Learning & Innovation Research Manager| Project Management Institute (PMI)Spain
Are you utilizing any specific checklists or protocols within your projects or company to assess your readiness for working with Generative AI data? I'm curious to know what strategies or tools you've implemented to prepare for integrating Gen AI into your workflows. Please share your approaches in the comments below! Saving Changes...
Although, I had an interest in GenAI I was unable to bring my company much support in this area. Yet, with this new training, I will definitely be providing GenAI solutions more readily in the future.
Thank you, Markus, are you implementing any of these items when integrating AI into your clients?
My personal readiness protocol involves using tools like Gemini, Copilot and ChatGPT to automate administrative overhead—such as generating meeting summaries, drafting stakeholder communications, and other company documents. Saving Changes...
Jignasa NaikVP Delivery| August InfotechSurat, Gujarat, India
Hello Claudia,
At August Infotech, our experience with Generative AI has evolved from simply using AI tools to building AI-enabled solutions for our own operational needs. This has shaped how we approach GenAI integration into software development and project delivery workflows.
Our practical checklist is:
1. Start with the business or operational problem. We first identify what problem we are trying to solve and whether GenAI can provide measurable value, such as reducing repetitive work, improving access to knowledge, accelerating analysis, documentation, development, or decision-making.
2. Identify the right point in the workflow. We look at where AI can assist or automate specific activities across requirements, research, development, documentation, testing, knowledge management, analytics, and reporting. We don't assume that every activity needs AI.
3. Establish a reliable source of truth. Where AI is being used to provide operational or business information, we need to define what information the system is allowed to use. For our internal knowledge solution, responses are grounded in approved organizational content, with source references and controls to prevent the system from generating answers outside its available knowledge.
4. Assess data and security before implementation. Client confidentiality and data protection are critical. We consider what information the AI system can access, where it is processed, and whether sensitive project information, source code, credentials, personal information, or proprietary data can be used safely.
5. Build guardrails into the workflow. We don't rely only on users to identify incorrect AI output. Where appropriate, the system itself should have controls for ambiguous requests, unsupported information, access permissions, and other defined boundaries. An important lesson for us has been that an AI system should know not only how to answer, but also when it should not answer.
6. Keep human accountability. AI-generated requirements, code, documentation, analysis, or recommendations are not automatically treated as final deliverables. The appropriate developer, QA professional, project manager, or subject-matter expert remains accountable for validation and the final decision.
7. Make AI activity traceable. For more advanced AI workflows, particularly agentic systems, we consider observability and traceability important. Teams should be able to understand what information, processes, or data contributed to an AI-generated result, especially when the output influences a business or project decision.
8. Measure the business outcome. We don't consider an AI initiative successful simply because people are using it. We look at measurable outcomes such as time saved, reduction in manual effort, faster access to information, improved delivery efficiency, adoption, quality, and decision-making speed.
9. Learn and continuously refine. AI capabilities and the risks associated with them are evolving rapidly. We therefore treat GenAI adoption as an ongoing improvement process, using practical experience, successful use cases, limitations, and lessons learned to refine our workflows and controls.
Our guiding principle has evolved from simply being "AI-assisted, human-approved" to:
"AI-enabled, rule-governed, traceable, and human-accountable."
For us, the real value of GenAI is not just generating content faster. It is about thoughtfully redesigning workflows so that AI can take on the right level of assistance or automation while maintaining appropriate controls, transparency, and human accountability.
We are working with GenAI tools customized for the usage of internal data. Our training tools for our users had just won an award since bringing the adavantages ans options to accelerate and improve analysis must be seen on any any desk. In the meanwhile first pathinders starts to work with agentic AI that could bring even more structured results for every day tasks. It is a must to reflect on theses tools Saving Changes...
PMI Team Member
