AI has quietly worked its way into almost every part of building software, not just the coding itself. This page lays out the 14 layers that now make up a typical AI-assisted development stack, from tools that write and review code to the systems that secure it, test it, document it, and keep it connected to the rest of your infrastructure. It also covers the layers most teams don't think about until they're already deep in: keeping track of what these tools cost, whether they're actually improving output, and which ones are approved for use in the first place.
Layer
What it does
Representative tools
1AI IDE / pair programmer
Generates code, explains repositories, refactors, debugs, and writes tests inside the editor.
CCursorGitHubWWindsurfJetBrainsGoogle
2Terminal coding agent
Reads and changes files, runs commands and tests, and works across an entire repository.
AnthropicOpenAIGoogleGitHub
3Autonomous development agent
Takes an issue or specification, works independently, and returns a pull request.
DDevinGitHubOpenAIFFactory
4AI code review
Reviews pull requests for bugs, maintainability issues, and missing tests.
CCodeRabbitGitHubQQodoGGreptileGraphite
5AI security
Finds vulnerabilities, insecure dependencies, secrets, and risks in AI-generated code.
SnykSSemgrepGitHubSonar
6Testing and QA
Generates and maintains unit, integration, and end-to-end tests.
DDiffblueQQodoTTestimMmablTricentis
7Documentation and knowledge
Creates documentation, answers codebase questions, and improves internal knowledge access.
SSourcegraphSSwimmMintlifyGitBookReadMe
8Context and tool connectivity
Connects agents to repositories, databases, tickets, documentation, browsers, and internal tools.
AnthropicGitHubAtlassianNotionSlack
9Model access and AI infrastructure
Provides model APIs, routing, gateways, caching, logging, and prompt management.
OpenAIAnthropicGoogleMMicrosoftAWS
10AI developer productivity measurement
Measures AI’s effect on cycle time, throughput, quality, and developer experience.
JJellyfishDDXLinearBFFaros AILLarridin
11AI ROI measurement
Connects adoption and spending to labor savings, delivery gains, revenue, quality, and business outcomes.
LLarridinJJellyfishWWorklyticsAApptio
12AI workflow process mapping
Maps where AI is used, how work changes, and where bottlenecks or automation opportunities exist.
LLarridinCCelonisSignavioUiPathMMicrosoft
13AI usage, cost, and token management
Tracks model usage, token spend, budgets, vendors, and allocation across teams.
LLarridinPPortkeyHHeliconeLLangfuseDatadog
14AI governance and compliance
Controls approved tools, data handling, model access, permissions, audit trails, and policy compliance.
CCredo AIHHolistic AIIIBMMMicrosoftOOneTrust
Brand marks are the property of their respective owners and are shown for identification purposes only.