And welcome to the Tray.ai session here at the AI Deployment Summit where we'll learn and see just how powerful and far-reaching AI is set to become for your entire enterprise across integration, infrastructure, and operations for SaaS and process apps. We have two great speakers with us this morning, Paul Turner, integration strategist, and Luke Smith, sales engineer. Gentlemen, welcome.

Great to be here, Vance.

Thank you, Vance. It's great to be here.

We're really glad to have Paul and Luke this morning. Paul advises enterprise customers on integration strategy and the best way to adopt AI for those tasks. He has more than twenty years' experience in SaaS, integration, and analytics. And Luke, for his part, works a lot with customers directly these days, focusing on making automation and AI easy to adopt for both IT and knowledge workers.

So we're really glad to have both these gentlemen here with us this morning to talk about rethinking integration in the age of AI. Today, Paul and Luke will give us a hands-on view of what's coming in the world where AI meets integration. First, we'll learn about AI-augmented development, where as many as seventy-five percent of projects will use AI coding assistance in just the next few years. We'll see how AI is also helping with integration modernization, process mining and discovery, and even operations and governance.

And to bring all these possibilities to life, we're gonna get some really great demos. So great speakers, great content. One last word: you can download today's slides. Just click the big red button under the view screen.

You'll also see some of the available downloads and even access to a free trial, all for you, just one click away. And we'd like to have you interact with our speakers. So to communicate a comment or question, just type it in the "Submit a Question" box right under the view screen. So with that, guys, let me turn to you and tell us about how we should be rethinking integration in the age of AI.

Right. Thank you very much, Vance.

So let's jump right into it. If your role is in — gosh — application integration, data integration, I think you'll enjoy what we've got for you today. So just a few things you're gonna learn: if you're a practitioner, I'm gonna share some really unexpected ways you can use AI in your role. We're gonna fast forward into the future to see what the AI wave will look like for integration practitioners, so you can plan accordingly.

And, you know, Luke has got a real nice demo, actually, where you're gonna see how you can use AI within your daily integration projects and how you can use it to solve new emerging requirements around infusing AI into processes.

You're gonna get some real thought starters from that demo as well. So hopefully you're gonna walk away with some real, practical ways to start thinking about adding AI into your integration life cycle and projects.

Okay. So often I'm asked: what does AI really mean for integration?

And when I speak with integration leaders at all kinds of organizations, the impact of AI really falls into three camps, and this cuts across everything from generative AI to classical machine learning.

The first is really providing augmented development — so a co-pilot, no-code development. This is really about assisting existing integration builders with all kinds of low-hanging, repetitive development and post-deployment maintenance work. You know, many integration builders — we all know, right? — we spend a lot of time on repetitive work: dragging connectors onto a canvas, defining the same initial logic for triggers and branches, the same optimization steps.

So really, the point of AI augmentation is to free you up from what I call "integration drudgery," so you as a developer can be more productive. And, you know, Luke's gonna drill in and smash some examples in the demo.

Gartner predicts about seventy-five percent of software engineers will use AI coding assistance by 2028, up from about ten percent early last year. We're already seeing really early benefits of using AI for augmented development. You know, we're seeing some data points come in from other parts of the industry — GitLab, Microsoft, ServiceNow.

The productivity boost for developers is starting to range from about thirty percent to fifty percent. So this is really about augmenting your team and freeing you up to focus on other activities.

And as large language models improve in reliability and consistency, you're likely to see this improve even further. So it's a really good time to start thinking about and getting more comfortable with augmented development right now

But to be honest, the productivity boost you can increasingly look for using augmented development is really an enabler to free up your time so you can be more prepared for the next wave, which is really adding generative AI and classical machine learning — so predictive and prescriptive machine learning — to your current business processes so they are smarter and more dynamic.

The business teams you support right now — whether marketing, sales, or finance — are looking at the processes they're running and asking, "How could they be better if they could integrate different kinds of AI into them?" And we're gonna explore those in a moment.

And it's not just about adding AI to your current processes — it's about opening up entirely new opportunities. Whether you're a sales operations team wanting to roll out a retrieval-augmented AI-based sales chatbot to handle common sales support tasks, or maybe your HR team wants to launch a whole new employee sentiment analysis program — there are whole new sets of tasks and projects coming down the pipe that really have AI at the core.

So you can really think about it as reinventing existing processes and handling entirely new requirements. And I really think this is the fun bit around AI requirements.

So let's start with augmented proposal development and what I call the new integration opportunity.

So this is the opportunity to really speed your development, really allocate your time, enable more talent in your organization to build, and maybe who want experts in integration.

So you can use Gen AI to help boost a whole range of development and life cycle activities from using no code to create workflows.

You can use it to help frame logic for API development. You can use it to assist in debugging. They even provide contextual recommendations around optimization.

And, also, use it to help describe what your integrations are doing for new developers too. For example, if you haven’t documented, you can use AI code pilots to assist with that.

