ZD 25.41: Is AI Useful Under Fire?
AI and other modern software helped me craft a huge trade show exhibit.
I spent the last few months executing a major trade show exhibit. Modern software helped a ton, I am excited to share what I have learned. Meanwhile in The Distilled Spirit find articles about China, Minerals, and Vibe Engineering. Stick around to speculate about OpenAI’s strategy, look at how Microsoft came to be and enjoy the story of Cash Money Records.
Different Decade Same Problems
One-off, large-scale booth coordination and builds are interesting projects. They require discipline, creativity, a bit of technical know how, and some combination of faith and flexibility to succeed. They also require an agile software stack to support them. I built a lot of booths until 2016. For reasons beyond my control, the work dried up for nearly a decade. Then I got the call — the band was getting back together for one last show. We just delivered and I feel great.
We were building a complicated, multi-stakeholder booth as well as an accompanying stage show. We needed to track a number of competing ideas for exhibits and suggested sessions and against a set of common themes. We needed to track a roster of contributors. We had a lot of call and meeting notes associated with figuring all of this out. Beyond the theoretical there was a huge layer of practical — we had to take what was selected and get it onto the right places on the right walls. We had to review and distribute designs and compositions. On top of this we had to manage the build — the numerous design deadlines, physically tracking donated gear delivery and return and managing registrations for hundreds of people with distinct groups and specific access or accommodation requirements.
A lot changed in the last decade. The band was a lot thinner. We no longer had a marketing or a registration team lined up and ready to go. But software had changed a lot too — most important AI was a practical tool one could employ, not pie in the sky or buried very deep on the back end. It proved to be an interesting laboratory in the state of AI and productivity software in mid 2025. Here is what I learned.
TLDR
Remote planning works now: Slack and Zoom made real-time + async coordination easy.
Miro = shared sketchpad: faster visual decisions with context in one place.
Airtable + Notion = AI powered sources of truth. Databases > documents.
AI copilots saved hours: Custom GPT factbook for venue rules, Perplexity for quick briefs, Copilot surfacing email action items.
Communications Tools Have Really Improved
Planning this sort of operation in 2016 required moving a team into a conference room for months. It was abundantly clear that we learned a lot during the pandemic. Slack provided a real-time communications platform that let you slip in quick questions without having to sit around the same table all day.
The last time we did a similar project, we were pushing the envelope by using GoToMeeting in the ‘10s. I’d say we had about a fifty percent success rate for getting callers to see the screen overall. Nobody used a camera once we got them online. The post-pandemic world is much different. Teams, Zoom and even WebEx are all much better than GoToMeeting ever could be. Things just work thanks to technologies like WebRTC and better platform maturity in general. It largely just works in 2025.
The most marked improvement is that core competency in at least one video conferencing platform is a necessary corporate survival skill. Reliable video conferencing saved a ton of time; being able to look at the same document on demand is an amazing power when planning an exhibit. At the very least, I was not chained to an office in the same way I was a decade ago.
When I was flying solo, built-in call notes were a life-saver. Leading a call while I kept perfect notes was impossible, so the auto-notes helped. This kept my action items from slipping through the cracks or off the radar. I like how Zoom approaches this, but I found Notion to be the winner. More on that platform later.
Miro is a Magical Sketchpad
A decade ago, I was marking up PDFs to give design feedback. But we could not really design visually nor collaboratively nor in real time. What we could draw was primitive and hard to ship around. I needed a great drawing tool. Enter Miro — a collaborative whiteboard application we adopted during the pandemic.
Miro is a downright amazing tool for collaborating on a visual project such as a large-scale trade show booth. Miro let us arrange visuals and annotate them in place. It was our sketchpad in the sky, and the one place we could effectively combine the design, messaging and partner threads with the physical space of the exhibit.
Truth in Airtable
For this build, I found that tracking everything is ultimately a data-centric operation. The tools of the bygone age had been lost or superseded. We did not have a registration system, CRM tool nor customer experience team to leverage. The old Access databases we used to manage speakers were not going to open on my Mac. Spreadsheets were never the right way to manage content and contacts.
Enter Airtable. It was another pandemic-era tool we adopted and kept for general use. It spans from complex spreadsheet-style databases to multi-tier Access-style applications. This ended up being the latter — I built a tool that tracked all of the details of the exhibit, with core tables for Exhibits, Sessions, Contributors, Assets/Shipments, and Deadlines. The range included the suggested exhibits, through developing requirements to delivery details. It really did enable a single source of truth, that could update across the board using relational algebra. Interfaces let you show the right people the right information.
In my experience, Airtable does have some limitations. First is that, to build a multi-tier data app, you need to exercise a lot of discipline in the tool. This in turn makes it harder to consume — you end up building fixed interfaces, so you need to anticipate what people want and how they want to use it, and update interfaces as needs evolve. In fast-moving projects this is especially challenging as divergent paths emerge very quickly, often before I realized I’d forgotten to expose a column or make a field editable. Sharing is also a problem — Mobile can be hit-or-miss, and external sharing often requires the recipient to have an Airtable account and explicit access.
Zen Through Notion
Notion is one of those things that keeps impressing me wherever I use it. This project was no exception. It has a very unique ability to be a great text editor and a passable relational database manager. Moreover, they have done an amazing job building very good AI into the product directly.
