Building an Effective AI Team: Strategy, Roles, Org Design, and Implementation

I am frequently asked about the need for dedicated AI teams, where such a team should report inside of organizational structures, what type of talent is needed, and how to get started. Here, I offer research-backed thoughts to address these questions.

What I Learned from My Doctoral Research of 46 Enterprise Uses of AI

I published research that examined 46 enterprise AI use cases from late 2021 through 2022. There were 5 clear themes identified that are needed to drive success with AI:

1. Value creation: You must pursue real business problem solutions, not "toy" AI
2. Focus on customer needs: This aids decision-making and prioritization
3. Collaborative teams with modern skills are successful with AI: Silos are deadly
4. Shifting culture to embrace failure and iterative learning is vital
5. The critical role of data: There is no avoiding this enduring fact

Based upon this research-backed grounding, let us dig into whether you should create a permanent AI team, who should lead it, the roles needed for success, and what steps to take to get started.

Should You Create a Dedicated AI Team?

Well, sorta. I recommend establishing an AI function that reports to the Chief Strategy Officer, or in organizations with forward-thinking technology leaders such as Chief Information Officers, this team can also thrive under their leadership. This function should be staffed with key roles using agile, business problem (value creation) specific teams. The organization would not have the same permanent team, although some members could see repeated tours depending on the value situation. Leadership, however, needs to be permanent and is often best aligned under the AI Strategy Analyst (definition below).

Why Ad Hoc AI Teams with Permanent Leadership Are Preferable

When it comes to building an AI team, many organizations face a critical decision: should they create a dedicated AI team? I suggest considering a hybrid approach that includes ad hoc teams with permanent leadership. This approach offers several advantages:

• Business value focus and clear accountability: By reporting directly to the C-level executive responsible for strategy, the ad hoc AI teams will be focused on driving business value and results, rather than technical outputs.
• Innovation: By forming teams tailored to each project, you allow for a diversity of perspectives, approaches, and ideas. This can spur more creative and innovative solutions.
• Flexibility: Ad hoc teams can be formed quickly for specific projects and disbanded once the work is complete. This allows for agility and flexibility in staffing based on need.
• Access to expertise: Ad hoc teams allow you to bring in the most appropriate internal and external experts for each specific project.
• Cost: Ad hoc teams are "on demand" and only staffed when needed. This allows you to avoid the fixed costs of a permanent team when there are no projects requiring AI.
• Learning: Each new ad hoc team faces a unique problem and sets of data. This forces continual learning and the avoidance of "rote" approaches that can happen on permanent teams.
• Resource optimization: Since ad hoc teams are temporary, enterprise resources like Data Scientists and Business SMEs can be assigned to multiple teams over time, exposing them to more use cases.

What Roles Should You Staff on Your AI Team?

Staffing your AI team with the right mix of skills and expertise is critical for success. Key roles to fill include:

• AI Strategy Analyst: Scan the market daily, stay current with tools, regulations/law, potential partners, synthesize intelligence into recommendations, and lead with the mindset of an entrepreneur who envisions creative solutions within the boundaries of internal governance structures. This role can be filled by an external partner, but it is preferable to make this the permanent leader of the team.

• Business SME(s): People who deeply understand the specific business problem to be solved (internal resource).

• Customer Advocate: Ruthless focus on customer needs to drive decision-making and prioritizing (internal resource).

• Data Experts: From both the technology side and the business side. Pick ONLY the data critical to solving the business problem and ensure it is in high quality condition.

• Enterprise IT Architect: Responsible for ensuring that the AI solutions are integrated into the organization's existing IT architecture in a way that is scalable, secure, and aligned with overall IT strategy (internal resource).

• Security and Privacy Expert: Addressing security and privacy concerns related to AI systems. This person must be both technically knowledgeable and working closely with the AI Strategy Analyst.

• Ethical AI Specialist: Ensures responsible and ethical practices within the AI team. This role develops and implements ethical frameworks and guidelines for AI development, promoting fairness, transparency, and accountability.

• Project Manager with agile mentality: Small, iterative efforts. Some efforts will not work. Learn and pivot (preferably well-respected internal resource).

• Business Value Owner: This person is responsible for defining measures of success and keeping the team laser-focused on driving towards value realization.

• Leadership: Committed to value creation, applauds learning as ROI, creates a culture that is (small) failure absorbent, abhors organizational silos, and champions/funds collaboration.

Note on Team DNA: All team members must possess natural curiosity, an appetite for continual learning, resilience when small efforts do not produce expected results, a collaboration-first work approach, and a desire to solve real business problems (versus tech for tech's sake).

Steps to Assembling Your AI Teams

Assembling an effective AI team requires thoughtful planning and coordination:

1. Clarify the business problem/opportunity: Make sure there is a clear understanding of the specific business issue the AI initiative aims to solve or opportunity it seeks to capture.

2. Identify and secure key internal resources: The Business SMEs, Customer Advocate, and Business Value Owner should be identified first.

3. Map needed expertise to roles: Determine which of the defined roles require internal resources vs external partners based on availability of expertise in-house.

4. Prioritize project management: Appoint an experienced Project Manager, preferably an internal resource, as early as possible.

5. Seek external partners selectively: Bring in partners to fill skills/expertise gaps that cannot be met by internal resources. Focus on temporary partnerships, not long-term vendor relationships.

6. Establish clarity of role/responsibilities: Make sure each team member's responsibilities are clear from the start. Develop a responsibilities matrix mapping roles to key tasks.

7. Gain governance and ethical clarity: Identify existing applicable organizational governance and ethical frameworks. Then, assess the need for added AI safeguards.

8. Foster psychological safety: The team should spend time building trust, rapport, and an environment where people feel safe contributing ideas and taking risks.

9. Establish OKRs for value metrics: Define the specific Objectives and Key Results that will guide the team's work and determine success.

10. Iterate quickly: Plan for an iterative process where the team assembles, does work, learns, and potentially makes changes to roles/members for the next iteration.