AI and the Digital BA—What’ It All About? Part 2

This is the second of a two-part article written with answers to some of the most frequently-asked questions I get about artificial intelligence (AI).

In part 1 I addressed some common terms and issues relating to AI as it is used in a business rather than technical context. In this article I will focus on the various roles the BA plays to help organizations with their AI initiatives. As with the last article, I will use a Question and Answer format.

Quick Review of Part 1

What is AI?

AI is an umbrella term that encompasses all digital technologies, like machine learning and predictive analytics, which are used to make predictions and recommendations using massive amounts of data. In short, it’s machines doing human tasks that range from simple to complex.

What is a digital business analyst (BA)?

A digital BA is a trusted advisor who helps organizations with their AI strategies. Rather than developing the strategies, they provide their advice about impacts to and value of AI initiatives.

What skills does a digital BA need?

The skills don’t change, but the subject matter is incredibly complex.

How successful are most companies with their AI efforts?

Not very. Most AI initiatives totally miss the mark and result in all kinds of issues, not the least of which is financial. A recent Forbes article details some of the resulting issues.[i]

What is digital fluency?

Digital fluency is defined as “The ability to interpret information, discover meaning, design content, construct knowledge, and communicate ideas in a digitally connected world.” [ii]

Part 2

What is the role of the BA on digital projects?

A digital BA can be involved in many aspects of an AI initiative. Some of the roles that a BA may play include one, several, or all of these:

• Strategic BA. In this role BAs help organizations determine the value and direction of the AI effort. Some of the specific outputs can include:
  • Business case on the value of the AI initiative
  • Recommendation(s) on the best strategic approach to the AI initiative
  • High-level implementation plan
  • Pitfalls to avoid
  • First look at state of the data to be used
  • High-level governance plan

• AI coordinator who implements the AI strategies. In this role the BA coordinates AI initiatives across project and portfolios.
• BA on a project(s) that is part of the AI initiative. Although this role is similar to any BA role, there are some differences. The BA will need at least working knowledge of, if not expertise in, AI.
• Business data analyst. In this capacity the BA may
  • Analyze the current data to determine how much is useable, how much needs to be cleansed, and how much needs to be collected
  • Recommend an approach to cleansing the dirty data
  • Help determine the data needed for predictive analysis and other AI functions
  • Interpret statistical analysis resulting from AI functions
  • Be an AI translator to facilitate communications between the data scientist and the business stakeholders.

What’s the difference between a data scientist, data analyst, and BA who works a lot with data?

These 3 roles can be confusing. At first glance we might not recognize differences or understand why the distinctions are important, but they are. I discussed the possible roles of the BA above, so here is a brief description of the other two.

Let’s take the easy one first—the data scientist. Not that the role is easy, it’s just easier to explain why this one is different from the other two. The data scientist is the most technical and needs the most expertise. About three-fourths have master’s degrees in mathematics and statistical analysis. Over half have Ph.Ds.

Data scientists create the predictive models. They determine what the machines need to do in order to meet the business objectives. They decide which algorithms are best given the objective of the AI initiative so that the machines can be trained to learn. Having said that, unless there is good governance and substantial input from business stakeholders and decision-makers, those algorithms have the potential to be created with built-in biases. Likewise, they may not be the best ones to solve the business problem.

The data analyst. This is really a subset of the BA role. I described some of the high-level functions above. On AI projects it’s necessary to focus on the data because it’s so integral to the success of the effort. Machines learn based on historical data. Issues like dirty and redundant data, as well as ownership of the data aren’t easy and require a strong facilitator and influencer to resolve. This data analyst role is such an important role that IIBA has created a new certification—the certification in business data analysis (CBDA).

What are some of the business and technical pitfalls that the digital BA should be aware of?

Here are some of the big ones:

Strategic

• Beginning with AI as a solution without a defined problem
• No real AI strategy
• Unrealistic expectations of what AI can do for the organization

Data and technology

• Dirty data
• Business processes don’t support the technology
• Weak security

Organizational and communications pitfalls

• Siloed and cumbersome business architectures
• Inflexible organizational structures
• The data scientists create the business rules
• The data scientists talk directly to the business and the business does not understand
• Confusing roles on AI projects
• Built-in biases in the algorithms

In Part 3 of this article, we will explore other aspects of how BAs can help organizations get the most value from their AI initiatives. Some of the topics we will cover include the need for governance on AI efforts, the recognition of the importance of the AI translator role, the digital PM, and more. 

[i] https://www.forbes.com/sites/insights-kpmg/2019/12/10/data-governance-is-risk-number-one/?utm_source=TWITTER&utm_medium=social&utm_content=2937780067&utm_campaign=sprinklrForbesMainTwitter#90dd59b91c81

[ii] https://www.slideshare.net/RobinAshford/guiding-learners-toward-digital-fluency