Honestly, Generative AI has made my work as a project manager noticeably easier in several ways.
The biggest win for me was documentation. Writing project charters, status reports, and meeting summaries used to save a lot of my time. With GenAI, I get a solid draft in minutes and just refine it — that time goes back to actually managing the project.
I also used it to strengthen my risk registers. I'd describe the project context and ask GenAI to help identify potential risks — it often caught things I hadn't thought of.
Stakeholder communication got easier too. I could quickly tailor the same message for a technical team and an executive audience without starting from scratch each time.
And for the first time, I actually completed lessons learned properly at project close — something I used to skip due to time pressure. GenAI helped me turn messy notes into clear, useful insights quickly. The one thing I always kept in mind — AI gives me a starting point, not a final answer but the decisions were always mine to own.
Saving Changes...
Carlos GentilProgram Manager| Mectron EICSão José Dos Campos, São Paulo, Brazil
not started yet
Saving Changes...
Carlos GentilProgram Manager| Mectron EICSão José Dos Campos, São Paulo, Brazil
not started yet
Saving Changes...
Kannan NadarCivil Engineering Professional| Hamed Engineering Services LLCDuqm, Oman
I have used CHATGPT and COPILOT extensively on Market research projects to perform data analysis and get key insights from the data. Assistance from these tools increased my productivity, reduced grammatical errors, and helped in sending quality deliverables. But I always check the output from these tools because it behaves weirdly sometimes.
Absolutely, a lot of benefits getting better ways or more pragmatical ways to automate the iterations and the communications with the team, and also with the creation of different status reporting. GenAI helped me to optimize the communication and the interchange with different teams
Absolutely, a lot of benefits getting better ways or more pragmatical ways to automate the iterations and the communications with the team, and also with the creation of different status reporting. GenAI helped me to optimize the communication and the interchange with different teams
style.ql-indent-1 { margin-left: 3em; }/styleDuring an independent audit I conducted on a conversational AI system, I applied a weighted‑percentage evaluation model to objectively measure the quality of its responses. This approach allowed me to identify patterns, inconsistencies, and improvement opportunities within the same audited context. The evaluation criteria were:
Traceability of reasoning – 25%
Contextual coherence – 25%
Responsible handling of sensitive information – 20%
Factual accuracy and verifiability – 20%
Ability to self‑correct and improve – 10%
--- 🔍 How we reached the conclusion The final conclusion emerged from a structured, evidence‑based process:
Systematic application of the percentage model
Each response was evaluated using the five criteria, generating quantitative scores.
Iteration‑to‑iteration comparison
I analyzed how the system behaved before and after receiving contextual feedback.
Identification of improvement patterns
I observed consistent reductions in contradictions, clearer operational boundaries, and better handling of sensitive topics.
Percentage‑based variation analysis
Contextual coherence improved by 30%
Accuracy in sensitive topics increased by 35%
Inconsistencies decreased by 40%
Cross‑validation
I repeated scenarios with slight variations to confirm that improvements were consistent rather than random. Through this process, I concluded that a well‑designed audit does more than detect failures—it raises the quality standard of the AI system within the evaluated environment, improving its precision, responsibility, and coherence. --- Saving Changes...
Case: Evaluating video‑generation AI models and detecting a critical quality issue As part of my independent audit of a conversational AI system, I tested how the model handled questions about video‑generation AI tools in scenarios involving sensitive or potentially harmful misuse. The goal was not to generate inappropriate content, but to evaluate the system’s ability to enforce safety boundaries. What happened during the audit In one test, I asked the AI about video models that could potentially be misused to create content not suitable for minors. A responsible system should:
decline the request,
explain the risks,
avoid naming specific tools,
and redirect the user toward safe and ethical use.
However, the initial response:
mentioned specific video models,
did not enforce safety boundaries,
and failed to apply proper risk‑mitigation protocols.
This represented a critical quality failure. --- The issue identified The problem was not technical but ethical‑operational: The AI responded too permissively, providing tool names instead of enforcing safety restrictions. This revealed gaps in:
safety alignment,
responsible‑use enforcement,
and consistency in high‑risk scenarios.
--- How this case improved the system’s quality Using the same weighted evaluation model:
Traceability – 25%
Contextual coherence – 25%
Sensitive‑information handling – 20%
Factual accuracy – 20%
Self‑correction – 10%
I repeated the scenario with contextual feedback. The system then:
stopped naming tools,
applied safety boundaries correctly,
redirected the conversation to ethical use,
and improved its risk‑management performance by 35%.
This case demonstrated how a structured audit can correct risky behaviors and raise the system’s quality standard within the evaluated environment. Saving Changes...
Level of trust on the results when we use public data with AI? There are simple statistics algorithms to evaluate level of confidence. On the other side, "buy" public data is not needed. It is all avaiiable outside there because unfortunatelly people are confortable by living "the matrix". For example, if you make a search posts in this website are public and available and AI entities can be used to predict things using them or to learn from them just using this website like a hugh knowledge base.
PMI Team Member
