If it feels like a new AI tool launches every 48 hours, it’s tough to know which ones actually matter for your real estate business.
On last week’s episode of The Walk Thru, Tessabella Jelten joined hosts Jason Cassity and The Broke Agent to walk through three AI platforms they’re testing right now.
Instead of debating the future of AI in abstract terms, they focused on questions that are on a lot of agents’ minds:
- What AI tools are actually worth my time (and money)?
- How are agents using them?
- Could these tools make a real difference in my income this year?
From instant landing pages to automated inbox management to a self-running AI assistant that builds internal systems, the discussion stayed practical.
Let’s start with the first tool Jason put to work.
#1. Kimi AI: Build a Landing Page or 40-Page Mailer in Minutes
The first tool Jason walked through was Kimi AI, a design-focused platform powered by multiple large language models (LLMs) working in the background. His use case was practical and hands-on. He used it to build real marketing assets for his business.
Example #1: Custom Seller Landing Page.
Jason prompted Kimi to create a landing page for homeowners looking to sell quickly. He dropped in a link to his Google Drive with brand assets, colors, fonts, and logos. Within minutes, the platform generated a fully functioning, multi-page website aligned with his branding.
As Jason explained:
“I gave it the prompt, ‘Can you build me a custom landing page for homeowners that are looking to sell their homes quickly?’
“I gave it a link to my Google Drive with my brand assets, and literally like seven seconds later, I had this full-service website. In the matter of, I’d say, five minutes, I was able to get a fully functioning three-page website that people then can go to, submit their information, and get a cash offer on their own.”
Instead of waiting days for a designer, he exported the files and sent them to his web team to plug into his existing site and support a Google PPC campaign.
Example #2: 40-page Luxury Mailer
Jason received a high-end Chase Sapphire Reserve booklet in the mail. He photographed each page, uploaded it as inspiration, and asked Kimi to mimic the format for a San Diego lifestyle mailer featuring local restaurants and highlights.
The result was a polished, graphic-designed 40-page draft that he plans to send to his database after minor edits.
“With this high-end elevated mailer of graphic design that would have taken me or my assistant days probably to create something like this, it took five minutes. I still have to do a couple of edits. I have to put my headshot in there and stuff like that. But I now have basically a 40-page graphic-designed mailer that I’m going to then send off to my printer and send out to my clients.”
So, to recap, here’s what you can do with Kimi AI:
- Test niche landing pages without waiting on a developer
- Create elevated print marketing without hiring a designer
- Move from idea to execution in minutes instead of weeks
The bottleneck shifts from production to implementation. And that’s a very different business model.
Next up, the tool Jason used to get to inbox zero overnight.
2. Twin.so: An AI Inbox Assistant That Runs Every Hour
The second tool Jason demonstrated was Twin.so, an AI agent platform built to take action inside your existing systems. In his case, that meant Gmail.
He went beyond asking for suggestions and authorized Twin.so to log in and actively manage his inbox for him.
Jason set up an agent to:
- Identify spam and newsletters
- Unsubscribe from unwanted emails
- Flag property-specific messages
- Draft replies to emails that required a response
- Archive and label relevant conversations
Then he scheduled it to run automatically every hour.
Here’s the prompt Jason used:
“I’d like to create an agent who patrols my inbox, deletes and unsubscribes from spam and newsletters, and drafts replies to anything that requires a reply from me.”
The result?
“It uses the LLM and it actually drafts the reply, puts it in my draft folder for me, and then unsubscribes from all the newsletters. And I’m at inbox zero now. And I did this last night… and I have it set up to, every hour on the hour, do a scrub.”
For agents, this is about eliminating one of the biggest daily productivity drains in the business.
An AI inbox assistant can:
- Reduce response lag
- Clear out marketing clutter
- Pre-draft routine replies
- Keep deal-related communication organized
Jason noted that tools like this typically run around $20 to $30 per month for mid-tier plans. For many agents, that’s far less than the cost of hiring part-time admin support, and it can reclaim hours each week.
