Change Management Considerations For AI And Data Science Projects

Track the progress, celebrate the successes, and help people adapt

To read the first part of this set of articles, please click on Fundamentals.

Your Change Management strategy and the assignment of the Change Leader and the Change Agent are critical for data science and AI projects.

For managing the change for those initiatives, we will need to track the progress, celebrate the successes, and help people adapt. To reduce resistance to change (which is happening in most AI projects), we can facilitate the right communication to gain a common understanding and make a clear picture of why the organization needs this innovative initiative. Communicating the required changes to have executives/employees buy-in the benefits is important. Remember, in most AI/ML projects, the Culture Change is the biggest challenge, not technology transformation. Having a plan to manage cultural barriers is important for successful adoption.
 

Following the proven guidelines introduced by either the Association of Change Management Professionals (ACMP) or Prosci would be a good change management approach. For this article, I choose ACMP’s standard, and add a number of recommendations for DS/AI/ML projects:

Evaluate Change Impact and Organizational Readiness: Assess the readiness and ability (very important for DS/AI projects), and the capacity for the transformation that the DS/AI project may result. Additionally, the impact the change needs to be assessed. Explain the urgency of change in the executive summary of your business case for DS/AI initiative. Consider the actual drivers for change (for DS/AI projects) in your organization.

Formulate the Change Management Strategy: Develop a high-level approach for change management with all stakeholders of the data science and artificial intelligence projects, including governance, risks, resources, budget, and reporting. This should integrate change management plans and DS/AI milestones into other activities. Adoption strategies must be a part of your overall project strategy, and forming a corporate academy for DS/AI can help support the change strategies, especially for important cultural changes.

Develop the Change Management Plan: Develop a detailed plan for implementing the change strategy that is suitable for transformations like AI or ML. The plan should cover communications, stakeholder engagement, training, risks, and integration with project management. For AI/ML projects, the best approach is interdisciplinary collaboration. Leaders need to be prepared for the change and have at least a basic level of understanding DS/AI/ML to be able to support the workforce. It’s critical for DS/AI projects to anticipate barriers to change. For example, Service Managers of your organization may disagree that computers can know better/more about what your customers need/want (what AI/ML initiatives can recommend). Try to find alignment opportunities that your project has with your organizational culture. Having a plan for barriers in change management can help in having a better communication with the employees and a better understanding of what DS/AI projects are executable. Additionally, in the early stages of the delivery, get feedback from users, observe the usage of the DS/AI result/output, as well as the behavioural changes, to identify and solve problems as you go forward.

Execute the Plan: To achieve the benefits of artificial intelligence or machine learning projects, we need to implement all the action items related to the change plan mentioned above.

Complete the Change Management Effort: Reinforcement is a key in AI/DS projects. Monitor the progress, measure results, and provide support and coaching to help the change sticks. For AI/ML initiatives, be prepared to spend for good reinforcement tactics as much as what you spend for technology. DS/AI based transformations take 2-3 years (by average) therefore keeping momentum is important. Advancing the culture of accountability in the departments that will use the new capability, having role-model leaders, change agents to facilitate adoption, and teams to capture the required corrective actions, as well as giving rewards for alignment with AI are all important for the success of the projects.

Knowing your audience and avoid jargon. Many executives complain that they don’t see the results of Data Science projects. Here is the root cause: The results are not communicated in an appropriate language, suitable for executives.
  

A scorecard that captures and shows project details in real-time for stakeholders helps in managing the project and keeping the alignment between the stakeholders. It should include:
 

Alan Bostakian is a senior consultant and analyst. He has worked TD Bank (Canada), Real Estate Council of Ontario, Government of Ontario (Canada), Global Association of Corporate Universities (UK), CPHR BC (Canadian HR association), and 3 Canadian colleges. His extensive experiences include Change Management, Talent Development, Corporate University Architecture, Training, Coaching, Certification, Project Management, and Research. Alan has a PhD in Business Administration as well as a number of certificates including Project Management Mastery (Stanford), Certified Training & Development Professional (CTDP), Registered Professional Trainer (RPT), Certified Change Agent (CCA), Change Management Specialist (CMS), Data Science (MIT), Executive Data Science (Johns Hopkins) and Internet of Things (MIT,). Connect Alan Bostakian