From SDLC to ADLC: The Enterprise Agent Development Lifecycle
summary
Enterprises must adopt an Agent Development Life Cycle that layers on SDLC and prioritizes outcomes, modular design and evals over traditional QA. Built-in claims: proprietary Palmyra models and an enterprise agentic platform; deployed with hundreds of Fortune 500s. Real result: a finance client turns a 2–4 week market commentary workflow into 8,000 reports -> five days annually, saving nearly $2M. New roles include agent owner, builder, champion, supervisor and practices like MVPs, trace-based evals and model fallbacks.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 2:10Housekeeping
- 3:33SDLC to ADLC shift
- 6:43AI apps vs agent systems
- 8:05Agent autonomy in enterprise
- 9:30Process mapping as prompt engineering
- 14:07Who builds enterprise agents
- 16:59Key roles in ADLC
- 20:35The agent development lifecycle framework
- 21:57Measurable agent outcomes
- 22:42Agent MVP design
- 24:27Scaling agents with design systems
- 25:55Evaluating adaptive agents
- 28:43Deploying and iterating agents
- 30:09Supervising agent behavior
- 31:13Five pillars for enterprise agents
- 31:57Q&A
- 32:39Collaboration without slowing innovation
- 33:41Testing non-deterministic agents
- 35:29Product managers in ADLC
- 36:30Multimodal data scaling
- 37:13Agents with legacy systems
- 38:58ADLC first principle
- 39:40Measuring ADLC ROI
- 40:43Fault tolerance for agents
- 41:28ADLC vs SDLC testing
- 41:48Closing remarks
related talks
33mThe Autonomous Enterprise
04mRebuilding Platforms for the ADLC - Justin Reock
33mAgent Experience
30mAgents Building Agents
28mAI Agents as Architectural Strategists | Aligning System Design with Business Goals
30mBuilding the Agentic Enterprise: Turning Data into Autonomous Business Systems
35mThe Agentic AI Engineer
01h 01mEnterprise Architecture 4 0 The AI Driven Future Preview