By Betsy Burton
What is AI Architecture?
Today, nearly every application and service provider are either already supporting artificial intelligence (AI) capabilities or promoting future plans. This means we are all going to find AI proliferating throughout our work, communities, and homes, one way or another.
It is going to become increasingly critical for organizations to support AI architecture as part of their enterprise and business architecture. But what does this really mean? And how do organizations support an effective AI architecture?
AI Architecture Defined
It is a discipline that focuses on creating strategy, guidance, guardrails and inspiration for the management and use of AI.
In other words, a person practicing the discipline of AI architecture, an AI architect, is the trusted advisor across the entire organization for if, when and how business and technology organizations should use AI.
It is important to not think about AI architecture as a “thing.” It is an ongoing discipline that includes creating deliverables that guide the usage of AI (Toolkit End User Principles For Use of AI). It’s not something an organization creates and then they’re done. Supporting it also is an ongoing effort as the business, people and technology continue to evolve.
The audience for AI architecture deliverables will include business leaders, technology leaders, end users, developers, service and technology providers, business partners and even customers.
AI Architecture Scope
The discipline of AI architecture must be focused on understanding the business strategy, the business ecosystem, people (customers, employees, partners), processes, information and technology. AI architecture is not just about AI technologies. In fact, I believe the biggest challenges with AI are going to be about information quality and integrity, ethics, change management, security, and governance.
Many organizations will appropriately support AI architecture as part of their enterprise architecture efforts; just like having a business architecture discipline within EA or solution architecture within EA.
However, even organizations that don’t have a functioning EA practice must support an AI architecture effort. The reasoning is that AI is becoming so pervasive and is affecting people process information and technology across the organization.
AI Architecture Level of Detail
We often think about architecture work being done at multiple levels of increasing detail, including a strategic level, contextual level, conceptual level, logical level and implementation level of detail. AI architecture must reflect and be aligned with the business strategy so there’s clearly work to do at a strategic level.
However, the vast majority of AI architecture work will be at a contextual, conceptual and logical level. Most of the implementation level details would be performed by individuals that are specialists in their specific areas.
For example, an AI architect might provide a business manager responsible for Human Capital Management with guidance for how they can take advantage of AI capabilities provided by Oracle or Salesforce. It would be up to the business manager to work with their service providers to understand the implementation level details; bringing the AI architect in as needed to help address issues.
Bottom Line
It is not an option. Your organization needs an AI architect and needs to support an AI architecture discipline. Organizations that don’t support AI Architecture, risk becoming inundated with AI that may or may not be valuable, wasting time and resources on AI efforts that don’t deliver value, and supporting bad implementations that lead to potentially disastrous business outcomes.
I will be doing a webinar on October 23, 2023, about AI architecture and the role of an AI architect, Decoding the Role: AI Architect vs. Solution & Application Architects. During this session we’ll be exploring how AI architects can deliver value within businesses, and we’ll be discussing some principles that could be in place to support AI.
I hope you will join us for this session, and I look forward to discussions with many of you about this critical emerging discipline.
Hear From Betsy LIVE on Monday, October 23rd at 10 AM PT | 1 PM ET
Decoding the Role: AI Architect vs. Solution & Application Architects
Adopting AI-enabled technologies is as much a people, process, organization, and strategy issue as it is about technology. The best AI-enabled solution will go nowhere if it doesn’t address a future-state business need and is accepted by users. Adopting AI-based solutions requires outside-in customer, user, and business thinking; not inside-out technology first thinking. Adopting AI requires a business and future-focused architect to help guide the integration of people, process, information, solutions and AI.
- Why is an AI Architect Needed?
- What are the needed Skills and Responsibilities of An AI Architect
- How is an AI Architect Different from a Solution or Application Architect?