Talk to fellow users of Intel Analyzer tools (Intel VTune™ Profiler, Intel Advisor)
5024 Discussions

Does Agile Work for AI ML and automation projects?


Does Agile Work for AI ML and automation projects?

Agile is failing hard when it comes to Automation AI. Up to 80% of AIML projects are failing, even in tech-forward firms that have implemented Agile in its most advanced and deepest forms. So, does Agile work for AI, ML, automation projects? And if not why not and what are the alternatives?

Millions of people worldwide and the vast majority of organizations implement Agile methodologies and Agile approaches on a daily basis to manage their various development projects. And for good reason: since the early 2000s, Agile has become a best practice for tech teams looking to accelerate development and delivery while reducing the risk that comes from rapidly changing environments, evolving user requirements, and uncertain customer needs. That’s what Agile is all about: focusing on the aspects of managing complex application development projects or running technology teams. Yet, Agile is failing hard when it comes to automation AI.

What are the 4 principles of agile?  Are they Relevant for AI?

Software projects never really have an end. New features can always be added, and new functionality delivered. 

As such, developers and technology teams didn’t want to wait many months or even years for a software project to wind its way through design, development, testing, and deployment. Stemming from this frustration the Agile approach was created to focus on tight, short iterations (two-week sprints) with a goal of rapidly producing a deliverable to meet immediate needs of the business owner. Teams can then continuously iterate as requirements and needs become more refined. 

The four core values of Agile software development as stated by the Agile Manifesto are:

  1. Individuals and interactions over processes and tools
  2. Working software over comprehensive documentation
  3. Customer collaboration over contract negotiation
  4. Responding to change over following a plan


Agile methodologies have forever changed the way organizations develop and release software functionality in a world where the pace of change continues to accelerate. But does Agile work for AI, Machine learning and Automation? If the examples of AI project failure are a case in point, the answer is no. 

The reason is that Agile methodologies are focused on functionality and user needs, but AI automation systems are entirely dependent on data availability and data quality. Even with the most advanced and rapid Agile processes, an approach that isn’t data-forward and data-centric will constantly lose out to the realities of data in the real world. Does that mean we can’t use Agile at all? Or can we adapt Agile methods to suit our AI and data-centric needs?


How do you develop a AIML automation project?

So, if using Agile for automation doesn’t work on its own, then what can the architects, project managers, owners, and leaders use when managing the AI automation and ML data projects? The answer is to leverage a data-focused approach that is also iterative and agile.  The best practice approach used by MO autonomous operations lead is the cognitive project management methodology.


One challenge with implementing cognitive method is that introducing something new and foreign created internal resistance. So the key is to provide a trusted approach that simultaneously delivers the expected results to the organization, using familiar and approachable terminology and concepts, and provides an approach for continued iterative development at the lowest risk possible. Because, at the end of the day, successfully running and managing an automation project with an appropriate AIML project plan should be everyone’s goal.

Some of these AI efforts are focused on natural language processing (NLP) and conversation systems, while others are focused on recognition or autonomous systems. Other efforts are internally-focused advanced predictive analytics, pattern and anomaly detection management, or highly personalized “hyperpersonalization” systems that aim to optimize an offering for each individual. 

Yet, we struggle to make AI Automation systems work by using dated processes, methodologies, and approaches that are application development-centric or aren’t rooted in the data-centric nature of AI. These traditional approaches face significant challenges when organizations attempt to apply them to the unique lifecycle requirements of automation and AIML projects. This is because what drives automation, AI and ML projects is not programmatic code, but rather the data from which learning must be derived. Therefore, what is needed is a project management methodology that takes into account the various data-centric needs of automation while also keeping in mind the application-focused uses of the models and other artifacts produced during an AI lifecycle, and also leverages best-practices experience around agile and iterative forms of project management. 


Dealing with the High Failure Rate of AI Projects

Despite decades of experience running major technology projects and with millions of dollars invested in advanced technology, we are experiencing a high rate of failure for AIML and Automation projects. With all the great technology we have and with so many highly trained developers and data professionals, why do we have such a high rate of failure?

      With our experience, we have identified 10 major reasons for why AIML Projects fail:  due to a wide range of reasons, such as a lack of availability and accessibility of data to a misalignment of the specific business problem to the attempted AO project solution.

  1. Applying application development approaches to data-centric AO
  2. Lack of sufficient quantity of data
  3. Lack of sufficient quality of data
  4. ROI Misalignment of AO solution to problem
  5. Lack of planning for continued AO, model, data iteration and lifecycle
  6. Applying proof of concept thinking to Org pilots
  7. Vendor misalignment on promise vs. reality
  8. Redundant processes in applications
  9. Redundant tools and technologies with many stakeholders


What we realized is that great technology doesn’t solve any of the above problems. Neither does simply throwing more people at the problem. The solution to the problems listed above is a robust, iterative method by which to reliably run Automation projects with a high degree of success. 

0 Kudos
0 Replies