Then get your foot in the door. Then upsell the ICM. NOTE: Jake already shouted me out and shared my Skool group here Verbatim what I just posted in my group is below the *** thingys. This is not self-promotion, this is me...helping you.... And about 150 of you already came over based off his post. ********************************************************************************************************* "First... @Carla Bosteder this was for you specifically, you asked, a few more did as well...but, don't say I never gave ya' nothin'. What I am seeing: Everyone wants to lead into businesses and sell "ICM". You and I both know they need it, but it's like walking up to a woman on the first date and trying to lift up her skirt. You're going to get slapped. Business owners are the same. You need to earn trust, so just because your friends will let you take a look up there and reorganize their entire business... It's a little bit hard of an ask to a stranger. You need to position ICM Folder as a high ticket upsell to an offer. Create a quick win, solve a problem, then they will trust you forever. They will even ASK you what else you can do for them. That is what the 3 laws is about. That is what the researching is about. That is what the customer's journey highlights. It's all in the classroom.... Here's an easier win you can get for business owners missing out on business because of late calls, calls while they're busy, etc and so on. Go look up the numbers - it's a real pain. It's just files n' folders, people: https://www.youtube.com/watch?v=_gV242QeLCw P.S. I built my front desk for my company as sassy'ish. I call her LIVE so you can hear that. The report I mention is in the description on YT and in the classroom.

Just off the phone with another business owner, an existing MSP client. We are starting a paid engagement on Tuesday. He has some crews on the road installing a product. Those crews are going to start taking before, during, and after photos of the jobs, to help illustrate the transformation that this product makes. He has asked me to come up with a way to help the guys on site manage the upload of the photos and related job information - hard working guys with not an awful lot of time or much IT inclination. I then suggested.... 1. Helping his guys with his initial request ➕➕ AND ➕➕ 2. Improving the quality of the information uploaded to make it more useful 3. Automating the process of those job photos being compiled in the correct order and a social media post about the job being drafted and presented for approval of both the photographs and the content of the social media post to the business owner (style based in his previous posting history) 4. And auotomating the posting to multiple social media accounts of the photos and the approved post content. 5. He doesn't know it, but I'll throw in automatic blog post creation as a bonus. I already have a fair idea of how I'm going to do it. The key point here is I didn't sell this to him. He called me for advice about a simple technical matter and I simply suggested that the rest COULD be done, and suggested a framework only (no specific tools or much detail) He then sold himself and spontaneously decided that, despite his own AI use, he needed us to do it for him because it souned a bit complicated.

Someone here has been trying to run an icm shaped system on gemini bc IT wont hand out api keys, and hit the wall where gemini code assist in vs code forgets the instructions and only takes one md file, and a Gem holds ten docs but cant write back to itself. gemini cli is the piece that closes this. u sign in with a google account over oauth, no api key involved at all, and the free tier is 60 requests a minute and 1000 a day. the reason it matters for icm specifically is that it loads context files hierarchically, the same shape the methodology assumes. it reads ~/.gemini/GEMINI.md, then the GEMINI.md at ur project root, then ones sitting in subfolders, and merges them in that order. u can pull other markdown into a context file with @path/to/file.md, so ur tree stays a tree instead of one pasted blob. /memory show prints the combined context that actually loaded, which is how u verify it rather than hoping. /memory refresh re-scans after u edit. and it writes back to disk, which is the bit a Gem cant do. for sharing with the team it just becomes the repo. they clone it, they get the same context files, nobody is hand maintaining a Gem. one honest caveat, the free oauth tier is tied to a personal google account, so if ur work account is locked down u would be putting work context thru a personal login. worth clearing that with IT first instead of finding out later.

Finished Playbooks 1.1. I decided to use the "Script → Spec → Build → Render" workflow to try and explain the workflow itself. The biggest thing for me was realizing the code really doesn’t come first. Once I had the script and visual spec figured out, the build became way easier to direct and refine. The hardest part for me was trying to make too much happen at once visually, especially in the refine section. Once I started separating each idea and letting the visuals happen in sequence, it got a lot cleaner. I ended up taking the Remotion render into CapCut and finishing it with the voiceover, music, sound design, pacing, and subtitles. This was my first real run through the whole workflow, and I definitely understand it differently now that I’ve actually built something with it. Which part of this workflow has been, or looks like it’ll be, the hardest for you?

The workflow: Hermes (orchestrator) routes every request. Skills already declare which model tier they need (the model: frontmatter is real, not decorative). When a request touches sensitive data, Hermes defaults to the local model — Gemma, via the Ollama instance Hermes already runs, currently idle — instead of escalating straight to Claude. 1. Local model scopes the request — reads the sensitive material, decides what needs coding. It only proposes; it never makes the final call. 2. A deterministic script does the actual encoding — swaps real names/values for stable per-entity codes via a codebook (table lookup, not AI judgment). . 3. A preflight check verifies nothing raw survived before anything leaves the machine — the check that can demonstrably fail on a bad run, not just a claim that coding happened. 4. Only then does it ask permission to send the coded payload out — your "greenlight and asks permission" step. 5. Claude works entirely on coded tokens — never sees a real name, real entity, real number. 6. The same local codebook decodes Claude's answer back to real names before you see it. That's the full round trip. Claude never sees a real name; you never see a code. What am I not thinking of?