Tools to Get Started
Tools to Get Started
You do not need to pick the perfect tool. You need to start. Here are the best options as of early 2026, organized by how much they can do.
Chat — ask a question, get an answer
Where most people start. Good for building and testing individual workflow steps: paste your inputs, refine your instructions, save what works. Free tiers available; paid tiers ($20/month range) unlock better models and longer context.
- ChatGPT — OpenAI. The most widely used. Large plugin ecosystem.
- Claude — Anthropic. Strong at long documents, nuanced writing, and careful reasoning.
- Gemini — Google. Best if you already live in Google Workspace.
- Grok — xAI. Integrated with X (Twitter) for real-time information.
- Microsoft Copilot — Microsoft. Chat interface plus deep integration with the Office 365 suite — it can draft in Word, analyze in Excel, summarize in Outlook, and build presentations in PowerPoint directly within the tools you already use.
Agents — give it a task, it works through the steps
These can browse the web, use tools, read and create files, and chain multiple steps together autonomously. This is where your workflows start running as real systems instead of manual copy-paste.
- Claude with Cowork — Anthropic. Agentic workspace within Claude that can operate your computer, browse the web, and work through multi-step tasks while you watch or step away.
- Codex — OpenAI. Desktop agent that can browse the web, use tools, and work through multi-step tasks on your computer.
Command line — maximum control
For those comfortable in a terminal or willing to learn. These tools run on your computer, work directly with your files, and give you the most flexibility for building sophisticated workflows. They are more technical but significantly more powerful.
- Claude Code — Anthropic. Terminal-based agent.
- Codex CLI — OpenAI. Open-source terminal agent.
- Pi — Open-source coding agent with an extensible tool and workflow system.
The tool matters less than the method
Every tool above can execute the principles in this guide. What matters is:
- You describe your workflow clearly.
- You identify constants and variables.
- You build incrementally (one step at a time).
- You test with real data.
- You refine based on results.
Start with whichever tool is most accessible to you. You can always move up later.