"Artificial Intelligence" is so 'yesterday.' What about good, old "Human Intelligence"?

    

A few years ago, someone asked me, “What does your company do?” I replied, “Oh, we offer a disruptive solution…” He groaned and said, “If I hear one more company saying they are disruptive…” “Disruptive” was a buzz word that was so overused that it quickly became void of value. Several years later, I fear the phrase “Artificial Intelligence” is quickly headed down the same path.

Don’t get me wrong; I am all for the pursuit of Artificial Intelligence (AI) assisting us – I spent several years back in the mid-1990’s developing AI techniques for CPM planning – those techniques were smart, really smart, yet in 2017, project planning is still largely a human-driven endeavor. AI is a good thing, but let’s have a reality check here.

What I am really pushing back on is the fact that it seems like every software vendor in the project management space is now claiming to use some form of “AI.” If you talk to a data scientist about AI, they will tell you it is nothing more than the next generation of computing technology and, more importantly, that the success of applying AI to a problem is predicated on asking the right questions. Mountains of data, but no idea what to ask of it.

Let’s embrace modern computing power to help make our tools smarter but let’s also not forget the core objective here, and that is to build a realistic project plan leveraging the expertise of our team members. We firmly believe that AI methods can help us build a sounder plan (and with greater speed), but this should not come at the expense of using good old human intelligence.

Where AI will help us…

Combining AI and Human Intelligence…

Accessibility: There is no shortage of data, and when the right questions are asked of the data it becomes not only accessible but useful and valuable.

Enablement: We address the concept of Augmented vs. Artificial Intelligence below. The role of AI is to feed better and more relevant information to the project team to support better decisions during the planning process.

Speed: Creating a plan based on historical as-builts, benchmarks, or standards is simply faster than beginning each new plan with a blank sheet of paper.

Feedback: When you combine AI with the ability for a human to rationalize the feedback, you end up with a plan that has been calibrated based upon your dataset, but also validated by the expertise of the project team.

Quality: A plan that has a basis in history, standards, and can be calibrated by benchmarks begins its life as a project in a much better state.

Closed Loop: As data is used and consumed by projects, human interaction actually teaches the AI engine. This closes the loop and creates a synergistic relationship between AI and the human taking advantage of its power, both learning from one another.

Augmented Intelligence, not Artificial Intelligence

When building a project plan, we are ultimately trying to predict, as best we can, the future. We are trying to objectively determine how long and how much a series of future events (in the form of a project) will take. We do that through the reuse of knowledge - knowledge of previous projects, knowledge of market conditions, knowledge about our ability to build and construct. That knowledge gets fed into a CPM scheduling tool which then calculates the answer based on your inputs. CPM is not a complicated science. However, getting the inputs correct is difficult.

Where I think AI does bring value is offering up knowledge (inputs to CPM) as part of the overall planning process, but this value stops at the actual building of the plan itself. AI suggests that the technology does it all, and that works when there is little to no variability in the expected outcome or end result (e.g., applying robotics to an assembly line is a good application of AI). What augmented intelligence provides is better information with more speed to the planner who can then rationalize and apply it to the plan. Take advantage of AI making suggestions, but don’t pretend that a software solution can spit out a meaningful CPM schedule at the push of a button.

Practical Application of Combined Augmented & Human Intelligence

Consensus-Based Planning

Good schedulers aren’t necessarily good planners. A good plan needs sufficient breadth and depth of input and knowledge to properly reflect reality and hence be realistic. The best project CPM schedules are those that have team member buy-in. We call this consensus-based planning. The problem with this though is that CPM tools don’t lend themselves to such collaborative plan development. Consensus-based planning should facilitate both creation and review of plans.

Plan Creation & Delegated Planning

Rather than a lead planner be responsible for the development of the entire plan, wouldn’t we be better off having the lead planner delegate specific areas of scope to different scope owners and let them build out the details? Those scope owners will better understand what is needed to achieve the plan more than anyone else. Let the lead planner own and control the overarching plan, but let delegated planners help share the some of the load.

Plan Review & Team Member Markup

Today, many projects carry out interactive planning sessions and/or schedule risk reviews which entail getting multiple team members and discipline leads together to ‘review’ the project plan.

These sessions have an inherent flaw – the plans reviewed are CPM-driven, yet the audience could care less or may not understand the science behind CPM (different types of float, precedence logic, leads, lags, the impact of multiple calendars, etc.). The audience does, however, know how to execute a project. So instead of forcing the team down the CPM route, why not just ask them questions such as “Do you believe the duration is achievable?” or “Do you think you can start on this date?” Then, behind the scenes, translate this back into CPM inputs.

Why not take this a step further and enable team members to each give their expert opinion in their own ‘sandbox’ version of the plan and then bring all of these ‘markup layers’ back into a single ‘consensus-based view’ of the project? They could even assess the impact of their contribution.

TMMViewforBlog.png

This approach of consensus-based planning makes it easy for team members to give their input and for lead planners to consolidate all of these inputs into a validated plan.

Calibrate & Validate

At BASIS, we have developed a solution that is a blend of “Augmented Intelligence” with “Human Intelligence.”

The ability to interrogate a Knowledge Library for sub-nets and common risks, as well as benchmark against standard rates and durations when building a plan, is a huge step forward for project planning. It is augmented not artificial intelligence that is driving this within BASIS. You as the planner still have ultimate control and decision as to what makes it into your plan. BASIS makes it easier by allowing you to apply these calibrations in real-time.

Add to this the ability for team members to review and contribute their expert opinion through their own personal markup layer and your plan evolves from being based on a single planner’s opinion to that of true consensus from those who know.

This approach offers the best of both worlds. We are absolutely taking advantage of better and smarter computing techniques by validating a project plan, and yet at the same time, we are making it easier for projects to leverage their ultimate asset - their team members’ expertise.

ConsolidationViewforBlog.png

This approach allows us to identify what areas of the plan differ from our original expectations. Even more, we can determine where the team is tightly aligned with one another, even if their opinion is different than our original estimate. For example, I am more comfortable with feedback from the team when they agree that the originally estimated duration is too aggressive or short. Alternatively, if the team has extreme variation in their opinion of how long an activity takes or when it can start, then I know I need to investigate further. It is an opportunity to have a plan that is truly validated by the team (human intelligence).

Conclusion

So today, when someone asks me “What does your company do?”, I tend to reply along the lines of, “We help organizations build more realistic and achievable project plans through augmented and human intelligence.” I think that statement is truer to what we are accomplishing rather than trying to “baffle with science.” At the end of the day, “Artificial Intelligence” isn’t even a new science; instead, it’s just another buzz word.

Comment on this Blog

Free White Paper Download:

How 'Pre-Planning' Drives Project Success

Pre-planning is the defining of scope and timelines for deliverables. 

Learn how BASIS differentiates between scope and work and helps you define "what we are building" before identifying "how we are building it." 

Preplanning White Paper Resources Image Rev2.png

Subscribe Here!