When you are in cardiac arrest, every lost minute decreases your chance of survival by 10%, so any chance of saving those critical minutes in response to a possible cardiac arrest victim is extremely important. The AI in this case analyzes words and non-verbal sounds that indicate someone is in cardiac arrest. It is able to do that after it has trained itself to spot warning signs by analyzing a massive collection of emergency call recordings over a period of time. In one study, it was found that this startup's AI was able to detect cardiac arrest with a 95% accuracy, compared to 73% for Copenhagen’s human dispatchers.
If AI can save lives like this – apparently 22% more accurately than humans can – can it save or improve project delivery? Probably. Is it cost effective or cost prohibitive? I'm not sure... likely cost prohibitive at the moment but I'm betting that will change in the not too distant future. How would you use it to improve project delivery in your organization? Well, this is where probably gets sort of out there and it's more like playing a project management “wishing” game than dwelling in the real world at the moment.
So, let's play. Let's look at five ways I've been considering - assuming project and PMO budget isn't necessarily an issue...
If AI can listen to 911 dispatch calls and recognize signs of cardiac arrest in potential call in victims, then surely it can be similarly trained to listen in on project status calls and detect confident and concerned tones in both the customer team and the project delivery team. Lots of comments happen during those critical weekly calls – certainly an AI tool could learn to detect concerned responses and probably even detect honest answers and when a respondent was holding back. Change it into a video conference and AI can detect facial expressions and non-verbal communications. That's when it can really provide some key benefits – though it would feel a little big-brotherish. Like the Planet P Project song... “they've got the cutest little cameras hanging everywhere... after awhile you just forget they're there...”
AI could listen to team meetings and collect all of the status update information and – assuming everything is configured correctly – update things like the project schedule, issues list, risk ledger, change orders and other key project information with latest and greatest progress and status info. From that, it could probably do a great job of creating a standard agenda and status report to use for the regular weekly status call mentioned above. I'm sure this sounds as if it were going to be replacing the human project manager, but that should never be the case, no matter how advanced AI becomes.
AI should be able to eavesdrop on status calls with the client and other key calls and emails and sort out discussions that seem to have requests that are out of scope. Having the AI “learn” everything about the project – including the detailed requirements as documented during early planning - to understand the current scope at any point in time would allow it to recognize discussions that were leaning towards work that would actually be out of scope for the project. It could then automatically create suggested draft change orders for the project manager and team to review and modify before presenting to the project client. With the input of proper costing and historical estimating and actuals for similar work, it could probably do a very good job of coming up with an initial estimate as well. Learning over time would only serve to tighten that estimate.
Anything we can have AI track on the project likely ensures more accuracy, more potential for help with solutions as the AI learns the issue tracking and reporting processes as well as understanding the goals and mission of the project, the milestones, and the technology in use. Think about it, the possibilities are endless. Would all of this change project management as we know it? Definitely. Would project managers become expendable? I don't think so and I certainly hope not. It will be a scary “Rollerball” type society if we allow too much AI takeover.
Just as AI could assist in issue tracking, it could be invaluable for identifying, tracking, assessing and managing the whole risk process throughout the project. AI learns and grows and as it takes in more data on one project and all projects in a company's portfolio, the data and history that AI could develop and retain on all projects could be invaluable to a company's success.
Summary / Call For Input
AI. Project management. Are we at a point where we are ready to combine the two? I doubt it. I think the overall cost is too prohibitive at this point – at least to a point where we can price it in as a genuine project expense and justify it to a client. If our organization is utilizing artificial intelligence currently and we want to incorporate it into the PM process then that's one option. But to seek AI at this stage and justify the cost of the technology and the personnel to develop, modify and incorporate through training and ongoing maintenance and oversight... likely that would be $500,000 - $1 million per year for awhile, if not more.
It's probably going to be hard or impossible to justify unless we are looking at something very high end, very cutting edge and very evolving – possibly like NASA's space program or any one of a number of medical fields, organizations and pharmaceutical suppliers. They may all have the money or the means to get it funded. Your average professional services delivery organization will not have that kind of money and time and resources in 2020.
Readers – what's your take? It's unlikely that many, if any, of us have actually touched AI technology so far in our technical careers or project management initiatives. If you have, I'd like to hear from you - please share any experiences you've had with AI and your thoughts on it's uses now and in the not too distant future helping organizations really win on projects... not just “do a little better.” Are there applications for it in PM? Share your thoughts and opinions.