PROPTECH SPOTLIGHT: Opportunities for Agentic AI with new Model Context Protocols

The Model Context Protocol (MCP), created by Anthropic, the developers of Claude AI, represented a significant advancement for AI integration in the real estate industry.

The open protocol helped Claude AI models integrate with third-party applications and was expected to lead to the exponential growth of AI-based tools across all AI systems. As a common “language,” MCP was destined to be a major enabler of Agentic AI.

The protocol had been applied to the Forms Live platform by Dynamic Methods, and Proptech Australia members were invited to hear more about how it had worked and the lessons for the proptech industry.

David Howell, CEO and founder of Dynamic Methods, and Proptech Australia president Kylie Davis hosted an informative webinar about its MCP Server Pilot Program and the opportunities for Agentic AI in real estate. This webinar was aimed at proptech companies and developers who were interested in building AI-enhanced solutions for property management and sales.

Key takeaways from the webinar included:

• Understanding the Model Context Protocol and its application in real estate.

• The opportunities to transform property documentation with AI.

• Participation opportunities in the MCP Server Pilot Program.

Dynamic Methods, a leading Australian proptech company, had been providing forms, contracts, and efficiency tools to the property industry since 2005. With over 8,500 real estate agencies and 50,000 users, Forms Live was generating more than 600,000 contracts a month.

Transcript:

Kylie Davis: Hello, everyone! If you're joining us, it's Kylie Davis from the PropTech Association here, or PropTech Australia, I should say. Sorry. Get my words from great to see you all. David, and I've just been having a very in-depth conversation behind the scenes to make sure I know what we're talking about today.

Kylie Davis: It's great to see you joining the call, please. Just I know. Look, this is prop tech. I know everybody's very quiet, but if you'd like to just say where you're dialing in, from what prop tech you're from, we would love to see it in the chat. It's always nice to know who we're talking to.

Kylie Davis: We'll just give everyone a minute or 2 to get sorted out, and and then we will kick off. We've got a great conversation to be had today.

Kylie Davis: All right. Everybody's very quiet.

Kylie Davis: Project.

Kylie Davis: Hello, Anita!

Kylie Davis: Thanks. Anita's joining us from Adelaide.

Kylie Davis: Nick Greenway. Thanks, hey, Nick. Great to see you guys? Joining the call.

David Howell: Hi Eric.

Kylie Davis: Eric. Hi.

Kylie Davis: thanks for joining in. Well, look! We might kick off Kylie Davis, president of the PropTech Association here with David Howell from dynamic methods and forms live. And we're going to be diving into a very juicy technical chat on Mcp and model context protocols. And and how we can use them, and Agentic AI in real estate.

Kylie Davis: But before we kick off I would like to acknowledge the wangal people of the Eora nation. The land on which I'm on at the moment in Sydney.

Kylie Davis: and recognise the traditional custodians of the land, and pay my respects to elders, past, present, and emerging, and David's dialing in from Adelaide. So any references to time you will see in this webinar, maybe half an hour behind. That's not the AI. But, David, let's kick off. I want you to just kick off by explaining what is a model context protocol.

Kylie Davis: And why do prop techs actually need to know about it?

David Howell: Okay, well, I'll I'll start with the basics and I'll give you a quick demonstration.

David Howell: So the way I explain the way I think about AI. So just to start with, I'm very conservative on AI, but I think that the tools, and particularly around the model context protocol, will enable AI to really be very useful, more useful than it's ever been. And here's a great example, right? So here I'm using Claude at the moment, and I don't have any tools enabled. It's

David Howell: Just Claude. So this is just a large language model. It has no access to anything outside of its own model. So if I ask Claude, what is the time?

David Howell: It's not going to know the time. In some some cases some of them will. Actually give you a time, but it'll be a prediction, and no idea how it's reasoned for that. But in Claude's case it says, I don't have access to real time information right. But if I enable a get time, Mcp. Tool.

David Howell: Just do that now.

David Howell: And I say, what is the time?

David Howell: What it's going to do? It does 2 round trips. 1st of all, it sends a list of tools to Claude and says, Hey, you've got access to all these tools. Claude will come back and say, actually, this tool, I want to use this tool. I feel like this is relevant to the context. You run the tool, locally or remotely, in our case with Mcp. Server. Send the result back to the back to Claude, and then Claude will include that in its response.

