Director, Learning Design & Development| PMIAsheville, NC, United States
Validating and checking outputs is critical when working with AI systems like Generative AI. Such validation approaches may include establishing clear criteria, implementing strong testing protocols, and continuous refinement.
In your experience with AI, what are some best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals?
ADEBAYO OLAJIDEGeneral Manager| Specific Tools and TechniquesIKORODU, LA, Nigeria
Be very sure of what you are expecting. Be clear and state the exact
Saving Changes...
rhonda hodgePM I| Georgia-PacificGa, United States
In reading several of responses, basically agree. It depends on the information provided, your ability to validate the information and the development of your prompts.
The old saying "Garbage In - Garbage Out"... Saving Changes...
Giridharan JManager| Anthem biosciences limitedBommasandra, India
Basically if we are experise in the respective subject and technically sound, it can be easily identified that the information is accurate or not. As per my knowledge, we seek the help of chatbots, where we do not find an appropriate word or detials required to complete a task or troubleshoot something. Therefore, we must know the infromation provide are correct and we need to validate it too. Saving Changes...
When using AI systems is very hard to set the precision or accuracy of the responses. I love bringing in the Agile mindset here pretty much imagine if you are mentoring someone you do a Q&A and based on the reponses of your Mentee you give the feedback so that Mentee can align his/her thoughts in the direction that we hint similarly review the AI responses and using our rationale judgement
1- Give Feedback to the AI system
2- Rework on your promp and be specific on what is expected
3- Keep it short and conscise, guage the responses and slowly we can tune the AI system in a way to get the best output
4- Now the Tech. Solution that comes in for accuracy is havig specific set of APIs that talk to real and accurate data sources or use 2-3 outputs of LLMs and then analyze and bring the best in output.
When approaching the challenge of precision and accuracy in AI responses, it’s important to think about this from both a project management and Agile mindset perspective. As PMP professionals, we know that continuous refinement, stakeholder feedback, and adaptive planning are central to achieving better outcomes. The same applies to working with AI systems.
1. Feedback as a Control Mechanism
Just as we monitor and control project work, we must treat AI responses as deliverables to be reviewed. Providing structured feedback ensures alignment with objectives, similar to how mentoring or coaching sessions guide mentees toward the right path.
2. Prompt Re-work as Iterative Planning
In Agile projects, requirements evolve and get refined through sprints. Similarly, reworking prompts is an iterative process where clarity and specificity drive improved outputs. Each iteration should narrow ambiguity and communicate expectations clearly to the AI system.
3. Conciseness and Incremental Adjustment
Agile emphasizes simplicity and incremental delivery. Keeping prompts concise allows us to inspect results quickly, make judgments, and tune the system iteratively minimizing waste and focusing on value.
4. Technical Enablers for Accuracy
From a PMP lens, accuracy is not just process it also requires the right enablers. This could mean integrating APIs that connect to verified data sources, or using ensemble approaches (reviewing multiple LLM outputs and selecting the most accurate). These solutions act like quality assurance gates in a project lifecycle, ensuring decisions are based on reliable information.
AI is a broader term. Generative AI is just an ancient model but everything "explode" when Google published the new architecture called transformer in 2017. So, with that said, take into account that generative AI is just "predictive test with steroids" just simplifying the model. With that said, two key points has to be taking into account when somebody works with AI: 1-human in the loop. 2-AI without Data (today called data science discipline or big data or whatever) is the same thing that live without oxygen. Talking about generative AI all related to technology has almost not impact with relation to all related to non-technological roles and activities. What you stated about accuracy and things like that are easy to implement because there are a lot inside disciplines like statistics. Most of them to make things "a priori" to prevent instead of cure. Few organizations taking into account that when generative AI environments are put in place almost a new business unit has to be created where roles like lawyers, linguistic, diversity and inclusion specialist must be hire to help on put it in place.
Good scenario, Generative AI functions with both the human and data considerations Saving Changes...
Janet BoydContract Manager| Delta VacationsPowder Springs, GA, United States
Being concise as possible along with providing details helps with accuracy. If it doesn't provide the necessary feedback continue to engage with the system and provide feedback. Saving Changes...
For AI result accuracy, I specifically apply the following under the very broad spectrum of aligning project execution with FIDIC suite of Contracts or any Contract Agreement reference (i.e. EPC, NEC) applicable to hybrid project management:
1. Define Your Goal Clearly
• Specify the exact question or task.
• Identify the desired output: summary, clause, risk table, etc.
• Indicate the target user (internal team, Engineer, Employer, Consultant).
• Example: “Summarize key risk allocation clauses under FIDIC Red Book for sewerage D&B;projects.”
2. Provide Sufficient Context
• Mention project type (STP, sewer network, water supply, etc.).
• Indicate contract type (FIDIC Red Book, D&B;, Framework Agreement).
• Include stage (Pre-contract, Construction, Post-contract).
• State perspective (Employer, Engineer, Contractor, or Consultant).
3. Evaluate for Accuracy
• Check if aligned with FIDIC principles and real clauses.
• Ensure technical and legal terms are correct.
• Verify realism of examples for your project setup.
• Cross-check against FIDIC 2017, company SOPs, and regulations.
4. Check for Relevance and Completeness
• Ensure the response answers your question directly.
• Verify all critical details are covered (timelines, risks, references).
• Request localized context if needed (e.g., Philippines PPP setup).
5. Cross-Verify with Human Expertise
• Review with Contracts or Legal team.
• Confirm practicality with Engineer’s Representative or PM.
• Never issue AI outputs without human review.
6. Refine Through Follow-ups
• Ask AI to recheck references (e.g., FIDIC Sub-Clause 20.2).
• Request multiple versions for comparison.
• Adapt results for memos or internal reports.
7. Keep a Record
• Save prompts and outputs for traceability.
• Label outputs by date and topic (e.g., '2025-10-18 Delay Claims Summary').
• Maintain for audits and management presentations.
8. Apply Human Insight
• Use professional judgment and site knowledge.
• Ask: “Would this hold up in Engineer’s evaluation or arbitration?” Saving Changes...
Ensuring GenAI outputs are accurate and aligned with goals requires to set clear evaluation criteria to the intent of the project, not just technical accuracy but usefulness and coherence and continously refine through feedback loops to learn from each iteration. Saving Changes...
David Enrique Velez BarretoFull-time Student| University of Puerto Rico, Mayaguez CampusMayaguez, Puerto Rico
As a senior project management student, I ensure AI results are accurate and aligned with project goals by setting clear objectives, crafting structured and contextual prompts, and validating outputs against reliable data or expert sources. I refine prompts through iteration and always maintain ethical and organizational alignment. This disciplined approach allows AI to enhance decision-making accuracy while supporting sound project management judgment. Saving Changes...
Validating and checking outputs is very important when working with AI. Here are a few situations where this matters: testing scripts, generating reports, creating automation, or making decisions based on AI results.
Even if we know what we want to achieve while prompting, it’s always a good practice to first review the AI’s response carefully. Before using it in real work, run it in a test or safe environment to make sure it works as expected and gives correct results.
PMI Team Member