First and foremost — and how Notion got the call to join the crew — was because I had combined the relational database and text editor into my primary note taking tool. It gave me the ability to take notes using a variety of templates and store them by project with some other metadata. Simply put, I have one big notebook I can easily filter for the right notes in the right date period.
Notion’s AI makes this more interesting in two ways. First, on the ingest side, their AI notes are amazing. The ability to combine the key things I highlight with first-rate voice capture is very helpful. The fact that I could use it even when it wasn’t my Zoom was extremely helpful. As was redundancy, I would often Zoom record and Notion record to make sure I got one good copy of the notes. From live events, I’ve learned: if you aren’t recording twice, you might as well not be recording.
Once I had the information, the AI came into its own. It really just worked — I could point Notion at this pile of more than 100 call readouts. I could ask it questions like “Which partner said they would need internet access?” and get cogent and accurate answers in a few seconds. Answers with a breadcrumb trail back to call notes for verification and color purposes. For someone wearing more hats than I should have, it was a godsend—an external brain, settling questions without engendering more questions and backing up your notions of the truth.
Custom GPT Factbooks FTW
I’ve found Custom GPTs to be an amazing no-code way to build a knowledge base. One thing that governs your life at these events is the exhibitor handbook. I had ChatGPT help me build a Custom GPT around those documents. I shared this with the team, and we could reference venue pricing and deadlines very easily throughout the process. I had something that could answer “how much is internet service in this venue, how do I order it and what is the deadline?” effectively.
Fast Research with Perplexity
When I was building a wide-ranging, industry-encompassing exhibit, I needed to understand the factual basis for far too many things to foresee. In a bygone era, I would spend lots and lots of time researching things about particular subjects to prepare for calls. Now, with Perplexity, I can get a very good rundown on most subjects in seconds. Or a minute or so if I want to jump to deep research mode. Answers are sourced too — meaning I can dive beyond to trust, but verify, the AI’s findings. It helped me not look dumb more than once, and for this I am thankful.
Copilot Finally Delivers
Email was my bane on this project. It’s the least-bad lingua franca for business communications. The reason it is the bane is things get shunted into so many threads and chains that finding the right answer, the current version of the truth, was nigh impossible without lots of manual searching. My hope and dream was that Microsoft Copilot could solve for this — after all it had my email.
I am extremely pleased to report that with the current Microsoft 365 Copilot, I can effectively search my email for “What open action items do I have in my email from $PROJECT_MANAGER” and get a real, accurate and actionable answer. Just as good as looking through all the email with much, much less pain.
I don’t spend much time in Microsoft land, but I could see this being very useful for people who live in Teams, Outlook, and OneDrive.
Don’t Forget the Printer
When I got onsite, a lot of this technology went out the window. I wasn’t in front of my computer; I was rushing around the floor with only a mobile phone. Connectivity might be spotty. Sometimes a piece of paper with the right information in your hand is what you need. Always pack the printer.
The Distilled Spirit
China and Rare Minerals
🔨 The China Factor ()
China has transformed itself over the last few decades. It features the best in clean energy and AI-enabled authoritarianism the world has to offer.
🪨 Anatomy of the Antimony Trade ()
This week’s edition of trade finds a way: Antimony oxides go to Belgium to become Antimony again.
OpenAI’s Big Play
🪟 OpenAI’s Windows Play (Stratechery)
One way to read the last week’s announcements is that OpenAI is building a platform not unlike Windows.
💬 OpenAI as America’s Super App ()
In China, you can do everything on We Chat. Can ChatGPT solve that puzzle in the USA?
📺 Forget About Plot-Based Media ()
AI-generated video apps like Sora2 work really well for endless scrolls. Who needs plots anymore?
Practical AI
🎶 Vibe Engineering ()
Vibe Coding really rewards and enhances good engineering practices. TDD wins again.
7️⃣ Seven Rules To Stop Overloading Your AI ()
A few reproducible ideas about how you can effectively engineer your context and have more successful experiences with AI.
📓 NotebookLM is Crazy Powerful ()
How to use Google’s NotebookLM as a personal learning system including personalized video flashcards.
📼 Produce Your Own Marketing Videos ()
An effective workflow for making content and avoiding the slop trap.
🤖 Using Agent Builder ()
A quick example on how to use one of OpenAI’s shiny new toys.
Product Thinking
🧠 Maximizers vs Focusers ()
Disciplined Architects and Opportunistic Strategists make a great alliance.
🚘 Charge Less for Lower-Rated Uber Drivers? ()
Is there an effective model for charging less for lower-rated drivers?
Business History
🪟 The Birth of Microsoft ()
How the outcome of a high school rummage sale was Bill Gates and computing as we know it.
💵 Modern Loan Origination (Bits about Money)
Why banks don’t make home improvement loans.
🕰️ The Timetable is a Product of the Railway ()
Railway networks move people, and people need timetables.
Other Interests
🐓 The Story of Cash Money Records ()
How the birdman became one of the richest men in the game. He flies in any weather.
🍄 Did Psychedelics Make Us Human (
)Exploring the theory of the stoned ape.
Did you enjoy reading this post? Hit the ♥ button above or below because it helps more people discover great Substacks like this one and it helps train your algorithm to get you more posts you like. Please share here or in your networks to help us grow!