PMI Infinity launches today! Thoughts on the new AI-powered tool?

Great to hear, Laura. Is PMI Infinity using the GPT-4 engine? I'm not a paid ChatGPT subscriber but I was able to use it so I'm assuming that either it is a 3.5 engine or somehow is not requiring users to have paid for access to the v4 model.

Kiron

Thanks Laura! I'm assuming once the current beta on GPT 4 has been completed, it would require users to have a paid subscription? That's a pity as it'd be great to see the improvement in predictions and responses with the more capable engine.

Kiron

Laura Schofield: This is a great tool and I am sure it will add value to many. I have a few questions though:

1) I did pose a question but didn't find the option to provide more detailed input via a pop-up window. Where can I find this?

2) When you move to GPT4, as paid PMI members, do we need to pay for the ChatGBT subscription to use the Infinity Tool?

I did ask PMI Infinity the following simple question: If the risk will occur for sure, what is the probability of occurrence? The answer was pretty good. I will try to ask more sophisticated questions later.

I tried it, but currently the training data appears to greatly limit the output.

I asked several questions ranging from technical subjects such as importance of money flow, how Monte Carlo simulation works, and what are derived requirements, as well as general questions like why certification is beneficial. Only 1-3 PMI articles were cited per response with an average of about 2.5.

Some parts of responses were interesting, others showed significant bias. Responses citing only 1 article went well astray, presumably to make a longer answer out of what was available.

I really like the fact that it cites the sources used, but that also reveals the very limited dataset. When the typical articles used are less than 500 words, and only 1-3 articles match the subject, it is a very limited set of training data to work with.

You should probably clarify on the architecture behind PMI Infinity. Some answers are "powered by OpenAI's GPT" while others are "powered by PMI proprietary content". I tried a line of questions that led to this peak confusion answer:

Me: Are you PMI Infinity?

Answer: No, I am not PMI Infinity. I am an AI language model assistant developed by OpenAI. PMI Infinity is a platform designed to support project managers in their professional development by providing them with a reliable and trustworthy knowledge base built using PMI's expansive catalog of proprietary project management content. My purpose is to assist users with their queries and provide helpful responses based on the available resources and knowledge.

Laura,

There are 7,488 articles in PMI.org’s article index since 1970 (taken from 9 different sources, of which ProjectManagement.com is not considered a source).

In ProjectManagement.com’s article index, we have over 10,000 articles since 2000.

From a run-rate perspective, ProjectManagement.com generated 340 articles last year, whereas the PMI.org domain shows only 47 articles.

Although “PMI Infinity” couldn’t tell me how many “PMI Proprietary” articles were scanned for the project, it appears that ProjectManagement.com assets were not considered for consumption.

If this is true, then the amount of knowledge “left on the table” is extraordinary.

Can you add further thoughts regarding the domains of content consumed under the PMI organizational umbrella?

I recently utilized PMI-Infinity to develop a project charter for a past project as used case, incorporating elements like background information, challenges, proposal, stakeholder details, and the business case in prompt. This approach, guided by industry best practices, resulted in a well-structured project charter.

The integration of PMI-Infinity in this process significantly enhanced efficiency, accuracy, and collaboration. The AI-driven method not only accelerated the charter's creation but also improved its overall quality. This ensured that the charter effectively guided the project, laying a solid foundation for subsequent phases like requirements analysis, detailed scope statement, and Work Breakdown Structure (WBS) development, all critical for project success.

I foresee PMI-Infinity's application extends to various project management areas, including scope, time, budget, resource, and risk management, support decision-making, problem-solving, and enhance communication and collaboration, while automating routine tasks.

I have prompted PMI Infinity with a classical tricky question about the project life-cycle phases and the answer was completely unexpected:
Prompt:
"Are the project life-cycle phases: initiation, planning, execution, monitoring and control, closure"?
Answer:
"Yes, the project life cycle phases are initiation, planning, execution, monitoring and control, and closure. These phases are a collection of logical groupings of related activities that usually culminate in a deliverable...." (!)

So, as the answer begins with Yes, it reveals the great confusion made between project life-cycle phases and the five process groups (raised in my question). Thus, there might be multiple wrong issued answers from different referenced Sources, as follows.
In this way, the interesting thing is the first Source reference, provided by PMI Infinity, by the end of the answer content ("Project Management and ISO 9001 -- An Integrative Approach Through Process Management Learning Library").
If someone searches for it, he/she is directed to the https://www.pmi.org/learning/library/integrative-approach-process-management-iso-7836 address, where it is posted the respective article dated in November 2001.
Inside it, it is revealed the first answer given in its first day of availability by PMI Infinity (DE means Dürr Environmental, Inc.) on the same subject:

"The phases of the DE Project Life Cycle include the following activities:

• Pre-Contract phase: Includes all sales and proposal activities as well as order booking, (a critical activity in a business environment as diversified as DE's).

• Planning phase: Includes activities such as scheduling, budgeting, process engineering, etc.

• Execution phase: Includes detailed engineering, manufacturing (management), field activities, start-up, etc.

• Closeout phase: Includes system optimization, testing, etc.

• Warranty phase: Includes all service activities addressing contractual warranty requirements".

Thus, PMI Infinity seems to be trained without inspecting and controlling the different reference Sources and frameworks of PMI Learning Library, no matter which content they reveal.

Actually I tried it for the calculations of days and it needs enhancements as follows:


it calculates the period for 1095 days after 1st of Jan 2022 as 1000 days which is misleading.
I asked "A delayed project with 30 days, the original duration was 1095 days and started on 1st of January 2022, on the 1st of January 2023 the project has an eot with 1 extra year, what is the updated completion date based on that?"
the Answer was "Based on the information provided, the original duration of the project was 1095 days, and the project started on January 1, 2022. On January 1, 2023, the project was granted an extension of time (EOT) of 1 year, which would add 365 days to the original duration. Therefore, the updated completion date would be calculated as follows:
Original completion date = January 1, 2022 + 1095 days = September 28, 2024 Updated completion date = September 28, 2024 + 365 days = September 28, 2025
Therefore, the updated completion date for the project would be September 28, 2025."
I asked "what about the 30 days of delay?"
the answer was "Based on the information provided, the project has a delay of 30 days. To calculate the updated completion date with the delay taken into account, we would need to add 30 days to the original completion date. Therefore, the updated completion date would be calculated as follows:
Original completion date = January 1, 2022 + 1095 days = September 28, 2024 Updated completion date = September 28, 2024 + 30 days = October 28, 2024
Therefore, the updated completion date for the project with the 30-day delay would be October 28, 2024."

I asked again "but is 1095 days after January 1st, 2022 is September 28, 2024?"

the answer was "Yes, that is correct. If you start with January 1, 2022 and add 1095 days, you will arrive at September 28, 2024. Therefore, 1095 days after January 1st, 2022 is September 28, 2024."
I asked again "what if it was 1000 days instead of 1095 starting from January 1, 2022"

the answer was "If the original duration of the project was 1000 days instead of 1095 days, the completion date would be calculated as follows:
Original completion date = January 1, 2022 + 1000 days = September 27, 2024
If the project was granted an extension of time (EOT) of 1 year on January 1, 2023, the updated completion date would be calculated as follows:
Updated completion date = September 27, 2024 + 365 days = September 27, 2025
Therefore, the updated completion date for the project would be September 27, 2025."