SDLC in the AI Era

Lifecycle phase

What changes in Build

Every change this guide lists for the Build phase, across all roles. Open one to see before and now, with sources, on the role page.

Product Manager / Owner
- Moving from a validated prototype to a real productCagan: prototypes are for learning; products must also handle scale and security
UX/UI Designer
- Working with developersCase study: AI code from screenshots matched static designs, not interactions
Software Architect
- Setting architecture standards for code that AI writesInfoQ: architects guide agents with prompts; Böckeler describes rule sensors
Software Developer / Engineer
QA Engineer / Tester
- Writing UI test automationPlaywright's agents plan and write tests from a request and a seed test
Security Engineer
- Reviewing code for security flawsIn Veracode's tests, AI code was syntactically correct but often failed security
DevOps / SRE Engineer
- Writing infrastructure as codeScan AI-written Terraform; well-formed code is not necessarily secure
Data Engineer / Analyst
- Writing SQL, transformation models and pipeline codeMost surveyed practitioners now use AI tools daily; organisations lag behind
- Building ELT pipelines with AI agentsOn a benchmark, agents largely solve extract and load; transformation improves
Project Manager
Scrum Master / Agile Coach
Engineering Manager
- Keeping delivery stable as change volume growsDORA: without strong testing, more change volume leads to instability
- Staying hands-onIn LeadDev's survey, some leaders are doing more hands-on technical work