This chapter focuses on the rollout of a project’s solution into production. We begin by discussing operationalization considerations in software and AI projects. We explore the typical roles and key activities of individuals. We then describe some of the operational procedures prior to and during limited production with the intent of having them completely implemented when reaching full production.
A successful rollout also requires careful and well-planned change management to assure wide adoption and deliver needed levels of service and support. So, we next discuss how we already created a strong change management foundation during the earlier stages previously covered in this book and the additional change management tasks that are particularly relevant to successful rollout of the production solution.
As our solution achieves production status, we should assess how successful the deployed solution is and use what we learn in the assessment to make corrections as needed. We should also assess the entire process used to reach this point and use the knowledge that we gained to improve how we go about identifying and solving problems in the future.
Finally, in addition to summarizing the chapter, we discuss what you should have learned by reading this book. We also call attention to the circular nature of the approach that we covered and its key stages.
Operationalizing the solution
Change management considerations
Assessing project success
Summary
Operationalizing the Solution
At the point that a solution is deemed ready for more widespread use and adoption, we must consider what it will take to make this possible. Many organizations identify and begin to solve operational challenges during their DevOps development cycles or when they deliver a limited production environment prior to moving into full production. As part of such efforts, they identify critical tasks and procedures that need to be put into place to deliver needed levels of service and support.
When we earlier put together the message used to sell the project to senior executives, we estimated staffing costs required to bring the solution into production. As we begin monitoring the incremental development and production efforts, we should uncover more information about who these key individuals are and the tasks that they must perform. We might also need to evaluate additional staffing needed to perform tasks that we had not earlier anticipated. A typical approach used during these evaluations is to identify who is responsible (R), accountable (A), consulted (C), and/or informed (I) for a variety of tasks (often referred to as RACI).
For example, the key activities identified for cloud-based PaaS deployment of software and AI solutions could include day-to-day operations, monitoring the solution, application and platform change management, application release management, performance tuning, and data protection. The people we might focus on in our evaluation include stakeholders and frontline workers, analysts and data scientists, cloud administrators, data administrators, developers, and IT managers.
The activities and people designated in RACI diagrams are influenced by the type of technology infrastructure utilized in the production deployment. For instance, a more extensive list of IT-related activities and staff would be required for deployment of an on-premise solution than that which we denoted for a cloud-based PaaS deployment. We would need to add systems administrators, storage administrators, network administrators, and data center managers to the list of critical staff. Activities such as infrastructure configuration, software patching, hardware and software updates, and data center infrastructure coordination and planning would be added.
RACI diagrams can also be extended to illustrate the delivery of specific levels of service and denote the individuals and their responsibilities in delivering them. Examples of areas of service-level focus include resource availability (and related technical implementations such as high availability and disaster recovery) and infrastructure security (including infrastructure, software, and data security).
Levels of detail provided within these diagrams can also vary. Some organizations choose to build RACI diagrams for each of the major activities that they identify and then break those activities into detailed tasks for which they also provide RACI diagrams.
Our next step is to document the procedures that are to be put into place to perform required tasks. The procedures are defined by considering previous procedures already in place within the organization, knowledge gained in prototyping the solution, and previous experience obtained by scaling other prototypes into full production.
Application deployment and updates
Auditing of devices and infrastructure
Business continuity and disaster recovery
Cloud and/or IT infrastructure policy
Communications access to devices, equipment, and infrastructure software and data
Data exchange policies and standards
Help desk support
Incident management and remediation
Networking and device management, monitoring, and maintenance
Networking and device standards for the organization
Physical access to equipment in data centers and plants
Security policies
The procedures are generally refined and tested during DevOps cycles or during a limited early production phase of deployment. Scalability testing likely also occurs during this time by simulating workloads and situations that will be likely when the solution is in full production.
AI projects require additional governance procedures be put into place as the accuracy of models often changes over time. Retesting of models leveraging more recent data and evaluations of the results of actions that have been taken should occur at regular intervals to determine when the models and actions need to be tweaked.
Change Management Considerations
Innovative software and AI projects, by definition, drive change. Managing change is an important aspect of assuring successful rollout of such projects. Best practices have been documented by numerous people and organizations who have expertise in change management over the past 70 years. Examples of such methodologies that gained a popular following include those defined by Kotter, Lewin, McKinsey, and Prosci.
Gaining support that change is needed to solve the problem
Gaining agreement that the envisioned change will solve the problem
Developing talents and skills needed to build, deploy, manage, and use the solution
Modifying organizational constructs and shifting responsible individuals where needed
Positively reinforcing those who successfully utilize the solution
The Design Thinking workshop is a critical early piece of the change management process. Remember that we gained support that change was needed and formulated a vision of the potential solution during the workshop. During the subsequent building of the prototype and then development of the solution, we began developing skills needed for successful deployment and management.
As we go into production, we must revisit the skills present in our frontline workers and among the management who will use the solution. If there are skills gaps, we will need to implement a training and rollout plan that closes the gaps and monitor success in utilization of the solution. As part of this effort, a popular best practice is to identify those with early expertise and assign them to act as mentors for those less experienced during the rollout.
The skills that are evaluated for frontline workers and management likely include business skills as well as technology utilization skills. In our supply chain optimization example, important business skills required might include those related to inventory management, supply chain planning, demand planning, transportation planning, order management, manufacturing, contracting, and/or financial management.
During the rollout, one of our goals is to continue to build solution utilization momentum. Senior management can help drive positive momentum by offering meaningful rewards and recognition to those successfully applying the solution. Those who resist change should be provided with additional guidance and training, and it should be made clear that they need to adapt and utilize the solution if they are to be viewed as a positive contributor to the business.
