https://withavi.ai/enWithAvi is an independent personal site by Avi about AI at work. It publishes practical guides, briefings and analysis for people turning AI change into action, with sources linked and open corrections. It is published in English, Hebrew, Russian and Spanish. Alongside Avi, five clearly labelled AI characters prepare drafts; Avi edits and approves everything before it is published.enhttps://withavi.ai/en/posts/first-automation-small-businesshttps://withavi.ai/en/posts/first-automation-small-businessSun, 11 Oct 2026 12:00:00 GMTPick the task before you pick the tool. Five short questions help you find what is worth automating first, and what is better left to a person.https://withavi.ai/en/posts/demo-vs-productionhttps://withavi.ai/en/posts/demo-vs-productionSun, 11 Oct 2026 12:00:00 GMTA demo is built around the cases that work. The product meets all of them. What to check before you trust an AI demo.https://withavi.ai/en/posts/n8n-test-ai-automation-valuehttps://withavi.ai/en/posts/n8n-test-ai-automation-valueSat, 10 Oct 2026 12:00:00 GMTTools like n8n are a way to start small. Before investing a lot of money in a big project, I always recommend starting small and failing small.https://withavi.ai/en/posts/make-ai-small-businesshttps://withavi.ai/en/posts/make-ai-small-businessSat, 10 Oct 2026 12:00:00 GMTMake does more than pass data between apps. The AI inside it summarizes, sorts, drafts and even decides. Here is what is there, and where I would stop.https://withavi.ai/en/posts/spec-for-ai-coding-agenthttps://withavi.ai/en/posts/spec-for-ai-coding-agentFri, 09 Oct 2026 12:00:00 GMTA clear task means less guessing for the agent. It does not promise correct code, so say how you will check the result too.https://withavi.ai/en/posts/small-eval-set-before-launchhttps://withavi.ai/en/posts/small-eval-set-before-launchFri, 09 Oct 2026 12:00:00 GMTYou do not need a platform to start testing an AI feature. You need about twenty real cases, a rule for passing, and a rule for when not to ship.https://withavi.ai/en/posts/model-better-process-nothttps://withavi.ai/en/posts/model-better-process-notSun, 04 Oct 2026 12:00:00 GMTA model upgrade is not a process improvement. Here is how to check whether AI really helps your team.https://withavi.ai/en/posts/more-context-not-always-betterhttps://withavi.ai/en/posts/more-context-not-always-betterFri, 02 Oct 2026 12:00:00 GMTA good agent isn't built by endless addition. It's built by deciding what to leave out.https://withavi.ai/en/posts/ai-said-it-confidentlyhttps://withavi.ai/en/posts/ai-said-it-confidentlyFri, 02 Oct 2026 12:00:00 GMTA fluent answer sounds like a correct one. They are two different things.https://withavi.ai/en/posts/ai-projects-back-on-trackhttps://withavi.ai/en/posts/ai-projects-back-on-trackThu, 01 Oct 2026 12:00:00 GMTThis is where we stop talking about capabilities and start talking about work.