Presentation Breakdown from Lloyd Skinner, CEO of Greyfly.ai
My Cliff Notes of his presentation of AI in Project Management.
One of the big mysteries for project managers is how AI is used in real-world project management.
I heard about Greyfly.ai from an AI course I took. When I saw they had posted a video for the Institute of Project Management, I thought the talk would have a few takeaways that would help me identify how AI is being used in project management.
So you don’t have to watch the whole hour presentation, I’ve taken notes and screenshots.
Of course, you can watch the video which is included above.
Introduction
Lloyd Skinner has over 25 years of project management experience and is the CEO of Greyfly.ai. They have built an Intelligent Prediction Platform that uses AI to produce project success.
Which Star Wars character are you?
Lloyd used the slide above to indicate your knowledge level of AI. He said that Ewoks would have used it in a work setting, but they may have played around with it at home or started to make it a part of their work. Those people may be summarizing emails or reports.
Princess Leia is when AI users start to get a little more advanced. These individuals start embedding AI in applications. They fully embrace data-first strategies. Most of these people are early adopters.
From what he has seen, some people have moved to the right side of the spectrum after 2022. Most people are on the far left side of the spectrum and have very little knowledge.
AI is becoming a partner with project managers
There’s a difference between embracing AI at the individual level and embracing AI for organizations. Lloyd says most companies are still very early in adopting the tech. Very few companies have gone beyond the initial experimentation or mobilization stage. Most companies rely on employee experimentation to bring about a shift to AI.
How AI unlocks project success
The above image shows a high-level project life cycle. The business starts at pre-approval or presenting a bid to win the job. After this stage, the company moves onto the delivery stage, where an asset is (hopefully) created. Those assets are “used” or “operated” after they are created.
Lloyd relates that AI is trying to understand the historical experience from that process, take that data, turn it into information, transform it into knowledge, and make it accessible as early as possible in the project life cycle. This knowledge also fuels the prediction engine so we can take the right actions or undertake the right interventions to create a successful project.
He maintains that in the project management realm, there are two different dimensions -
Effectiveness
Data-Driven decision making
Predicting project outcomes
Risk management and lessons learned
Quality assurance
Efficiency
Cost control and budget management
Automation of routine tasks
Resource optimization
Knowledge access
Notice that communication isn’t represented at all in the two categories above. AI needs to mature to be useful, and that’s where project managers assist.
The project manager being the human-in-the-loop
Lloyd commented on this chart and said that in the beginning stages of working with AI, we start by streamlining automation and workflows as we begin integrating it into our work. Chatbots can be used internally as assistants and to fast-track project work.
Robotic Process Automation (RPA) is a low-code workflow that is expanded across multiple systems. Microsoft Copilot is an example of this.
Lloyd mentions that as project managers in an organization, we usually are in the “Insight and Foresight” quadrant — its about prediction.
The far right bottom is the AGI we are all waiting for - the bionic or self-directed application (the thinking machine). But we’re drifting towards “digital twins”. It’s a mirror of the actual project - usually, in construction, they build a digital copy of the building.
You won’t be replaced by AI, you’ll be replaced by somebody with AI skills. — Lloyd Skinner, CEO of Greyfly
If the project is simple, consistent, or decision-tree-based, AI can assume more control over it. A good example is HR tools that review resumes. AI looks at history and can make simple decisions that start removing chunks of work for people and allowing them to focus on different things.
As project managers, you need some technical skills, like ChatGPT or using Gen AI. As PMs, we are likely to be innovators and leaders. It’s suggested that we start influencing and demanding that our organizations use AI.
Ultimately, soft skills and change management skills are the differentiators for PMs. Great intrapersonal skills separate the good PMs from the pack. You can influence management to embrace AI in their company.
The Intelligent Project Prediction case study
The first case study used Greyfly’s Intelligent Product Prediction methodology. The UK-based organization contracted them had about 3000 projects in its portfolio, 100 of which could be live at any one time and were worth hundreds of millions of pounds. This company was overspending by more than 70%. The problem was that it had difficulty delivering projects on budget compared to the baseline. The people supervising the projects didn’t know what to focus on since the reports were biased—the project managers didn’t like to deliver bad news.