You probably noticed that we illustrated many of the stages by using arrows that point in both directions. As many of these projects tend to evolve over time and roadblocks to success can emerge, there is sometimes a need to revisit earlier stages.
We believe that initiating change management steps as early as possible within the stages will result in more well-thought-out steps and greater maturation over time. As a result, the steps will have a greater positive impact on the organization.
Some experts suggest developing a communications plan during the production rollout stage. We believe that it is important to have ongoing communications procedures defined throughout each of the stages to convey changing project plans, project status, and solution focus. During production rollout, the communications plan should be fine-tuned with focus on the arrival of the solution for general use as well as its impact upon the broader organization.
We can utilize the earlier identified sponsors to assist in our communications to frontline workers during production rollout. We might also enlist additional key managers to support our communications efforts and help broaden the solution adoption.
No doubt, there will be some resistance to the changes that occur. We should have a plan ready to address such resistance, whether the resistance comes from some in management or frontline workers. Such resistance to adoption is sometimes uncovered as we assess the project’s success.
Assessing Project Success
As the project solution rolls out, it should be continually assessed as to whether it is addressing the business problem that it was meant to solve. Doing so is important if you hope to gain momentum leading to widespread adoption of the solution beyond a core group of believers.
Geoffrey Moore described the challenges in gaining adoption of highly innovative solutions in his book Crossing the Chasm. It is interesting to note that in 1991, when he wrote the book, Moore identified AI as one such technology that had failed to gain wide adoption and was seemingly stalled. As you probably realize, the status of AI adoption changed significantly over the past decade when languages, tools, and frameworks emerged aiding its development and usage. The number of trained and skilled individuals capable of using such tools also exploded, and they were enabled by the widespread availability of cloud-based resources and compute power.
Moore pointed out that a technology adoption life cycle can be defined in discrete phases. He described people and organizations leading adoption within these phases as innovators, early adopters, an early majority, a late majority, and laggards. Peak adoption occurs during the early and late majority phases.
A goal of most software and AI projects is to have the production solution gain usage beyond the groups of innovators and early adopters within the organization. A fair and balanced assessment that looks at the state of the project and how it has evolved along the way should be undertaken.
Did the solution deliverables match business requirements and solve the defined problem? If not, why not?
Is adoption of the solution by frontline workers and management growing and is the predicted business value being delivered?
Were project milestones met on time? If not, why not?
Did prototype development transition smoothly into production development and deployment? If not, why not?
Was the original budget for the project accurate? If not, why not?
Were skills found wanting, developed along the way, or are they still in short supply (with internal organizations and/or partners constrained in providing the right people)?
Were quality assessment practices put into place and did they improve the solution? How?
Were technical operations and change management programs adequately planned for and successfully executed?
It should be possible to claim success if the answers to these previous questions are of a positive nature.
If there are many issues, and especially if the business benefits are unclear or not as promised, you have more work to do on the solution. You cannot expect to see widespread adoption of a badly flawed solution or because of a poor rollout. If changes are determined to be necessary, it is time to have candid discussions with sponsors and frontline workers about the shortcomings that they see. The solution’s design, development, and/or rollout plan should be changed appropriately.
When you do find responses of a positive nature during the assessment, be sure to also gather endorsements from key executives, stakeholders, and frontline workers, data regarding solution adoption related to the success, and the ongoing measurable business impact attributable to the solution. All this information should be included in reports and presentations regarding the results of the project and should prove valuable in convincing skeptics within your organization.
Some might argue that software and AI projects are never complete as continual improvements will occur and question whether there will ever be a right time to present this information. We believe that when adoption reaches a point where stakeholders and frontline workers attest that the solution is delivering the significant business results that were hoped for, that is a great time to present this content.
Summarize the business results attributable to the problem’s solution and the early feedback obtained.
Provide an actual timeline of how the project progressed and came into production.
Describe challenges that occurred during development and rollout and how they were overcome.
Describe the important lessons learned.
Describe how the scope of the solution might have been limited (especially if there are deferred modification requests, next steps identified in responding to these requests, and additional funding needed).
You can use this as an opportunity to point out that a key to assuring success of the project was the Design Thinking workshop at the beginning. You might gain new sponsors for future Design Thinking workshops and ensure that it becomes part of the problem-solving methodology present in your organization moving forward.
Summary
Operationalizing the solution by defining key activities, steps, and procedures for people in various roles
Considering the role of change management during the rollout stage with an added emphasis on reaching a broader community
Assessing the success of the project using a defined set of questions, acting when needed to improve the solution, and using the responses to report results to senior executives and sponsors
If there was a bad miss in what was delivered in the project, a Design Thinking workshop might be held at this point to revisit the problem definition and solution. If the project proves successful, this is a great time to make sure that the workshop becomes part of the organization’s problem identification and solution development methodology going forward.
What Design Thinking is (including its history and common methodologies)
How to prepare for a Design Thinking workshop
Problem definition methods and tools used in the Design Thinking workshop
Solution definition methods and tools used in the Design Thinking workshop
How prototype creation is driven using information gathered in the workshop
How to move from prototype development to production development
How to operationalize the production solution and claiming success
Our hope is that this book provided you with guidance on how to identify problems in your organization and determine potential solutions. We also hope that it provided some practical thoughts on how to make these solutions real as you move beyond the Design Thinking workshop.
As a final thought, we would like to share this. Though software and AI proponents sometimes view the technology as something that only needs to be implemented to prove its value, it is far more optimal for an organization to understand the business problems present that need to be solved before starting any development efforts.
It is our hope that you now agree and can see the tremendous value in adopting a Design Thinking approach as part of your project development life cycle.