AWS Builder Center

Building Software Like Never Before With Agentic AI

AI generates code faster than humanly possible. Shipping better software? That's a different problem. Most teams adopted AI coding tools - few changed how they build software. This article is based on the talk "Building Software Like Never Before With Agentic AI" from AWS TLV Summit 2026, exploring how Kiro shifts the mindset from personal AI productivity to collaborative, org-wide impact - spec-driven development, team workflows, and governance. The goal isn't more code. It's better software, faster.

This page collects the research behind the talk "Building Software Like Never Before With Agentic AI" (AWS TLV Summit 2026) - why more code is being written than ever, yet less software is actually shipping.A full written version will follow after the event. Registration is still open - join us:  https://aws.amazon.com/events/summits/tel-aviv/ 

AI Paradox accreditation-

  1. "741% more code generated- by AI" / "+20% releases" -  https://www.nber.org/papers/w35275  (academic) /  https://qz.com/ai-coding-tools-code-volume-releases-gap-nber-study-061126  (easy read)
  1. "AI spend ×28", "innovation ratio flat" - DX State of AI Q2 2026 -  https://getdx.com/news/dx-releases-q2-2026-state-of-ai-impact-in-engineering-report/ 

The Waste accreditation-

  1. "52% of All Code Is Now AI-Generated" - DX State of AI Q2 2026 -  https://getdx.com/blog/ai-authored-code-has-nearly-doubled/ 
  1. "79% faster developers, same delivery pace" / "82% believes AI creates debt" - GitLab AI Accountability Report 2026 -  https://about.gitlab.com/press/releases/2026-06-23-gitlab-research-reveals-organizations-are-generating-ai-code-faster-than-they-can-control-it/ 
  1. "Refactoring ↓70%", "duplication ↑81%" - GitClear - The Maintainability Gap, 2026 -  https://gitkraken.gitclear.com/write_only_mode_ai_research 

The 10x developer productivity myth

Costs & Benchmarks

  1. Kiro Cost Analyzer - A serverless sample for analyzing Kiro Enterprise credit usage and developer impact -  https://github.com/aws-samples/sample-kiro-cost-analyzer&#x20 ;
  1. Agent Cost Bench - Measure cost, quality, and latency side by side for different models & CLIs -  https://github.com/aws-samples/sample-agent-cost-bench 
  1. AWS Bench - open-source benchmark for AI agents on AWS -  https://github.com/aws-bench/aws-bench 
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article