David Howell: So you can see here, obviously half an hour behind. But yeah, it is correct. 1235, 26.

Kylie Davis: 1235 download everyone who's on the call, you know, with me. So

Kylie Davis: Glad it's not daylight savings in Brisbane.

David Howell: So let's think so.

David Howell: Large language models by themselves have a limited scope. Then you add knowledge, base on them through rag, and then they have a better, better ability to to reason.

David Howell: But without tools they can't actually do anything.

David Howell: Okay. So a lot of people, probably in this chat would be used to coding with AI, and the term is Vibe coding, where you have another window open, and you'll be asking AI to come up with code, and you would literally copy and paste into your project rather than

David Howell: A gentic AI, where AI can actually autonomously do things with the tools that they've got.

David Howell: So

David Howell: Not all large language models support this, either. So if you, you need to make sure that large language models can support the tools function, not just Mcp, not Mcp, but tools. So things like Gemma from Gemini doesn't actually support tools but all all the major open source and all the major closed. Source ones do.

Kylie Davis: Right? So just give us a little bit of a background on Mcp. It's an anthropic tool. Who are the guys that make Claude? But is it?

Kylie Davis: Does that mean it's locked in? You can only do stuff with Claude, or what's the.

David Howell: No. So so anthropic came out with the model text protocol specification last year.

David Howell: And I kind of monitored it and checked out. I thought, well, it's it's got a lot of potential.

David Howell: And it was when that the models were coming out with tools and the ability to take Json and expect it back, and things like that. But there was no standard protocol for these tools, you know. Each each Llm. Was doing their own thing. Openai was, etc. Now what happened was Mcp. Came out which and the open source community really jumped on board because it gave them a standard that they could work towards.

David Howell: But of course

David Howell: It's a it's a protocol. It's a specification. As a client and a server. So you needed you really needed both. You know you need the servers before you can do the clients. And then, you know, once you've got the service you need the clients. I wasn't sure whether it was going to take off. However, it's particularly this year. All Major AI companies have jumped on board and are now collaborating with anthropic

David Howell: To improve and support the specification. So right now, if you use Gemini, you can use Mcp tools. If you use Chatgpt through their Api, you can use Mcp tools.

David Howell: Obviously, Claude, and there's a whole suite of open source projects, libraries, frameworks which support it as well.

Kylie Davis: So what we're seeing here in real time, I guess, is the development of a well, a protocol or standards, or, you know, guardrails for how to create

Kylie Davis: Tools and code, for that's informing AI and telling it what to do is that.

David Howell: Effectively. Yes, yeah. And then that's the thing. So the way I treat AI is like they're like the apprentice.

David Howell: So they they know a lot of information, but they may not know how to use it. So the advantage of Mcp and that the specification is you can provide a set of tools to AI, with, along with various explanations, prompts, resources, and and another option called roots, where you can give it

David Howell: Kind of the guardrails or the guides on how to use these tools as opposed to just giving it a list of tools and expecting it to work it out. And I think that's the big difference. And that's where the Mcp protocol or the protocol really has the power in that

David Howell: I can build tools

David Howell: or 3rd party developers like the prop tech industry for them to use, and they can have some confidence that the AI will use the right tool at the right time.

David Howell: Yeah. Going back to the apprentice scenario.

Kylie Davis: The rental thing, or whatever.

David Howell: Correct.

Kylie Davis: Yep, okay.

Kylie Davis: Okay, so how? And so, how is, how do you see? Agentic? AI, you know everyone's talking about Agentic. AI. Now, it's kind of the flavor of the month we've moved away from. Just AI. But how do you? How important do you?

Kylie Davis: How do? How important do you see it becoming in the real estate space.

David Howell: Look. I think.

David Howell: As I said earlier, I'm very conservative on AI always have been. I think it's great for things like coming up with documentation marketing plans. You know, things like that where it has a lot of information coding. It's very good. But you know it's kind of the vibe coding that everyone's doing at the moment but the agentic AI and the ability. Now through tools and specifically through data standards like Mcp, you can start to build

David Howell: Autonomous AI that can assist and make things more efficient.

David Howell: So now that's very generic. But think about so.

David Howell: Obviously, we're building these these tools for ourselves as well. So we we are accompanying that dog food, our own

David Howell: Technology. So that, like the.

Kylie Davis: We say, drink your own champagne. We don't say, eat our own dog food. That's terrible.

