A short, 30 min talk in which we’ll look at how Generative AI can positively impact the entire spectrum of SDLC.
First we will scout through the phases of SDLC like product research, requirements gathering - just so that we don’t naively equate Generative AI to Github Copilot and assume it is only applicable for coding.
Half-way into this short talk, we’ll spotlight on software development. And how can we systematically approach software development.
People are still figuring out what those 10% skills are - whether it is language proficiency or critical thinking or critiquing something.
But there is no running away from an immediate recalibration.
There’s an explosion of tools in this area, and for every phase. Leading the pack is ChatGPT and Copilot.
Not only are there ground-up Generative AI tools such as ChatGPT, but existing well-known tools/platforms like Vercel, a hosting platform or Harness, a CD platform - all have Generative AI integrations.
So this is definitely a wave
Codeium, Tabnine, Copilot,
In each of the SDLC phases, there are multiple sub-phases and activities under them that go on.
Most of them have some applicability of Generative AI.
-User research
- Market research
- Regulatory research
User research is understanding behaviour, needs and motivations of users so that we build the right things, the right products.
https://docs.google.com/presentation/d/1AFaYo_prC2CKlRw3F4XVpOjuJ59yJ-MrUrecdbuxRk0
User interview strategy formation - finding user personas, structure and content of interviews
Summarizing results from reviews in Google, FB etc.
Product value articulation - both for investors, and also for the common people
From a broad problem statement, you can go top-down, i.e. to identify key features.
From there you can drill down to agile user stories.
Give a bunch of stories, create a domain model for it
Refine through iteration
Take the domain models created in the previous slide. Use that as a starting point.
Take the user stories generated before.
Generate the acceptance criteria for each story.
Iterative: First create the DTO, then create the response class, then a controller, then a service.