experience.

The teaching went well beyond GenAI prompts. We worked through the six-phase AI project lifecycle: business understanding, data understanding, data preparation, model development, evaluation and operationalisation. I used concrete examples to connect each stage to decisions a sponsor, data scientist or delivery lead would need to make.

• Establishing whether AI is the right solution to the business problem
• Data ownership, quality, stewardship and readiness
• Training and inference pipelines; overfitting and drift
• Precision, recall and F1 rather than headline accuracy
• Governance, explainability, fairness and life after deployment

One room, several different technical languages.

The first group brought together 12 professionals from organisations in banking, telecommunications and other sectors. Some worked closely with data and AI; others came from innovation, product and business transformation. The same concepts had to be technically sound and intelligible across that range.

We made the sessions interactive and drew on real delivery situations. The conversations improved the teaching: a data scientist challenged the technical assumptions, while someone from the business side challenged whether the proposed use case would create measurable value. Both questions belong in a well-run AI project.

Measured results, and a useful lesson for the next intake.

All 12 enrolled participants completed the five-day course. The BIBF evaluation received 12 responses, recorded an overall score of 4.73 out of 5 and rated the instructor 4.69 out of 5. All 12 respondents said they would recommend the course. This is feedback on the preparation programme, not a claim that participants have passed the separate PMI certification exam.

The feedback also showed where the format needs work. Twenty-five hours across five consecutive late-afternoon and evening sessions was demanding for a technically dense subject. Several participants recommended spreading the learning over more days, with shorter sessions or a morning schedule. I agreed and passed that recommendation to BIBF. Good course delivery should improve the next cohort, not merely close the first one.

A small first cohort, a meaningful start.

What matters to me about this assignment is that we introduced a structured approach to managing AI projects in Bahrain, rather than just another introduction to the technology. BIBF's first PMI-CPMAI programme is complete, and the institute has publicly recognised the inaugural cohort.

I am grateful to BIBF for trusting me with the material development and delivery, and to the participants for the questions and experience they brought into the room. Their next step is the formal certification examination; this programme was designed to help prepare them for it.