David Howell: Well, the Api we built for for 3rd party is the same Api we use. And so we're, you know, Api 1st driven, and that goes the same with AI.

David Howell: So we're we're building a forms live assistant.

David Howell: And our primary goal is to make our customers more efficient, which has always been our goal. And

David Howell: We can do that various ways. The 1st way is with these tools that we're building is around

David Howell: Giving AI the ability to analyze and report back to customers on what's happening with their forms.

David Howell: And you couldn't do that. You couldn't do that before. You know, before these tools came out.

Kylie Davis: Right, because I saw an ad for salesforce like, while I was walking down past Barangaroo yesterday. And it basically said, You know that the the salesforce agents is that this is what AI was supposed to be, that these are the little you know bots that run around and do things for us.

Kylie Davis: So what I just want to get clear on. So we've got and and apologies for the really techno guys on the call. If I'm being a little bit slow around it, but we've got we've got the protocol that's giving us the guidelines and the and the sort of standards to write to. But then it has to sit on top of

Kylie Davis: A data set or an Api, I'm assuming of some kind. And so what you guys are proposing, or what you you guys have set up at forms live is that you've you've got the anthropic Mcp sitting on top of your Api. And so what's the benefit to other prop techs in being involved in creating tools around that like, is it is is that, yeah, what's that benefit.

David Howell: Yep, so.

Kylie Davis: And.

David Howell: The the answer to that is, at the moment anyone could build an Mcp. Server on the top of an Api.

David Howell: And I suspect some are doing it, and they're probably doing it through tools as well just through other means. So, for example, through Gemini have different set of tools, and they're consuming our Api already. But using it in an AI scenario. And

David Howell: What I want to do is discourage that. And I'd rather than go through the Mcp server, because what we are doing is we're building endpoints for AI, which will be far more useful for you as opposed to you, having to try and

David Howell: Provide the context to the AI around that some of the Api endpoints we have available. So, for example, you know, with forms, you get a list of forms through our Api, what we're providing through the Mcp server is the same ability.

David Howell: but with more

David Howell: Context around what the data returned is and how it looks and what that means. So then AI can understand that and make a better decision or provide a better experience. So what's in it for

David Howell: The the industry is really, it lowers your risk of

David Howell: Doing something through the forms. Forms. Live platform that you didn't want the AI to do so. We can help.

Kylie Davis: Okay.

David Howell: Build those guardrails you were saying before, to make it easier to implement AI, and with the use of tools, with forms like.

Kylie Davis: So we're kind of. I guess we're kind of at at a

Kylie Davis: observation point that in the past, the way the road the prop tech has gone down is that people have worked out how to do something. And then, in your own prop tech, you've gone off and invented it for your own prop tech. And and therefore we've seen replication of capability created a million times across, or, you know, 165 times across the industry.

Kylie Davis: And so and so what we're kind of recognizing as an industry, too, is that? Well, if we all called our data the same thing, we could actually start to speed up. How we iterate, you know, iterate things.

Kylie Davis: And I guess what you're saying is that we've got. We've got the Mc. We've got the Mc. Protocol sitting at the top. You've got that connected to an Api underneath that. But what you're doing is building that bit in the middle. That is saying, okay, so if you're going to use our Api with these things, this is what the translation layer or whatever. So you use this rather than trying to invent your own translation error, and then come up, maybe with something else.

David Howell: Correct. Correct. And

David Howell: Yes, yeah, that's it. In a nutshell. We're not. We're not telling you what AI to use. We're not telling you what models to use. We're not telling you how to implement things. We're just saying, here's the tools that we can. We can give you. So you can build what you need to build around agentic AI, or autonomous tools and workflows and things like that. So I think people people get very confused. People think AI is workflow.

David Howell: And I isn't.

David Howell: AI is magic. But yeah, oh, yeah.

David Howell: But that's the thing like you just throw a list of tools, and it suddenly knows how to do everything. And that's not the case, you know, like, you need to be very specific and give it a very narrow context on to stop it from getting out of control. You know, you hear about hallucinations and all those types of things. But yeah, so that's

David Howell: We. We want to make that as easy as possible, so that you know the whole point. This I mean, this is why I love prop tech. Australia is, you know, the people in it. They want to do innovative things, and I want to be able to provide the tools to be able to do those things. I don't even know what some of these you know.

