Cognitive PM methodology: The conceptual nut bolts
Thinking machines are no fluke or fiction but could be formidable human ally that, if used ethically, can help augment human living experience. Thinking machines with cognitive capabilities can provide value-added support across various fields including (but not limited to) medicine, engineering, science, business and human day-to-day living. However, effectiveness of thinking machines lies in their ability to mimic some form of human cognition processes and be able to learn, decide and act independently.
While artificial intelligence (AI) technologies such as neural networks, machine learning (ML), deep learning (form of ML), and natural language processing (NLP) have evolved over past decade or so, yet the cognitive capabilities of thinking machines are still quite limited. What it means is that concentrated efforts are needed to build technologies that have the cognitive capabilities including understanding and comprehension, natural language processing and data mining. To this end, project management can play a very important role.
Project management (PM) is a knowledgebase that provides guidelines on a structured way of doing things through an array of methodologies, systems of best practices, and standards. So far, such guidelines in the form of traditional, agile (e.g. Extreme programming), and hybrid traditional-agile (tragile) methodologies and systems of best practices have proved to be useful for delivering projects of all types (from simple to complex) and sizes. But at the same time, there is always room for methodological innovations. In particular, the complex nature of thinking machine technology projects underlines the need for the development of new PM methodologies. One such methodology is the Cognitive PM methodology.
We define Cognitive PM methodology (CogPM) as a knowledge architecture (composed of frameworks / models of project implementation, activities, practices, learning and skills) that guides delivery of cognitive technology projects through a multi-disciplinary fusion of neurology, psychology, computing, robotics and management science.
Having defined the boundaries of CogPM methodology, the question then is what are the conceptual nut bolts of the methodology? What the key concepts, elements, activities, and techniques are that encapsulate the make-up of methodology?
To answer the questions, below we propose some conceptual building blocks of the CogPM methodology. The proposed CogPM methodology revolves around two key aspects: (a) what needs to be done (activities), and (b) what is needed to get it done (skill competencies).
The reason for choosing skills over knowledge for the second aspect is because skills refer to the application of knowledge which, in hindsight, is a more tangible competence. Moreso, since knowledge is a precursor to skills, so focusing on skills is beneficial as it covers knowledge anyway. People having skills (rather than just the knowledge) are more likely to be capable of dealing with the dynamics of project work in the time of, both, calm and crisis.
As a starter, here we define and discuss three building blocks of CogPM methodology as follows. The idea is to initiate a thought-building process on the role of PM in cognitive technology projects (hereinafter referred to as Cog-projects).
1. Critical activity groups (CAGs): The CAGs are the groups of activities needed to deliver a CogPM methodology-based project. We have proposed five CAGs which also correspond to existing nomenclature for processes used in traditional and agile PM methodologies and frameworks. Keeping CAGs aligned to the existing knowledge base will help in mapping of CogPM methodology conceptual building blocks with other methodological approaches, and thus facilitate cross-transfer and cross-fertilization of knowledge as well.
2. Cognitive Project Lifecycle (Cog-PLC): Since we are proposing a methodology, so it is pertinent to define project phases. For this purpose, we draw upon the existing ‘Cognitive Project Management for AI Methodology CPMAI)’ (Cognilytica, n.d.) to define a four-phase Cog-PLC.
3. Skills competency areas (SCAs): Skills competency areas (SCAs) are the grouping of skills needed to successfully deliver Cog-projects. For defining SCAs, we draw upon the existing knowledge on skill categories defined by the global skills and competency framework for a digital world (SFIA, 2018). Using skills categorization of SFIA framework as a reference model, we propose six SCAs for CogPM methodology.
The conceptual nut bolts or building blocks of Cog-PM methodology
1. Critical activity groups (CAGs): We define five CAGs that broadly cover all the activities that should/must be done to complete a Cog-project. The five CAGs are well aligned to traditional and agile approaches of project delivery defined in the book by Haugan, (2011). We briefly explain each of the CAGs below.
It is pertinent to note that CAGs should not be construed as project phases as CAGs are simply the grouping of activities that can be used/repeated as many times as possible in a project lifecycle. A phase in a Cog-project can use any number of CAGs (one, two, three, four or all five) as required based on the need of the phase. CAGs are meant to simplify and help clearly see all the useful activities in a particular basket of activities (e.g. realizing) which will allow choosing the ones needed for the project work.
- Envisioning: This group includes the activities that help form the vision of the cognitive system that needs to be developed. It will be firming-up the idea, need and consideration of various neurology, psychology, computing and robotics issues.
- Conceiving: This group includes all the activities that include mapping of human brain cognition power, cognition modelling, software/hardware planning, creating design blueprints of the cognitive systems, and PM planning etc.
