|
Over the past three months I’ve built six working tools for our websites without writing a line of code myself, and at the end of this email I need your vote on what to do with them next. If you missed Part 1, it covered the AI tools I actually use. This one is about what I’m making with them. What “vibe coding” means for a non-engineerVibe coding is describing what you want in plain English and letting AI write the software. It sounds like a developer thing, but it isn’t anymore: most people doing it today have no programming background. For me, it’s the first time I can turn a framework I’ve used with clients for years into something you can click on. What I’ve built so far
| 01 |
Executagility KPI diagnostic
This was our first application, and to be fair - I didn't build it, our CTO did. He used Codex and Cursor and leveled up the version we had a developer and designer create for us a couple years ago. Five minutes, no cost, and a clear picture of where your team’s time is going. |
| 02 |
Prioritization Matrix
I did this one based on a spreadsheet template that I developed over a decade ago and have refined over the years. It sorts competing initiatives by meaningful criteria to the business, so the team has a way to prioritize that is logical, rather than risking the "loudest" voice in the room setting priorities by default. This is also helpful as a PROGRAM Management tool when your company plans to execute multiple initiatives (Rocks, projects, whatever you call them) at the same time. |
| 03 |
Activation Chart This and the next tool are offshoots of the Prioritization Matrix. As you check the box in the above to "Activate" any given project on your list, then you can move to the next tab in the tool to see details for the Activated projects only. Then you can fill out a SCOPE template for each one as well, which helps teams think through the project boundaries, assumptions, risks, and mitigations. |
| 04 |
Team Load
The Team Load Visualization tool is the final tab of the sequence that starts with the Prioritization Matrix. It shows risk areas by name for each resource (human being) that you have assigned to any project. This is the one leaders tend to ignore when prioritizing, because they don't always think about which individuals are helping with which projects. Too often, the same person is juggling more than 1 project on top of their Day Job, and that's when things break down. I should know as I am often the bottleneck for projects in my own company! The great thing about the Team Load visual tool is that you can now SEE who may be over-taxed, and you can then re-prioritize their projects so they know what is the very most important one to work on if (when) time becomes scarce. |
| 05 |
Conversational versions
The same frameworks, but you talk them through instead of filling in boxes. I am telling you, this is EYE OPENING. It reminds me of the Gloria Steinem quote, "The truth will set you free, but first it will piss you off!" I ran it on my own projects and was reminded of all the things which can distract strong execution (seasonality, other project load, big customer needs, resources from different departments or that are outsourced, etc.). Man it is annoying to realize that the path to the future vision will incur roadblocks and potholes as you traverse along it! But better to have eyes wide open than trip without warning and not be prepared to get back up. |
| 06 |
Project Charter & Operating Agreement
For larger projects, these are the two template documents that we use at AJC to prepare the team for what we’re doing and how we’ll work together. At some point, I may combine them with the other tools, but for now they are stand alone - again, built very easily by uploading templates I had already created. I iterated until I felt they could be completed in chunks by Section Headers, and then the entire document can be copied or exported. |
How I’m actually doing itThe way I am doing this is by creating a project in Claude, uploading my templates, and asking for an interactive, user-friendly web-based tool that is dynamic and automatically prompts the next step while maintaining inputs that are already provided. Then I copy the code and paste it into a new page in Squarespace that I create for this specific use with my normal header and a block for "code." Sometimes I do see things like bold, color, font, size, verbiage within the code that I want to change, and since I did become marginally dangerous in HTML in the early 2000s, I can make those edits manually. If I do that, I copy/paste the code back into Claude so it captures it for next time.
| |
A guardrail around our MVP Iterations
Everything I’ve built is a practice tool, not a database. None of them ask for anything confidential, and we are not saving the inputs in a database. You need to export to text, pdf, or screenshot to retain your work as once you close the browser, it is gone. That decision is intentional, because we are truly in beta mode and I don't want my MVP and iterations to be bogged down by security concerns. When researchers scanned 1,645 apps built on one popular vibe-coding platform, 170 of them were exposing users’ personal data. If these tools take off, we will go through the proper channels to keep personal data safe (as an aside, we HAVE done that for the Executagility KPI which does connect to a database; and that is where we have invested money with professional developers and now our CTO has taken over the back-end administration, so I'm comfortable there).
| What people are telling meThis quote is from a Chief Growth Officer for a 35-person global research firm in the healthcare space after he went through my Vistage presentation on using the Prioritization Matrix, Activation Chart, and Team Load tools to evaluate the upcoming quarter. "I walked through the practice exercise with my leadership team and there was interest. The timing is good as we are evaluating workload by project, initiative, and role." Another quote from a CEO of a 70-person specialty clothing design and retailer on the Prioritization Matrix is "That's a neat tool!" I'm not promising that my AI-coded tools solve the world's problems, but it has been great to demonstrate the type of work that we do at AJC in an interactive way that anyone can use on the web. The question I can’t answer yetWhat have I learned from all this? Turns out that developing a first pass of our tools using AI-assisting coding in Claude really was not that hard. It does take some vision, some motivation, some trial and error, and I constantly come up with small tweaks to make the user experience better. The problem is the guidance for using these tools because without that, the tools are just like that brand-new set of golf clubs my husband gave me for my birthday the year after we moved to Florida- standing alone in a corner without being used (not even once!): fun to have, but not practical. Anyone can ask an AI to create them a prioritization matrix. What they can’t get from a blank chat window is the method behind it: which questions to ask in which order, how to tell a real priority from a wish, and when a “yes” from leadership doesn’t mean yes. So I’m weighing four options: 1) I could package my methodology to run inside OpenAI (apparently already has a marketplace) or Claude (surely will get one), so the AI’s suggestions follow the AJC approach. 2) I could build a small subscription product to sell via SaaS model. This seems very unlikely to succeed on its own. 3) Or we at AJC could coach the use of our tools with a moderate subscription fee so companies can learn to fish, but not be left on their own completely. Kind of like what I eventually plan to do with my golf clubs - hire a pro to help me learn. 4) Don't bother trying to sell my tools at all because most companies want to build their own. I’d rather find out now than spend a year building something nobody wants. That’s the same advice I give clients. Would you use the AJC methodology (with or without coaching) inside your own AI tools? Tap one. It opens an email to me with your vote already filled in. Add a sentence about why if you have ten more seconds. The why is what I’ll actually use. Talk soon, Andrea Andrea Jones · Principal, AJC Company
|
Next in the series · Part 3 · November 2026
I cloned myself. It hasn’t made a dime.
Why I built an AI version of me after seeing Brian Halligan do it, why I bought andreajones.ai, and the honest ROI so far.
|
|