David Howell: These tools that we're building. I'm sure

David Howell: There's someone out there who's gonna build a really cool solution that we just have not even thought of.

Kylie Davis: Yeah, and look, I just want to say to everybody who's watching, because I am conscious of my lack of technical expertise in this space because I'm a bit more of a big picture girl. But if you've got any specific questions, or you want to add something to the conversation. Please pop it into the chat, and we'll and we'll definitely incorporate it in. So so how do you? So I just want to go back to David, though. So

Kylie Davis: What are the sort of things that you see a possible

Kylie Davis: By other prop techs using this? You know this translation layer with the forms. What are the sort of things that you see? A possible that would be beneficial to the other prop tech as opposed to forms like, because because at the moment it's like, Oh, we built this and you should build it, because then it's not.

Kylie Davis: I know that what you're trying to do is not sort of say, this will benefit us. It's more that this will benefit the ecosystem, and you can do stuff faster, and I just want to dive into that. If I was another prop tech, what would the benefit to me be of

Kylie Davis: Creating creating code off the of off? What you've what you've built for me.

David Howell: Yeah, so.

Kylie Davis: I mean I.

David Howell: Would hope that everyone in in this webinar is at least playing around with AI

David Howell: And looking at use cases potentially within within their own business or their own solution.

David Howell: And you know, don't. Don't forget Kylie. Really, at the end of the day these tools by themselves.

David Howell: Useless.

Kylie Davis: Hmm.

David Howell: So that you know, unless someone's using them, that they have no value. So this isn't really about. I mean, obviously, I want to be involved in those innovative projects that use cool technology to be able to provide really cool solutions, you know. And and you know, I'm a developer at heart. So that's what I want to do. So

David Howell: Yes, the the selfishness part of it is that we'll be involved in that project, but also the the great part of it is, you don't have to worry about compliance. You don't have to worry about forms, and and you know, and on top of that. We're then giving you these Mcp server tools which allow you to interact with forms live through an AI experience.

David Howell: So yeah, and that could be

David Howell: Like what we're doing with with having a mobile assistant, you know, and think about remote assistance. It's the same. So virtual assistance, the same scenario. Some of those tasks that they're doing are repetitive tasks, and they're they're doing that for efficiency reasons. Some of those will be taken over by gent AI, you know it's inevitable. So but.

David Howell: For example, form 5 assistant will in the morning, when you, when you speak to it, it will tell you.

David Howell: Give you a kind of a headline summary of what's happened over the last 24 h, you know. And what forms are out for signing and you know. Do you want to follow up? Well, let's say there's a there's a form out for signing for Anita, and that form has been that I can analyze that and say, Hey, this is a rental agreement. It hasn't been signed. Most rental agreements get signed within 24 h. Perhaps you should give Anita a call.

David Howell: So you can see that. And and AI can do that in, you know, a matter of milliseconds

David Howell: And really improve your efficiency. So it's not about

David Howell: Making you redundant or removing. That's about improving, improving the experience.

Kylie Davis: Yeah. So I mean, we're getting to a place I'm assuming in the very near future where the ability for a sales agent to simply start talking to their phone and say, You know, tell me what I've got. Not not just what have I got on, Siri? What's in my calendar? But actually, you know, how many contracts have I got outstanding, you know, who's overdue in terms of signing like? What do I need to pay attention to? And I guess we've seen that in AI in some degrees of like. Well, here are the leads you should focus on.

Kylie Davis: But you have to log into the leads platform to do it. But I guess we're going to be seeing

Kylie Davis: The and that.

David Howell: That's that's.

Kylie Davis: Turning into major assistance sections which are pulling a gent from lots of different parts of a real estate agents.

David Howell: Yeah.

Kylie Davis: Platform.

David Howell: Absolutely. And and I hope there's other prop techs out there who will adopt the protocol and provide tools as well, because once you start combining them. Exactly, that you have a very powerful potential

David Howell: To do to change the experience. And I think I agree. I believe long term, the user experience will change dramatically, and that most interactions will be via

David Howell: Verbal and and audio and things like that, while you're, you know, rather than having to, you know, like you said your assistant says, Hey.

David Howell: Do you want me to send this form to Kylie to sign, you know, and it's like, Yes, please. And so it just goes into it because because it knows how to do it.

