Role of Business Analyst in AI/ML
Business analysis is the art of bringing business requirements together to draw a perfect software to satisfy the human needs with Information Technology.
Without putting you straight into deep water, I’d share some things I found during my literature study. Hope that’d help you to understand the ‘background’. Here we go!
When the organizations are gearing up, they come up with a need of streamlining their existing business processes with current organizational processes. There they look in to,
- Business Process Automation (BPA)
- Business Process Improvement (BPI)
- Business Process Reengineering (BPR)
When the organizations started giving technological ‘touch’ to their existing business processes, they started looking for ‘new technologies’ that could help them to make decisions based on more accurate data and make them stand out among the competitors.
That is where Artificial Intelligence came on to stage.
The concept AI was planted by philosophers who attempted to describe the process of ‘human thinking’ as the ‘mechanical manipulation of symbols.
There are TWO MYTHS that we commonly believe about AI.
- AI will replace human
- AI will compete with us and then we will lose our jobs
This is where the resistance comes to welcome “AI” to organizations.
That is where the organizations need business analysts to step forward.
Organizations expect Business Analysts to do below tasks in the context ‘AI’.
Understand the domain
Regardless of whether it is AI, a business analyst should understand the domain before coming up with solutions to address customer pain points. Techniques such as fish bone diagrams, 5 whys support BAs to identify the root causes of the pain points and address them. If we are treating the problems that are in the surface without looking into ‘root cause’ the solution will definitely be a temporary solution which we can simply call a ‘patch’. The most important thing that you should know is, ‘permanent solutions’ will only be coming from a person who knows the ‘domain’.
A hospital identifies that most of their patients die in ambulance while coming to hospital. Reason identified was that the ambulances do not reach the patients on time to save the lives. So the hospital wants their IT team to come up with a solution to solve this. The BA suggests a system that could instantly alert nearest ambulance driver in an emergency.
A marvelous solution yeah??? No, it is not.
I’ll take you through the 5 whys for this problem.
Problem: Patients die in the ambulance before reaching the hospital
Why? Ambulance does not reach on time to get the patient.
Why? Current number of ambulances cannot satisfy the demand.
Why? Only two ambulances are functionating. All five other ambulances are broken down.
Why? The ambulances were not repaired on time.
Why? Management has not invested for vehicle repairs as there was no proper mechanism to predict the future demand.
Solution: An AI solution that could predict the future demand based on historic data and give insights to management to invest on repairing the old ambulances or buying new ones to save more human lives.
AI is a huge investment for a company. A BA who does not know the ‘context’ cannot come to this level of analysis. Knowing the domain is very important before looking for root causes. This is common for any industry.
This does not mean that you should be able to code (Well… being able to code is an added advantage). When you are gathering and analyzing requirements for AI based solutions, it is good to have some sort of a sense about the technology as you will have to come up with requirement documentation for the solution proposed. If you do not the ‘answer’ for the ‘problem’, you are not really ‘solving the problem’.
While writing functional specification documents for AI based solutions, a BA will have to wear a couple of hats.
> Business Analyst
> Domain expert
> Business Architect
If the above three roles are blend in a proper manner, that BA is ready to hands on an AI/ML project.
Identify the gaps between project POC and project objectives
It is recommended to come up with Proof of Concepts (POCs) AI/ML solutions (predictive models mostly) for a couple of reasons.
> It supports to deliver more concrete and immediate value
> It identifies potential problems and finds efficient solutions
> It compares different approaches without spending too much time (less struggles)
> It supports to get feedback and improve the data set
So the business analyst should be able to understand whether the POC is aligned with the business objectives, otherwise the organization will have to drive the project to a different direction to achieve the required objectives.
There are a lot more things to talk about, but I am not going to make this article longer anymore.
As business analysts, improve your……….
- Idea generation capability
- Creative thinking capability
- Technological soundness
- Flexibility to fit in with new technologies
- Risk taking capability
- Ability to adopt to change
……….in order to fit in to the project world with AI/ML.
As I specified earlier, AI is not going to steal your job if you are capable enough to challenge the innovativeness and creativity within you.
AI never replaces you! It supports you to perform better!
– Written by Maleesha Ashanshanie
Cover Picture Source : Design photo created by rawpixel.com – www.freepik.com