The answer was project prediction. Greyfly applied prediction to projects that had not yet started and those that were live. Machine learning recognized patterns between projects and began to predict budgets and outcomes. What used to take teams of people was done very quickly. Greyfly worked to improve data maturity since there’s a direct relationship between data maturity and project success. They defined the prediction accuracy - it showed what the system predicted versus what happened on the project. With fine-tuning, they got a 95% accuracy of a potential project before starting it. This led to tremendous project success.
Data is the most important when dealing with AI, and when Greyfly onboards a customer —they do a data maturity assessment. It benchmarks the quality of data in the client organization. Any time actuals or forecasts were brought to monthly checkups, they would work with the PMO to ensure any data issues were addressed. Greyfly would then work with an optimum mix of project features to increase the probability of project success. If you can do that before the project is committed, the project can have greater value.
The client company uses an API to plan its projects and monthly forecast actuals. It compares these to the predictions and budget spending, which is done at the individual project and portfolio levels.
During the Discovery phase, the company needed to improve its data maturity (data is the heart of any AI project). Discovery can include defining the enterprise data model and the data rules.
Greyfly works out the custom prediction data model for each client. The data prediction model that Greyfly builds from is the base model and is customized. Greyfly built the back end in Azure with a PowerBI front end. The hardest part is the first data load and preparation stage. The data is then run on many machine-learning models, and the best-performing one is chosen. This may change every month since the data scientists are fine-tuning the ML models while maintaining accuracy.
Integrating it into the workflow is key. In the monthly progress report, the project manager may say their estimate is different from the prediction system. The client will use gate reviews to see what is needed to complete the project.
Intelligent Project Lessons case study
AI in project management is all about unlocking knowledge. When lessons learned are captured, they are usually put on a shelf and unused. Accessing records may be difficult, and making sense of them may not be easy since they aren’t in a templated format or maybe ad-hoc.
Greyfly worked for a client that used a product called Lesson Flow to archive the lessons learned from their projects. It contained 5000 lessons across 200 completed projects. To help the client use these lessons learned, Greyfly needed to integrate search parameters and make lessons accessible.
Greyfly weeded out the duplicate lessons and used Gen AI to simplify them so they were much more straightforward to understand. This was also applied to the mitigation lessons since risk mitigation strategies are valuable lessons. Employees could use a chatbot to prompt this database for the appropriate lessons for their project.
You may have a department in your company called “investment control.” These are the people who appraise business cases and approve the viable ones. Greyfly’s AI predictor model is doing quite a bit of work for them since it’s picking out the projects that would be successful. It’s using the data from the lessons learned to pick the projects that will help the company.
Lessons when implementing AI in project management
Executive Sponsorship — The right sponsor needs to be in place. They are at key organizational levels, but they are budget owners and can influence their peers.
Data Quality and Availability—This hasn’t been the case probably in projects you have done in the past, but you will spend 60% of the time in AI projects dealing with data.
Technical complexity—The project manager will face some additional hurdles, but they are not insurmountable with the right skills and capabilities.
Ethics and compliance — Make sure the data is correct and is not being manipulated by internal or external factors.
It works! Data driven insights. — Remember that you are given data driven insights into your project and it probably hasn’t been done before.
People resistance — People will resist change and you as a PM will need to deal with that backlash.
Budget and resource constraints — You should not have a limitless budget and approach it like any other project.
Interventions feedback — How did you handle any roadblocks and intervene to make the project successful?
AI is increasingly powering a new way of PMO, project, and program management
Technical and people barriers
AI enables data-driven decision-making
AI is starting to be proven and enables potential savings
Requires transformation (technical and change) projects
Closing thoughts
There are many lessons here to unpack, but you can see how Greyfly is using AI in its workflows to help clients parse and integrate data.
Once you have figured out a methodology, you put AI to work and extract the data to determine which projects will be successful and how to increase the chance of success on future projects.
Firms like Greyfly have a head start since they have worked with AI for over 5 years.
But with persistence and learning, you could be the Yoda of AI project management in your workplace!
AI-Driven Tools for PMs
Greyfly.ai - The firm this article was based on
AI News PMs Can Use
Artificial Intelligence in Project Management: Disruptions, Risks, Advantages, and Adaptation
Cool ChatGPT Prompt for PMs
10 ChatGPT Prompts for Project Managers (link at the bottom of page)