Most advisors default to ChatGPT for everything, but the AI landscape is much bigger and moves every week. This is a fast tour across twelve categories (frontier chat models, research, image and video generation, browser agents, voice and meetings, presentations, AI email, no-code builders, agents, workflow automation, and scheduling) covering what each one is actually for and which tools are worth knowing right now.
Sign in to view this.
Sign inSign in to view this.
Sign inMost advisors I talk to default to ChatGPT for everything. Not because it's the best tool for the job, but because it's the one they know. The actual landscape is much bigger than that, and it's moving every week.
The intent of this workshop is to give you a fast tour of what's out there, what each category is actually useful for, and where it fits in your day. Not a deep dive on any single tool. If there are any tools you would like me to go deeper on in a future session, send me a message.
These are the everyday "thinking" tools that almost every other AI workflow sits on top of. ChatGPT, Claude, Gemini, and Grok are the four you'll run into, and the choice between them is increasingly about ecosystem (Google vs Microsoft vs standalone) rather than raw quality.
ChatGPT has the largest brand and the most familiar interface. Claude is the one I reach for when I care about quality of writing, long-document analysis or the latest features. Gemini is the obvious pick if your business already lives inside Google Workspace and care about price. Grok is fastest and least filtered, not the most useful but if you need to sound more authentic it can be a good option.
If you actually want to know which model to use for a given task, the place to go is LM Arena. It runs anonymous, blind, head-to-head battles between models on your prompts and computes Elo ratings from millions of votes, the same scoring system used in chess. There are separate leaderboards for text, coding, image, and video, so you can see which model is best at the specific thing you're trying to do based on real human votes, not just an arbitrary task. Worth bookmarking.
Open source is also closing the gap fast. GLM-4.7 from Zhipu and DeepSeek V4 are both genuinely competitive with most well known models now, and they matter if you care about cost, privacy, but require more set up to get all the benefits.
These tools compress the time it takes to get up to speed on a destination, a supplier, a long document, anything you'd otherwise spend an hour Googling. All of the big platforms have their own version of a deep research mode now which you can enable in chat. This is generally what I reach for as the quality is usually better than special tools and I don't need another subscription. But there are some other tools that are interesting in the category.
Perplexity is the fastest way to get a sourced, cited answer, basically a search engine that synthesises instead of returning ten blue links. NotebookLM from Google works on documents you upload (PDFs, contracts, supplier guides), and it can generate a podcast-style audio overview that's surprisingly good for the commute or the drive between meetings.
This category has matured fast. The new models render readable text inside images, hold visual style consistent across multiple shots, and edit existing images while keeping lighting and depth intact. Honestly though, they essentially all do the same thing. You type a prompt, you get an image. The differences show up at the edges.
Midjourney is still the best for adding your own taste and inspiration. Nano Banana Pro (Google's Gemini 3 Image model) is great at readable text, taglines, or layouts. Seedream 4.5 from ByteDance is excellent at editing and compositing across multiple reference images. GPT Image 2.0 launched in late April and instantly took the number one spot on the Image Arena leaderboard by a record margin and is the one I use for everything. Recraft is the only one of these that outputs editable vector files, so it's the right tool for logos and icons.
This is the category that genuinely changed in the last six months. Models can now produce realistic, multi-shot, audio-synced clips from a prompt or an image. The single most useful thing for travel: most of these now accept an image as input. You can ground the scene in a real photo of a destination, a hotel room, or a face, and have the model animate from there. That's what makes them actually usable for travel content, instead of generic stock-looking video.
Veo 3.1 from Google is now free for any Gmail account, which was a watershed moment in the industry. Seedance 2.0 from ByteDance and Kling 3.0 from Kuaishou are the two I've seen the best community results from lately. One of the people in the workshop made a birthday video for a friend with Seedance that was indistinguishable from a real production. Runway Gen-4.5 is the top paid option for cinematic quality. Sora 2 is the model from openAI that got a lot of hype last year but the web app shut down on April 26 and the API is shutting down in September, so probably not the one to build a workflow around right now.
This is the next frontier and the one I'm most excited about for travel specifically. Instead of telling the model what to do in chat, the model takes actions for you, in your browser or on your computer. It clicks, types, fills forms, runs through multi-step workflows.
The reason this category matters more in travel than almost anywhere else: most of the software we use, the supplier portals, booking platforms, consortia tools, doesn't have an API. They give you no way to plug other tools into them. Browser agents are basically the only way external systems can work with them, because the agent just uses the same interface a human would.