Next, the most advanced tool discussed in the episode, a self-running AI agent that goes beyond inbox cleanup and into system building.
3. Claudebot (OpenClaw): A Self-Running AI AgentThat Can Build Systems for You
The third tool discussed was the most advanced, and the most complex: Claudebot, also referred to as OpenClaw.
This is where the conversation shifted from task automation to true digital assistance. Tessabella Jelten explained that she treats it like a virtual assistant that lives on a separate machine and runs continuously in the background.
She interacts with it through Telegram and has trained it to focus on one primary objective: removing bottlenecks from her life and business.
As she explained:
“The best way that you can look at what Cloudbot can do or OpenClaw… is think about a virtual assistant, especially in real estate. Most of us have all at some point hired a virtual assistant or gone through that process.”
She emphasized that this tool requires more setup and more training than something like Kimmy or Twin. It does not come pre-packaged with guardrails. It learns based on how you communicate with it and what you expect it to take ownership of.
“I spent a lot of time teaching it about me and how I like things to be done. And I continue to teach that. And so whenever I’m speaking to it about something, its core mission is to go and figure that out.”
One of her most practical examples was not flashy at all. It was operational. She runs social media for multiple agents, and formatting content for client review was taking hours every week. Instead of continuing to do it manually, she had Claudebot build a custom Google Chrome extension that automated the formatting inside Google Docs.
“That has easily saved me at least five plus hours a week just in that one solution. And I didn’t do anything. I didn’t know how to code. I didn’t do any of that. It literally did the whole thing for me.”
She also described using it while driving to voice-note every task and idea in her head. The bot organized them into categories:
- tasks it could complete
- tasks that required more input, and
- tasks that still needed her involvement
It researched, asked clarifying questions, and created action plans.
Claudebot can also spawn additional agents for specific roles inside a business. Tessabella has explored dividing responsibilities between human employees and AI agents so that her in-person team focuses on work that requires human judgment and presence.
“Every time you identify something that’s a true role within the company… spawn off a new agent that will solely focus on that and will have different objectives, different personality, different like all of these kinds of things. In theory, it could build an entire organization for you to run a lot of the heavy-lift tasks.”
This is not an entry-level AI tool. It requires technical setup and thoughtful oversight. Tessabella was candid about the learning curve and even the cost of early experimentation.
But her larger takeaway was this: used intentionally, this kind of assistant can reclaim serious time and remove repetitive operational friction.
Start With One Bottleneck and Reclaim the Time
For agents, the practical starting point is identifying one recurring bottleneck and testing whether a tool like this can eliminate it.
In fact, the most important theme from this segment wasn’t how advanced the tech is. It was what agents should actually do with it.
You don’t need us to throw more AI tools in your face. And if you’re feeling mind-boggled by the massive and growing menu of AI tools, you’re in excellent company.
The Broke Agent put that feeling into words perfectly:
“All this does is it gives me anxiety that I haven’t learned any of this, that you guys are already 10, 20, 30 steps in front.”
Tessabella’s advice was grounded and practical. Agents don’t need to master every emerging platform.
They need to solve one time-suck at a time.
“I don’t think you have to be on the forefront of every single new thing that comes out. Think about trying some of the tools like what Jason talked about, that will get you bits of your time and energy back.”
Jason brought it back to leverage. The goal is to offload repetitive, administrative, and production-heavy tasks so you can focus on the parts of the job that still require you.
“You’re not going to get replaced by AI, but you’ll get replaced by someone using AI.”
For most agents, making steady and meaningful progress with these tools likely starts with something simple:
- Automating inbox cleanup
- Building marketing assets faster
- Streamlining repetitive formatting tasks
You don’t need a custom server setup or a digital team of spawned bots to benefit. You need to identify one clear bottleneck and be willing to test a tool that might remove it.
That’s the real opportunity here. Not hype. Not fear. Just better leverage.
