Building complete and responsible AI systems: the importance of the purple unicorn role
Enterprises struggle to build complete AI systems. Product managers claim that IT doesn’t know business. Tech leads claim the business doesn’t know technology. Missing is the purple role in between.
Many enterprises struggle to build complete and responsible AI systems. At a high level, you have product managers moving forward with AI use cases and claim IT doesn’t know the business. Fair enough- that is true in many cases.
But on the other side of the coin, IT moves forward on building the data science technology stack, and claim the product manager doesn’t understand the technology. This is also true in many cases.
So what’s missing, when each side blames the other as the bottleneck for building complete and responsible AI systems at enterprise scale? I opine the purple role between business and IT is the missing link.
I’ve been on all 3 roles in my various careers and industry verticals as a data professional:
Business side as a SME on labeling corpuses in the early 2010s for AI psychiatric classifier
Startup side as wearing all hats to build AI prototypes in the medical and real estate verticals and pitch to clients
Technology side as the tech lead for data science stack and implementation of AI governance for AI projects
Because of these various experiences, perspectives and viewpoints, I can see how end to end AI solutions can be held up due to bottlenecks. Small startups are easy- you can wear all of the hats, as prototyping and pitching are the main activities to stay alive and keep your funding runway intact. At large enterprises, you have many different groups responsible for each of the hats you wore in the startup. You can’t do it all by yourself anymore. You need to not only collaborate with each of the AI stakeholders at the organization, you need to have someone orchestrate all of those stakeholders.
A main integration point is the connection between business and IT. I opine you need that purple role, between business and IT, to move forward on building complete AI systems. This purple role also helps to make these complete AI systems responsible by integrating AI governance that usually comes from the data organization, where their mandate is to govern AI systems at the Enterprise.
You might say that this is the role of an Enterprise Architect (EA), but even if an organization had one, their expertise in AI is usually diluted by the other more pressing matters at the organization with CIO, CTO, CDO, and CEO all bringing forth their technology initiatives and struggles to the EA, unless they are an AI- first company.
So if you don’t have an EA with a special focus on AI at the enterprise, then this purple role connecting all of the stakeholders together to build complete and responsible AI systems is critical. This liaison between the AI product manager in business and the AI/ML technical lead in IT is a critical role indeed. If the AI product manager and AI/ML technical lead do not connect and align, then the product manager will guide the building of local AI applications in a business silo, while the technical lead will build an enterprise AI/ML platform that nobody in business will use.
Without a connection between the product manager and the technical lead, your organization’s pursuit of building complete and responsible AI system will be difficult to attain, if they ever get there.
So who is qualified to be the purple person between the AI product manager and the AI/ML technical lead? The answer- this purple person is a unicorn, where they understand both the AI product manager role and the AI/ML technical lead role. In addition, this purple unicorn needs to understand AI governance and its necessity in building complete, enterprise AI systems in a responsible manner.
If you don’t have an EA with a focus on AI, then look for this purple unicorn to help your organization build complete and responsible AI systems at enterprise scale.
If you're looking for support from me, here are a few options:
Enterprise Data Science Consultancy: With my consult team comprised of a Senior Data Scientist, Senior ML Engineer, Senior Data Engineer, and Senior Cloud Engineer, we will help you architect and build your Enterprise Data Science platform, and transfer knowledge to your IT team to maintain and optimize it. We will also overlay an MLOps framework to manage the AI solutions you build on this platform. If you don’t have an MLOps team, we will help you build one. Please get in touch about this consultancy here
Coaching and Mentorship: I offer coaching and mentorship; book a coaching session here
This article is featured in a data newsletter: https://www.linkedin.com/pulse/july-data-news-castordoc-1e
Tell your Data Scientists they need to care about and learn the business operations. How does the company make money? What are the largest sources of expenses? It's frustrating to see a Data Scientist jump into a pile of data without first talking with stakeholders to understand their problems or what decisions need to be made faster or more accurate. Leveling-up the DS team can have higher returns than trying to locate and hire the unicorn employee you describe.