Kylie Davis: Yeah, cause we do still have a really disjointed experience with AI, I mean, I mean, and I know it's very early days, like, you know there's the gemini inside my my, my forms, or my, you know document folder, which is different to the one inside a document which is different to going into notebook. Lm, which is different.

Kylie Davis: And I love how sometimes you'll be, and you you forget, because they're all called Gemini. So you go and do something, and it'll go. Oh, no, I can't do that, and you're like.

Kylie Davis: Yep, you did it yesterday. Oh, no, that was the other one. So have you guys.

David Howell: Got a Eric's just asked a question. It's an interesting one. So sorry we're going off script to be so basically.

David Howell: You heard of a scenario recently where a company like Uber release tools to book a ride which allowed, you know. Hey, siri? You know book an uber on. Sorry wasn't talking to you.

Kylie Davis: Sure. Shimmy, yeah.

David Howell: Do you think the Uber business model as users no longer?

David Howell: Look, the the big thing is what's happening with search right now.

David Howell: Everyone's using AI for search. You know, web search has changed dramatically or is changing dramatically, and for that exact reason that you're interacting in a different way. So is the risk of losing eyeballs on our products, or and it's something to worry about. That's a very good question, Eric. That's a very good question. I don't have a lot of consumer eyeballs on my product, so I probably don't need to worry about it. But there are certainly others out there that do I?

David Howell: Yeah, I I think.

Kylie Davis: But that's a paradigm shift, isn't it? Because, yeah, yeah.

Kylie Davis: Measure eyeballs, we measure eyeballs on our sites or on our products, because we because that is the way

Kylie Davis: Using, you know, version, whatever that we recognized how to

Kylie Davis: Identify. Was this the right, you know, was this valid or interesting? And so if it was valid, if it was interesting, was it valid? And if it was valid, did it answer the question? And if it answered the question like, you know, that's where we move people down the funnel. But we're getting to a point where, with AI where we'll look. The question got answered or the task got done. So does it matter about the eyeballs like this number of people engaged on the AI. This number of people got the answers they wanted and did the thing that they needed to do. So it's almost like.

Kylie Davis: Yep, beating or compressing the whole thing quite a lot quite quickly. Yeah.

David Howell: Well, I mean, look, listing portals are going to have to

David Howell: Embrace it, because I assume you would. You've probably already starting to see a lot of AI scraping. And

David Howell: Yeah, in my opinion, you should be embracing things like this because you could use AI. You know, you could have a consumer who's using AI to search for property. You know I want to find a property in Adelaide that, you know, has these characteristics, and it uses a tool to find it definitely. It's definitely on the cards.

Kylie Davis: So what's can you give an example of a specific workflow? I'm kind of just envisioning suddenly this future where I basically say to AI, you know, find me, you know the best properties that are 3 bedroom townhouses.

Kylie Davis: Close to the water, and it tells me where they are from all of the portals not just based on my personal preference. And it's not just not just the red team or the green team. So you've just really you've recently introduced Mcp to forms. Live. What? What's it allowing you to do that you couldn't do before before. So

Kylie Davis: Tell the time.

Kylie Davis: Yeah. So look.

David Howell: Thank you

David Howell: Whether I can give another demonstration around forms. But my, so we've got a pilot group and anyone who who wants to join. Please reach out to me on Linkedin, and and you can join up. It will give you staging access or sandbox account to our forms. I've platform along with the ability to connect to

David Howell: Our Mcp server.

David Howell: So you can play around and and and try those tools out for yourself.

David Howell: Now. What we're doing. Our 1st stage is really around read-only type tools

David Howell: And more around the metadata of the forms as opposed to the content of the forms. Now, the biggest fear. And one of our questions we were talking about. Kylie was around. Fear, you know. What do people fear?

David Howell: Everyone's everyone fears AI. Everyone fears about data. You know that there's going, you know, are you using data to train? Are you using data? This? Can it scrape all this? And you know what we're doing is making sure that we're very

David Howell: Safe and secure around that. So having a sandbox account

David Howell: And having being able to play around with these tools, which, like I said, I read only at this stage. So if I were to jump into my demonstration.

David Howell: Just gonna turn on the tools.

David Howell: So I've only got a a few tools here. I've got. Get templates, get forms by address, get forms by agent, get reminding, signing status and get form by name. So I'm just gonna ask it a question.

David Howell: I'm gonna say, do I have any forms?

David Howell: 4, 48.

