Hey everyone, I have been in the lab like usual, working on some pretty cool things. I want to update everyone to let you know what I am working on for Opulence AI and for you guys! This past month my focus has been all about creating a reliable marketing channel to get consistent inbound customers. We created 3 easy to digest products that we are currently marketing, 1. Change Order Shield: This is a tool that allows contractors to draft up project specific change orders ready for a client signature all from their voice and their phone. 2. The Site Reporter: Creates daily logs and weekly client updates by having an agent SMS text crew members to gather what happened on each jobsite 3. Proposal Master: This product lets contractors know when a client is looking at their proposal so they can give a call to those prospects in a timely manner. I realized that in order to make a paid marketing campaign work, you need to have products to sell. Who would have thought. So that is why I have been pushing my development team to get these products bomb proof while I have been building out the marketing funnel. I have learned a lot. First of all I have been only using AI to create engaging video ads. These ads have been generating leads, no sales yet though. I believe this is because the first ads I created had a very unstructured offer, with out clear landing pages with checkout options. I have attached a new ad that I have created to this post, believe it or not I built and AI system that fully created and edited that video. So this time around I am recreating my whole campaign with 15 new ad creatives aimed at targeting specific user pain points and translating those into highly engaging AI video ads. Creating these ads is far too time consuming, so I did what any self respecting AI agency owner would do. I automated it. From the video concept research to the script creation, the video generation and the editing AND (you won't believe this) even the posting and monitoring to Facebook.

I created this community from the lack of communities that discuss brand new techniques for creating complex systems that withstand the harsh test that society has for AI solutions. Let's face it... AI already has a controversial name. Right now we are split between hyper optimistic AI engineers, rightfully so (that was ironic), and the hyper pessimistic majority of the world. That means I will be creating courses and resources in this community that dive way deeper than explaining the 10 best nodes in n8n... I will be creating videos and resources that will help you take your AI engineer skills to the next level, which means, the competency to look at ANY system in the world, and having the ability to build a custom AI powered system that can 10x it. Recently I have been creating a production ready estimator for one of our clients and it has gotten me diving deep into operationalizing evaluation systems, that make it easy to twist and turn the knobs to get your system to perform up to its intended use 100% of the time. So I leave the decision up to you guys, would you want to see a course on Algorithmic System Building OR Evaluation System Building ➗ Algorithmic System Building: Building algorithmic systems takes fundamentally probabilistic LLM technology, and forces it to be more deterministic without losing the core strengths that AI gives you. This is all about looking at a problem that a business brings to you, and creating a repeatable equation and weighted scoring system that allows the LLMs to take messy unstructured data, and turn it into weighted variables that are plugged into an equation that ultimately makes the decision for the business. This is how to build a system that reasons like a human but executes like a calculator consistent, tunable, and reliable every single time. 📝 Evaluation System Building: A course on how to think and write evaluation tests so that you can guarantee to clients that you will get the system working to a measurable output. Building evals has been identified by Sam and Dario (open ai + anthropic hotshots) as one of the most important skills for engineers. They’re how you prove your build works instead of hoping it does. Without evals, you’re flying blind you don’t know if your estimator, agent, or workflow is getting closer to “accurate” or just spitting out pretty answers. An eval takes the goal you want (like predicting cost, scheduling correctly, classifying a task) and creates a measurable test that tells you exactly how far off your system is from perfect. It’s the difference between guessing and engineering.