Harnessing the power of AI in Your Business

During my technology education, I was fortunate to participate in a Summer of Statistics Seminar that covered three semesters of graduate-level statistics in just eight weeks. The class met for four hours a day, five days a week, and began with around 360 students. However, the Socratic Method of teaching was employed and students who were unprepared were dismissed for the day. By the end of the course, only 36 students remained. These students became statistical experts. I received an A in the class and the knowledge I gained has provided a solid foundation for my statistical reasoning and application over the years.

One of my proudest applications of statistical reasoning was the Studio Resources Forecasting Tool that I developed while working at Rhythm & Hues Studios, a feature film VFX studio. The Forecasting Tool was software that "data mined" the studio's accounting and work tracking records to generate recommendations for which staff members should be assigned to specific tasks for a given client job with a given deadline. In modern terms, I created a deep learning model of the studio that could predict the most suitable staff assignments for a given client job and deadline. The forecast included:

1. Only considering active and available staff in recommendations,
2. Taking into account physical dependencies, such as computer equipment, work spaces, active and backup disk storage requirements, and shared computation resources,
3. Managing the client job's critical path, including "intelligent" staff recommendations in the event of staff loss or partial loss,
4. Examining staff skills, competencies, and learning rates at the individual task level, aggregating into a comprehensive department and company capability matrix,
5. Simultaneously tracking multiple client jobs and their associated tasks as they progress through the company,
6. Allowing for the easy creation and use of "what-if" scenarios by studio management for client and investor pitches, active work negotiations, and internal staff disputes.

By the time the Studio Resources Forecasting Tool was a mature tool it had been debugged and battle-tested across nine major release VFX heavy feature films.

A significant factor in the success of this tool was the availability of task tracking and work quality records. In addition to the forecasting software, these records were critical for generating accurate forecasts. While I will not delve into the statistics and reasoning behind why it is important to track the following information in your business, here are some key data points that you should consider in order to take advantage of today's AI technologies in your own business:

1. For every activity that generates revenue for your business, itemize all of the tasks required to complete the job;
2. For each task listed in #1, itemize the task's requirements, including personnel, equipment, time, and any cleanup;
3. For each task listed in #1, place the tasks on a timeline;
   
   1. If the same resource (person, hardware, workspace) is booked for the same time frame:
      
      1. move the task forward in time OR
      2. reassign the specific people, hardware, and/or workspace until there are no longer any time conflicts;
4. Identify the series of dependent tasks that compose the minimum time from client work start to client delivery, or the Critical Path of tasks to delivery;
5. For every resource (person skill, hardware item, workspace), assign a quality rating;
6. For every type of task, assign a difficulty rating;
7. For every task, track the following:
   
   1. who performed the primary work and the time spent,
   2. anyone who provided supporting work and the time spent,
   3. an A-D quality rating of the work performed.

If your business is using any type of task tracking management software, it is likely that concepts like the Critical Path are already in use. However, most task tracking systems are primarily intended to help management oversee their staff. Modern AI and its trained algorithmic nature can provide much more useful and insightful information about a business and its internal operations. In order to take advantage of modern AI, it is essential to have this or similar data about your business.

If you do not have this information already, now is the ideal time to start collecting it. Follow the numbered list above and gather the information described at each step. Use a tracking system of some kind and include both hours and accounting ID numbers in the tracking. It is often easiest to track hours worked in the company accounting system and completed task quality ratings externally, as systems can vary from company to company. The key requirement is the ability to bring both forms of information from past jobs into the forecast (schedule) for new and proposed client jobs. 

The reason that the quality of completed work is included as a metric is because collected ratings on past work enables the business to forecast expected time and quality when that task is performed by a new employee versus an experienced employee. When considering client jobs with a significant volume of separate tasks, having past time and quality ratings for each employee provides an excellent foundation for forecasting. When there are multiple jobs and a correspondingly larger workforce, it is easy to see how this information on an individual employee level can be used to map out a series of dependent task chains, and then maintain them dynamically - which is exactly the type of work that AI excels and human staffed operations management struggles.