The main names are ChatGPT Atlas, Comet from Perplexity, Arc, Strawberry, and Claude Computer Use. My honest take is that you can get away without downloading a new browser at all. The Claude Chrome extension is really good right now, especially since the 4.7 vision update, and it works on top of the browser you already use. Strawberry is the one I'd pick if you specifically want to build lists, because they've added integrations (writing into Google Sheets, for example) that keep the agent from having to figure everything out from scratch.
Browser use can still be unreliable and costly so I wouldn't use it for everything and definitely not for critical workflows without thoughtful setup.
This category turns conversation into structured data. Calls and meetings become transcripts, summaries, and action items without you taking notes during the call. The first thing to know is that Google Meet, Microsoft Teams, and Zoom all ship a built-in AI notetaker now, so for a lot of advisors the question isn't "which tool" but "do I want a dedicated one on top of what's already there."
Granola is the one I use. It records and transcribes without putting an obvious bot in the call, which is way less awkward for client conversations. Cluely is different, it listens during the call and gives you live prompts on the side of the screen, useful for high-stakes calls if you can handle reading mid-conversation. Wispr Flow and Willow Voice are personal dictation tools, hands-free, optimized for typing into things while you're walking or driving. To me Wispr Flow is essential for anything I do
Itinerary decks, supplier proposals, and internal training are the heaviest design surface in an advisor's day, so this is a category worth knowing. Gamma is the fastest end-to-end, you give it a prompt or an outline and it builds a finished deck in under a minute. Claude Design (from Anthropic Labs) has the tightest iteration loop, you refine the deck conversationally and it edits in real time, and it exports to PPTX and Canva. Canva AI is the most familiar because most advisors already have a Canva account but the quality is not great. Pencil is more of a design canvas than a slide builder, it was build for AI and human collaboration so you stay in control without having to do all the work yourself.
For an advisor with hundreds of supplier and client emails a week, the inbox is the single biggest time sink. The AI-native email clients rebuild it around triage, drafting, and summarisation. Superhuman is the premium, keyboard-first option and it'll auto-draft replies in your voice. Shortwave is the more affordable AI-native client, built on top of Gmail, with a Ghostwriter that learns your tone. Fyxer is the cross-platform pick, it works inside Gmail, Outlook, Teams, and Zoom, and it bundles a meeting assistant on top.
This category collapsed the gap between "I have an idea" and "there's software running" from weeks to minutes. For an advisor, the practical use is internal tools, custom landing pages, simple booking calculators, one-off automations, without hiring a developer. None of these require you to actually write code. Cursor and Claude Code are the more developer-leaning ones, very powerful but they assume some technical comfort. Lovable, Base44 (recently acquired by Wix), and Replit are the ones I'd point a non-technical advisor at first. You describe what you want in plain English and you get a working app deployed at a URL.
Agents are the platforms that let you create or use a custom AI assistant. This is where "I wish I had a VA who watches my inbox, drafts itineraries, and chases suppliers" actually becomes possible. Claude Cowork, the one I run a lot of my own automation on, is a desktop agent built for non-developers, it works against your file system and apps and runs scheduled tasks unattended. Manus is fully autonomous and the AI has its own browser to play with. Lindy is the easiest entry point with a visual workflow builder and a generous free tier. is the enterprise pick if you care about compliance and data residency. (from Wordware) is newer and interesting because it's built around learning your preferences over time. is the AI agent you can text over WhatsApp and iMessage and made thousands of people buy mac mini's.
This is the older if-this-then-that world that the AI wave is now eating into. The leading tools have added AI nodes, but the foundation is still the same: connect your apps, define a trigger, define the steps, let it run. Zapier is the largest by far with 7,000+ integrations and now has Zapier Agents on top. n8n is the open-source self-hostable alternative, more powerful but a steeper learning curve. Gumloop is the AI-native one, where LLMs are first-class building blocks in your workflow rather than bolt-ons.
Tools here either help you book meetings or help you defend your time. Calendly and SavvyCal are the booking-link tools, with SavvyCal's calendar overlay being a genuinely better experience for group scheduling. Reclaim and Motion are the AI calendar assistants that auto-schedule tasks and protect focus time, recalculating your day in real time when meetings move. For advisors juggling discovery calls, supplier meetings, and focused itinerary work across time zones, this is one of the easier wins to bank.
That's the wide tour. There's a lot in here, and I don't expect anyone to use most of it. The point is to have the map, so you know what to reach for when something specific comes up, instead of stretching ChatGPT to do every job because it's the only tool in your hand.
My personal recomendations:
Deeper sessions on specific tools and workflows will come in the workshops after this one.