David Howell: Gordon rides Park

David Howell: Now in in

David Howell: In a real workflow scenario. It's probably going to be telling you what's going on and asking you what it wants to, what you want it to do more than anything else. Of course.

Kylie Davis: Yeah.

David Howell: Change the hockey.

Kylie Davis: The joy of a live demo.

David Howell: Yeah, let's try again.

Kylie Davis: Apart from what have been the biggest

Kylie Davis: Did you type it right? Didn't put a typo in.

David Howell: There we go. So it's saying it knows it wants to use this. This tool get forms by address.

David Howell: So yes, I found a form for 48 Goner Road, Clarence Park. It's a sales authority created by David. It's been sent for signing.

David Howell: It's currently waiting for Anita and myself.

David Howell: I might ask Anita if she can sign it quickly.

Kylie Davis: So so the tool that you use there was the tool of get address, or get form by address.

Kylie Davis: Form by address, so that we provide that tool, you, too.

David Howell: To to via the Mcp server. So it knows the AI knows to successfully call that function. It needs this information, and this information is gathered by the user, it may already have it, or it might actually ask for more information.

David Howell: I need to just sign that form for me. So.

Kylie Davis: Good evening.

David Howell: Oh, sorry! Can we check again.

Kylie Davis: Yep, still asking permission. We like that.

David Howell: It's gonna try. And this will be interesting whether

David Howell: We're gonna work on that one.

David Howell: Oh, there we go.

Kylie Davis: There we go. It's a live demo. Folks.

David Howell: Yeah, so and it's now waiting for me. Not yet signed it.

Kylie Davis: Yeah, fantastic. Okay, so very, very simple.

Kylie Davis: Yeah. So what were the challenges that you like? What were the biggest challenges that you had in introducing it, or setting it up.

David Howell: So, as everyone would probably agree, the biggest challenges we have around these types of tools is around security.

David Howell: So authentication, authorization, and all those types of things. So that is something that we're

David Howell: Yeah, obviously very serious about. So that's why we're offering a sandbox account only at this stage.

David Howell: But

David Howell: Yeah, that as usual, that's where all the time gets taken to. To make sure everything is safe and secure.

Kylie Davis: Yeah, okay, so so how are you going to support the projects that are being involved in it? Like, what? If if people want iterated within your sandbox. What's what are you? How are you going to work with them?

David Howell: So basically, the 1st thing as you join up, you'll get a survey and would ask how you plan on using the tools. Briefly, or what what experience you have with AI like, I said. Then you get set up with a sandbox account and the ability to to jump in and get started straight away. So there's nothing stopping you from there. Then we're really looking for constant feedback on

David Howell: What other tools you want to be able to use what you know. What's the use case that you want? How can we help you? Make those tools more useful for your scenarios, because at the end of the day, if that if if we do a good job from that point of view, then we'll have a good experience. On

David Howell: Both ends of it from a developer's perspective and also from a a a consumer. Customer. Perspective.

Kylie Davis: Because I can imagine that this would be quite valuable in terms of

Kylie Davis: Crm reporting, or in terms like, or just reporting how, how, what agents and principals or Pm's are seeing on a daily basis. But then the step after that, then becomes okay. So, reporting on this, now, you need to do that. And this, is the.

David Howell: And that's exactly it, like like AI right now, is very good at analyzing. It's good at summarizing.

Kylie Davis: Telling me.

David Howell: Not terribly good at math, but it's getting better at it, so that there's all these scenarios which are currently very good at. And so you combine that with the use of tools, and it just changes changes it completely. You know. And yeah, it.

David Howell: For example, you know, we're looking. We're looking internally at our our dev team. We're looking at autonomous tools to be able to analyze incoming support tickets. See if there's any an issue and be able to lodge a ticket into our repository.

David Howell: Things like that which just help us give us efficiency gains like, I said. It's very good at that type of role at the moment. So we can do it. And we can do that all on premises.

David Howell: So you don't have to. The data is not leaving.

David Howell: Your premises. It's not leaving your computer. If in those, in some cases and

David Howell: Yeah, you're still getting that efficiency gain.

David Howell: I think, Salesforce said. They're not. They're not hiring any more developers

David Howell: Because they've got a 30% performance gain out of them by vibe coding and and using AI.

Kylie Davis: That's interesting.

Kylie Davis: Wow.

