Floyi 1.7.5 Release Overview

All right. So we have a few big updates here for Flowy. First, we have ChatGPT is now one of the AI search engines in the AIERS Analyzer. And then we've also added generative intent as the 5th search intent. And we'll get into that in a second. And then we're also grabbing the fan-out search queries from all the AI search engines themselves. So we're not creating something where we're trying to estimate or guesstimate what the AI search engines are using. We're actually getting it from the search engines themselves. Let's get into it. I've purposely used this search query. What's the best diet for a man in his 50s who works out 3 times a week? I wanted to see what a YMYL sort of prompt topic would bring us. And you can see here. In the user intent, our AI agents have seen that the primary intent is actually generative intent. And we have a nice little handy intent definitions legend down here. And you could see generative is getting AI to search, synthesize, and respond with a tailored ready-to-use content. So it's not just, it's not just finding the steps and giving you information so that you do it yourself. You're actually asking AI and ChatGPT or Claude or Gemini, whatever your favorite AI search engine is, and chatting with it and getting a plan for the best diet for a man in his 50s, right? And the secondary intent here is also informational. A couple other things we've actually added on here is the list ratio, list to content ratio. And the average words per sentence and the number of AI citations from each AI response. So coming down a little bit further into our content blueprint, you could see a great little content brief that will help explain what you should be creating based off of all the AI responses. So we're not just giving you the responses from the AI. That's simple, right? You can do it yourself and not that you want to, but you can open up 8 tabs. Then you can put in the query yourself and track each one. And then you can analyze it, but our, you know, Flowe does that all for you. So based off of all the AI responses sources, we're distilling everything down into a nice content brief that you can use to help you get cited. What's the unique angle for this piece of content? What kind of things should we create? So you could see here, the format is a comprehensive guide with embedded personalization tool, TLDR, plus persona calculator, plus downloadable 7-day meal plans and grocery lists, plus an expandable evidence grounding block. It's a lot, right? But your target, the person asking this initial query, what's the best diet for a man in his 50s who works out 3 times a week? It's pretty specific. And for an AI to want to cite you, you've got to get pretty specific yourself. Now let's come down here a little bit more. So we had priority keywords and we had key entities before. These are all the keywords and entities that the AI search engines would return to you. So we have the core keywords, the ones that are commonly used in the responses. We have the supporting keywords. Less commonly used, but still used organizational entities and brand entities that are mentioned throughout. But here's the cool stuff. The fan-out search queries used. You can see all these different search queries that are used. We see the tools that gave us fan-out search queries are ChatGPT, Perplexity, Claude, Gemini, and OpenAI Web. Just to be clear, OpenAI Web and ChatGPT are 2 different sources. So OpenAI Web Search is the API and ChatGPT is ChatGPT. So you can see all these different queries that are being used. And then you can see the related searches that are fetched from AI Overviews and Bing Copilot. Obviously AI Overviews and AI Mode, we can't get them to give us our, the fan-out search queries yet, but the next best thing is the related searches. What else? Other big updates is GPT-5 and 5 Mini here. All right. So we didn't immediately add GPT-5. We didn't just plop it in into Flowe. We actually did a lot of testing because this is a different AI model, LLM model, than the previous version. This one's a reasoning model. So there are certain levels and parameters There's a verbosity and there is a reasoning effort. In the reasoning effort, they're pushing, promoting a lot the minimal reasoning effort. And, you know, obviously it takes less time. It's very quick, but the quality of the output is pretty minimal too. So we actually did a lot of tweaking. There are areas depending on the tool and the feature that you're doing in Flowey, we either use the low or the medium. Reasoning efforts. So if you find yourself using GPT-5 and it feels slow, it's because we're actually using a higher reasoning effort. So the longer it takes, it could be the reasoning effort's a lot higher. So it's really thinking about what to output. Try different things. We still have all the previous versions, the GPT-4.1, 4.0, The Minis and then 03 and 04 Mini. We might remove the 03, 04 Mini eventually because the output from the GPT-5 Mini is actually really good. We've done so much tweaking with the parameters in setting up GPT-5 Mini that, you know, the 03 Mini and 04 Mini don't really make sense anymore because they also cost more credits. And then we still have the Claude and Gemini. But for everything that we're doing here with creating topical maps and the research and the analysis, the models here are the ones that do really well. All right. That's it for this Flowe update. Sign into Flowe, try them out, hit us up if you ever have any questions.