The first read
Where to start with your first AI agent
The two or three things worth doing first with your agent, and what to skip. An honest starter guide to getting a return, not a mess.
Read the starter guideInsights / field notes for trade businesses
Read what matters before you hand over the work.
Practical guides for choosing the first office job, checking the return, setting the boundaries and understanding what your AI employee will cost to run.
The reading desk
Choose by the decision you need to make.
There is no required reading order. Pick the question that is holding up the next decision, then come back for the rest when you need it.
Evidence and outcomes
Work out what is worth handing over.
See the difference between AI that answers and an agent that completes work, then compare the jobs other operators have already tested.
- AI agents, explained What an AI agent actually does Not a chat tab you type into. What an AI agent actually does for a business, the jobs it takes on, and how it gets sharper over time.
- Use cases What could your AI agent do for your business? Real jobs agents are doing for businesses, with the numbers. Ad spend, overdue quotes, enquiries, research. Find the one that sounds like your week.
Control and risk
Decide what it can do, and where it must stop.
Read this path if customer messages, sensitive systems or a second decision-maker are part of the conversation.
- Risk and guarantee What if your AI agent doesn't work out? What happens if your agent doesn't deliver. Our honesty on the call, the control you keep, and the guarantee behind it.
- Security whitepaper Request the OpenClaw security paper Request the OpenClaw security and build-process paper. We'll email it as soon as possible.
- Why a platform Why a platform, not just ChatGPT A chat tab answers; an agent does the work. Why a platform beats a chat window or a DIY setup, and why owning it matters.
Cost and setup
Check the economics before the build.
Understand what drives cost, which model each job needs, where the agent should run and when a cloud, local or combined setup makes sense.
- Pricing and ROI What an AI agent costs, and what it pays back Why AI costs run away on a bad setup, how an agent compares to hiring or doing it yourself, and where the return actually comes from.
- Reference AI model guide: which model for which job Pick the right AI model for your agent. Compare price, quality and best-for picks across the leading text, image, video, text-to-speech and speech-to-text models.
- Infrastructure guide Local or cloud: choosing where your agent runs Should your AI agent run in the cloud or on a Mac in your office? How each option affects speed, cost, control, and where your data sits, in plain terms.
- Architecture guide Cloud and local: the best of both worlds Run a cloud gateway for always-on reliability and an optional workstation on your own Mac for local control. How the two-piece setup works together.
Day-to-day use
Picture how your team will actually use it.
Start here if the practical questions are about technical skill, messaging apps and what changes for the people doing the work.
- AI agents, explained How technical do you need to be to run an agent? If you can brief a new hire, you can run an agent. Why running one is as simple as sending a message, and which parts we handle for you.
- Messaging guide WhatsApp or Telegram for your agent WhatsApp or Telegram for talking to your AI agent? How they compare on setup, reach, group chats, and reliability, so you can pick the right one.
When reading is not enough
See the work happen, or bring us the job that keeps getting stuck.
Watch a practical session first. If you already know which process is costing time or money, walk us through it and we will tell you where AI can help.