Kylie Davis: So I so, Rob, I saw that you put your hand up so I don't know if that was a that was a mishit. But if anyone's got any questions or any comments that they want to add into the add into the chat, please let us know.

Kylie Davis: David, do you think this is something?

Kylie Davis: Do you think this is something that's only for smaller prop techs who are looking to iterate or like? Do you see that it's something that big end of town could also be in, interested in.

David Howell: Look. I hope everyone's interested to me. This is a level playing field. I think.

David Howell: If you're small and nimble, you probably have a bit of an advantage at the moment. The larger you know, the the

David Howell: Larger teams have a bit more

David Howell: Paperwork to get through. You know every again. Everyone's kinda afraid of AI from, you know, a legal perspective and things like that. So and and yes, you know, it is certainly something we need to consider.

David Howell: But at the same time.

David Howell: You know, we we should be, whether we, whether you well, whether you like it or not, AI is going to be a big part of

David Howell: A tech industry and you need to embrace it.

David Howell: And you know, why not make use of it? So it makes everything easier. So yeah, it's.

Kylie Davis: You, too, that we're using this early stage to set us try to come up with the with the, with the protocols that actually get us all

Kylie Davis: Pulling in one direction rather than trying to. You know, a million different variations of the of the technology. So what's your vision for the future of Agentic AI, and and how this will work with with forms live, and how do you see it impacting the industry.

David Howell: Well, I look, I think, what Eric kind of touched on was is spot on in that

David Howell: It's gonna be less about eyeballs and more about interactions. And those interactions are probably from AI, but still driven from customers.

David Howell: So you know, in in one way.

David Howell: A customer may never use forms live in the, in the long future in a long term future, providing compliant forms and all those the workflow around. You know. For example, today I did a webinar for New South Wales around the termination notice. It's become very complicated. It's a real pain, but you have to do it, and you have to do it this particular way. So.

David Howell: But I can see that long, you know, and the vision is

David Howell: That customers won't be actually doing those things. It'll be driven by AI, but in a actioned by customers. So to me, AI is a customer experience or or a user experience. So it's a ux solution as opposed to just you know it as opposed to it, being

David Howell: Something that

David Howell: Can come up with, you know. Come up with marketing plans and things like that. It's it's it's a way to interact with with.

David Howell: Your device.

Kylie Davis: And I guess one of the biggest complaints we have as an industry at the moment, not just in prop tech, but in all areas of our lives. Right? Is that too many tabs are open on my computer for me to do. I need to understand all of these different programs and platforms in order to go about my day to day job. But I guess we're starting to move into that space where the number of

Kylie Davis: Pages I have open starts to go down, because I've just really got my assistant there, and my assistant knows well my my agentic AI knows which

Kylie Davis: Pages or platforms need to be accessed, and and what tasks need to be done, based on the request.

David Howell: Yes, I think I think you're right. I don't know if anyone uses chat, gpt, and chats to them in the morning when you're driving to work, or or things like that, but you can have conversations with it. And it

David Howell: It sounds it's very fluid. And you know you combine that with

David Howell: Being able to do things, it becomes.

David Howell: It's yeah. You're you're not. You're not. You don't need to be in front of a screen. You don't need to be even looking at your phone. You know your.

Kylie Davis: Moving into Star Trek.

David Howell: Yeah, glasses. And I can talk to. I can talk to Meta. And I can ask him, you know all these things. So

David Howell: Yeah, it's definitely going to be interesting.

Kylie Davis: Oh, we do have another question. Should we catalog our Mcp tools as an industry? So we can connect our tools together better. And how can we design them to be more interoperable, and make them speak the same language?

David Howell: Another great question from Eric. Thank you.

David Howell: Yes. Look that. We obviously think very similarly, because that is something that I've been wanting to do is have a catalogue of Australian Mcp. Servers. There is. There is a number of websites out there that

David Howell: Do have all sorts of, you know, Mcp. Servers that you can use most of them around devtools, you know. So Github and Gitlab, and things like that. But there's.

David Howell: I think, but with the advantage of Mtp. Is it kind of creates that standard? So the interoperability

David Howell: And to a degree and so the AI helps

David Howell: You speak the same language anyway. So if you adopt the standard, I think it will help

David Howell: Solve that problem anyway. But I'll be very interested to know I mean, has anyone? I don't know if anyone in in this webinar have have. Are they looking at creating Mcp servers?

David Howell: Have they used any client tools yet?

