AI Universe · Experience

Experience zoom 2

Where do people meet it? 5 concept buckets. Each one is a job you can learn once; underneath it, who does that job on each platform.

Concept

Chat and assistants

A conversation window people type into.

web chatmessaging channelsvoice

Physical: who does this job, where

Open sourceChatwoot, Open WebUI
AzureCommunication Services, Bot Service
AWSAmazon Connect, Lex
GCPDialogflow CX
VendorsChatGPT, Claude apps, WhatsApp

Concept

Copilots in apps

AI inside a tool people already use.

inline suggestionsside panelsactions in the app

Physical: who does this job, where

Open sourceCopilotKit
AzureMicrosoft 365 Copilot, Copilot Studio
AWSAmazon Q
GCPGemini in Workspace
VendorsNotion AI, Salesforce Agentforce

Concept

Conversational analytics

Asking questions of business data in words.

questions to certified metricsnever free SQLshow the working

Physical: who does this job, where

Open sourceSuperset + a semantic layer
AzurePower BI Copilot, Fabric data agents
AWSQuickSight Q
GCPLooker conversational analytics
VendorsDatabricks Genie, ThoughtSpot

Concept

Coding agents

AI that reads and writes code with you.

completionagentic editsrunning tests

Physical: who does this job, where

Open sourceAider, OpenHands, Cline
AzureGitHub Copilot
AWSAmazon Q Developer
GCPGemini Code Assist
VendorsClaude Code, Cursor, Codex

Concept

Background agents

Agents that run jobs with nobody watching.

schedulesautonomy levelsreports to a person

Physical: who does this job, where

Open sourceyour harness + a scheduler
AzureFoundry Agent Service + Functions
AWSAgentCore runtime + EventBridge
GCPAgent Engine + Cloud Scheduler
VendorsDevin, Claude Code in CI

Zoom in further

← The AI application (agent harness) · Assurance & operations →

Compute & infrastructureDataModelModel accessThe AI application (agent harness)ExperienceAssurance & operations