Blog
AI and Workday, from the inside.
What we see building an AI agent for Workday consultants — the problems, the gaps in generic AI, and how real tenant integration changes the work.
Latest
How to choose which AI model to use in kiweely
Match your Workday task—from quick WQL lookups to deep Extend architecture and executive decks—with the ideal AI reasoning engine.
In this post
A practical guide to choosing between ChatGPT (GPT-5.6, GPT-6-Astra), Claude (Sonnet 5, Opus 5.5), and OpenRouter models for Workday tasks.
All posts
Newest first · 17 posts
-
Consistent, template-based deliverables for your Workday project
Store firm and client branding once. The AI agent generates perfectly styled project documents and decks grounded in tenant facts.
-
Rapid ideation and prototyping of Workday Extend apps
Describe what your business needs, let kiweely build and validate the app, and test it live in your Development tenant.
-
Debugging Workday Extend apps and orchestrations in kiweely
Traditional Extend debugging means jumping between Developer Site, tenant logs, and Orchestration Studio. kiweely connects the whole chain.
-
Debugging Workday Extend web services: Postman vs. kiweely
Postman tests endpoints in a vacuum. kiweely tests them inside your tenant's security, schema, and business context.
-
Claude Code with a Workday MCP server vs. Claude in kiweely
The answer depends on the job and the tools connected to Claude. Here is how to choose the route you can verify.
-
Claude Code for Workday Extend: what changes when it can reach your tenant?
Claude can write an app draft. Tenant access makes it possible to check what Workday accepts and what a user actually sees.
-
Use your Claude subscription with kiweely for Workday
A practical first run: connect Claude, choose a Workday project, and check the result in the tenant.
-
Workday Extend vs. configuration: how to decide
Concrete example requests, the custom object vs. Extend business object distinction, and what each path costs to own, licensing aside.
-
How to connect Claude or ChatGPT to Workday
No route fits every job yet, and Workday's own connectors are still in early access. Here is what each option can and can't reach today.
-
Multi-client AI for Workday consultants: keeping your clients' data separate
Context bleed and credential confusion, and how pinning each conversation to one client keeps every other client's tenants out of reach.
-
What happens after ChatGPT writes your WQL query
The 10-second win and the 10-minute reality: why the gap between AI-generated text and real tenant data is the problem that matters.
-
The Production tenant problem: why AI and enterprise systems need a trust layer
Four principles — visible actions, explicit consent, environment awareness, and honest verification — and how kiweely implements each one.
-
Why Workday Extend doesn't need to be a developer-only tool
Functional consultants already have the knowledge. What they lack is the mechanical ability to express it in JSON metadata and CLI commands.
-
From intent to deployed app in one conversation
A walkthrough of the full kiweely journey: describe an app, watch the agent build it, approve each step, and end up with a live, tested app in your tenant.
-
The 7 challenges Workday consultants face when using Claude for Workday work
Seven structural limitations — from stale knowledge and the copy-paste gap to multi-client chaos and the verification problem.
-
Why generic AI falls short for Workday Extend development
Six structural differences between a general-purpose chatbot and a purpose-built Workday co-pilot — and why integration, not intelligence, is the gap.