Kylie Davis: Great question for the group we've had. Jeff and Robert both put their hands up, but I don't

Kylie Davis: Haven't had any questions come through.

David Howell: But look at so it's.

Kylie Davis: How do we reach out to? You? Have the guys reach out to you. If they've got questions, follow up.

David Howell: Please just hit me up on Linkedin. I think it's probably the easiest way, and I'll get you on the pilot if you want to be, or happy to chat about this

David Howell: Anytime like I, said I.

David Howell: Until recently I was very conservative on AI, and I think this is going to be the turning point for

David Howell: Useful development of of AI solutions.

David Howell: And yeah, I would love for the prop tech industry and prop tech to Australia to kind of help. Drive it, you know, and

David Howell: Not put things behind paywalls, and, you know, allow the innovation to happen as opposed to trying to control things.

Kylie Davis: Yeah, and and another apart from the apart from well, apart from the privacy, it's a big one. But are there other risks that you see, see that we need to be mindful of.

David Howell: Look, it's it's probably it's, in my opinion, is no different to say providing an Api. You know you're you're you are giving the ability for someone to interact with your platform. You, you know there are things you need to make sure you've got working very well out of box like throttling Api throttling, because AI can do things a lot faster than than humans. So you don't want them to, you know, accidentally hit an endpoint a million times, which you know, in case costs and things like that.

David Howell: But outside that

David Howell: I believe that that it is very similar to the Api scenario. So you know the usual things. And this, which is why we have a pilot to start with, as well to make sure that we're all using it for the right purposes and the right reasons.

Kylie Davis: Cool. Well, guys, thank you so much if we haven't got any other questions or or comments. Thanks for a great conversation, David. Using the Mcp or the Mc protocol, I'm conscious. If I say, Mcp, protocol.

Kylie Davis: Yeah, yeah.

Kylie Davis: Utology. Yeah.

David Howell: I'm.

Kylie Davis: But coming up with some common well, having some common standards and guidelines around, how we're using it in this space, having that translation layer sitting over the top of an Api. And if you're looking for a place to sort of start to experiment with this.

Kylie Davis: How big, how big is forms live, David and and Rei forms. How how big is your market? Share that? And the number of forms that you've got going through your system.

David Howell: Nationally, we're doing around about 800,000 forms a month at the moment.

Kylie Davis: Okay.

David Howell: And that's still growing. And I think we're at between 7 and 8,000 agencies using using our platform. So

David Howell: Yeah, a lot.

Kylie Davis: That's a hell of a lot. So you've got a good data. So you've got a good Api data set that that's

Kylie Davis: Sitting.

David Howell: I think there's a lot of potential around Crm integrations. Crm, you know, things like that. And and obviously with the with the listing portals as well. How we could really change fundamentally change, how our customers use our platforms.

Kylie Davis: Well, any kind of workflow automation that's requiring a form or a contract right?

David Howell: Yeah, absolutely. I mean any, any, any interaction, in my opinion. There's

David Howell: There is potential for it. Which ones are useful, or which ones give the most efficiency, I guess, is the question. But yeah, obviously, with forms, forms are a pain in the ass. And that's why we we exist, we try and make it as easy. Yes, we make it as easy as possible, so that you don't have to worry about, and that when I say you I mean the prop tech.

Kylie Davis: Come on!

David Howell: In industry, because and that's you know, that's I really enjoy

David Howell: Being a part of the prop tech industry and prop tech Australia and talking to like-minded people.

Kylie Davis: Awesome. Well, look, guys, if you've got any further questions, reach out to David and thank you so much for being part of our prop tech spotlight. We are going to be doing more of these. So if anyone on the call has got an idea around, you know, this is where we can really geek out on the on the technical stuff. It's where we can discuss issues around, you know, data. AI Apis privacy. Aml.

Kylie Davis: You know, cyber security, all the sorts of fun things that are going on at the moment. So if you've got a topic or an issue that you really want to see us discussed, and and you want to bring some stuff back and share it with other people in the industry. Please reach out to me. That would be great.

Kylie Davis: Thank you so much for everybody on the call. Thanks, David, and thanks to dynamic methods and forms, live, and looking forward to seeing everyone at the prop tech awards. On the 3rd of July we will be announcing finalists tomorrow and looking forward to that. See, you guys, thank you.

David Howell: Thanks. Kylie, thanks. Everyone.