- Realizing: This group includes the activities needed to build the prototype, and the final build of the cognitive system. It includes procurements, stakeholder management, risk, quality assurance, technologies development and testing.
- Governing: This group includes activities that must be executed for project governance such as management, control, leadership, mentoring and team developments. As Cog-projects are complex, therefore, the people will be expected to possess a variety of self-governance and self-leadership skills to handle project governance effectively.
- Concluding: This group includes activities to review the quality of system developed, completing handover tasks, recording experts’ opinions on the development process, providing debrief to the client on the risks and shortcoming, forecasting technological evolution based on the developmental experience, and writing a review on team’s capabilities and experience on the project delivery.
2. Cognitive project lifecycle (Cog-PLC): The proposed Cog-PLC is aligned with PLCs used for AI or ML projects. Majority of AI-focused PLCs have a similar configuration. Therefore, to build knowledge on the existing understanding, we use the conceptualization of PLC defined in CPMAI™ (Cognilytica, n.d.) and Richey (2019) to propose a Cog-PLC. Given the complex nature of Cog-projects, every Cog-PLC phase includes multiple sub-phases as proposed below.
- The project design (Phase 1): This phase involves multiple sub-phases including (but not limited to) (a) relevant business understanding, (b) data identification, (c) cognitive requirements development, and (d) cognitive model engineering.
- Data preparation (Phase 2): This phase includes various sub-phases such as (but not limited to): (a) data cleaning, (b) feature engineering, and (c) data labelling and augmentation.
- Model Fitting (Phase 3): This phase includes a variety of sub-phases including (but not limited to): (a) modelling assumption and algorithm selection, (b) model development and iteration, and (c) model evaluation and validation.
- Model operationalization (Phase 4): This phase also involves a number of sub-phases including (but not limited to): (a) model deployment, (b) Post-processing and Visualization, and (b) real-world model monitoring.
3. Skills competency areas (SCAs): For the Cog-PM methodology, we propose a shift from knowledge-based to skills-focused approach, because it will help project staff having tangible capabilities to work on Cog-projects. Such a shift is also expected to help the delivery of PM education and training to produce a workforce that has tangible PM skills. Using ‘The global skills and competency framework for a digital world’ (SFIA, 2018) as a reference model, below we list six SCAs with corresponding skills as part of Cog-PM methodology.
- Business, technology foresight and strategy skills: The skills covered under this category include (but not limited to): innovation and creativity, visionary leadership, cognitive technologies foresight, data analytics and management, business strategies development, entrepreneurship, and resource sourcing.
- Change management-focused Portfolio, Program and Project management skills: This category includes skills such as (but not limited to): 3 PM (portfolio, program and project management), new organization design and set-up, requirements engineering, organizational capability development, organizational knowledge management, and change implementation and leadership.
- Development and implementation skills: This category includes skills such as (but not limited to): technology architecture design and implementation, UX design and implementation, data modelling and algorithm design, software/hardware development and implementation, system integration, and system installation
- System operation related skills: This category includes skills such as (but not limited to): service design, service delivery and management, ongoing service support, problem troubleshooting, incident management, and system security management.
- Quality assurance and risk management skills: In this category skills include (but not limited to): quality and risk benchmarking, quality conformance assessment, digital forensics, risk identification and prioritization, crisis analytics, issues management, and safety assurance.
- Relationship and engagement skills: This category includes skills such as (but not limited to): vendor relationship and satisfaction management, experts’ engagement, stakeholder engagement and satisfaction management, User experience management, communication management, and customer care management.
The intelligent machines that can think, learn and adapt have been a matter of immense interest among thinkers, technologists, futurists and entrepreneurs, just to mention a few. The scope of support to economic and social activities that such machines can provide is simply un-quantifiable. It is, therefore, imperative to build thinking machines that mimic some level of human cognition to help solve problems and support human activities.
Given such needs, we have proposed a new methodology i.e. Cog-PM that is aimed at contributing to the innovation development process of Thinking machines with cognitive capabilities. The proposed methodology takes a skills-focused approach to management of Cog-projects. As part of the methodology, we have proposed three conceptual building blocks encompassing critical activities and the skills needed to complete such activities for successful delivery of Cog-projects. Needless to mention that the proposed building blocks are aimed at initiating a discourse towards forming a full-scale conceptual make-up of Cog-PM methodology, and hence should not be taken as exhaustive or final.
Cognilytica, (n.d.). CPMAI: Methodology for Implementing AI Projects Successfully, https://www.cognilytica.com/cpmai-methodology/
Haugan, G. T. (2011). Project management fundamentals: key concepts and methodology. Second edition. Management Concepts, Inc.
SFIA, (2018). The global skills and competency framework for a digital world, Version 7, https://sfia-online.org/en/sfia-7
Professor Jiwat Ram
© 2020 Jiwat Ram, All Rights Reserved.