If you provide a no-code natural language experience, that really helps with democratization as well. So that there are folks that can start building that maybe aren’t familiar with more complex integration tools. And so that helps reduce integration bottlenecks and takes the pressure off your team.

Ideally, you want to get to a point of code, low code, no code. Teams building business logic together using AI to augment their process, sharing artifacts, reassembling across some multiple business teams that could work together.

So let's kind of get tactical. And so I just want to give you some examples of where the state of the art is in augmented development.

So on the right hand, you know, you can see a traditional low code integration flow that you can build using drag and drop. But by using a Copilot, right, in Tray we call this Merlin, you can, for example, use natural language. So here we see create a new Google Sheet and get a name and email from Salesforce and add them to sheets. And in this case, the Merlin co Copilot will lay that initial workflow down for you, and Luke's gonna cover that in the demo. Now, you know, of course, you're gonna need to run through and configure your parameters and get over the finish line, but, really, it's all about, you know, getting the initial framing done to help you focus on what matters. In a cloud environment, your business team might start with no code building out using Merlin, and then you might switch to low code to bring it over the finish line and maybe even incorporate some code-defined lines as well. So it’s really about a combined kind of blended experience.

So, you know, another example here is explainability. Everyone's talking about automation scale, you know, hundreds of integrations, automations, and microservices.

But how do you ensure future teams can manage, you know, change management? So, well, you can use a Copilot here, for example, to explain how your integrations work. So here we're asking, you know, what is this workflow doing? Which is really important if your team is engaging together and people are jumping in on workflows. You can use AI to aid in explainability and narratives.

Performance as well. Things like parallel processing and figuring out concurrency integrations can be a real head scratcher. But you can use, for example, GenAI for, you know, pure performance recommendations.

But here, we can see, for example, Merlin analyzing a flow and providing recommendations on how to parallelize and how to improve the performance.

And you can also use a Copilot model to support data integration activities as well. So everything from writing code to insert into a workflow, writing regular expressions, inserting data into new services. So here we see an example of that. Merlin here has added a semi-configured connector to an existing workflow and highlights that the builder must finalize some parameters. So, you know, the thing I would emphasize, we've covered things like performance optimization, laying initial framing for a workflow, narratives.

Here, we're doing a DEI task, but Merlin's not gonna do all the work for you. Right? You know, GenAI is not gonna do all the work for your building, but it's gonna help facilitate your development process. It’s like an extra member on your team, right, that’s helping you do your work

You know, it also supports operational tasks as well, what I call life cycle management. So things here, for example, maybe you want to write a cron command for scheduling your integration. And here, you can use a Copilot to help build that cron expression for scheduling your flow. So lots of different applications at different parts of the development process.

Okay. So let's jump to the second part here. So the second is infusing process with AI services. So this isn't about making your development work more efficient. This is about making the process you build and manage more intelligent. So you're gonna start seeing more of these AI services coming down the pipe in your business teams. So the key is to make it easy and maintainable to consume AI services within your processes.

As you can imagine, your business team support is likely trying to figure out how to boost their productivity and performance by AIifying their processes.

So if you think about many integrations you have, there is a huge opportunity to upgrade them. So, you know, here I just listed out all the different integrations that we see across the organization.

But for example, you know, you can automate, you know, extraction and classification, maybe tapping into Google Vertex to maybe upgrade and automate document-heavy processes like order entry. Or you might want to add predictive analytics. You may be consuming AWS SageMaker to add predictive decision making to services and service and escalation tasks as well. Or take your customer renewals process and add sentiment-based conditional operations to it as well.

Or, for example, do your lead import process, incorporate a large language model to help with data enrichment as well. So there's lots of different opportunities. Right? Providing with iPaaS can reach out to their services.

There's lots of different ways to upgrade your existing processes with AI.

So, for example, here, we can see what we call the AIification of the existing process. You see a Tray workflow on the left, and we're using OpenAI calling out as part of the workflow to aid with text extraction.

You might, for example, use other services like Anthropic or maybe Google Vertex, IDP, Telstra document processing, or Amazon Textract to extract data, all incorporated into your workflow. And what we do at Tray is actually we provide a template library and a growing set of templates that show you how to infuse AI into those processes.

So this is another exciting area, which is whole new opportunities that GenAI opens up. So just providing some examples here, and these actually have pulled these in from Gartner that has some coverage on new emerging services.

When I think about, for example, using retrieval-augmented generation, your internal sales materials, and a large language model. So, for example, your sales operations team can deploy a sales assistant, for example, or automate your competitive intelligence work using classification sentiment. So automatically pulling data from Salesforce and then using an ALM to help form, say, sentiment analysis and classification on your competitive. Or your accounts receivable team might want to automate customer outreach using personalized collections emails, HR employee self-service, integrating chat with HR back-office apps to help with an HR chatbot experience.

We're seeing even ecommerce, for example, using Gen AI services and OpenAI Vision to automate running product descriptions and pushing to your site. So the point here is that AI opens up a whole brand new set of integration and automation opportunities that you really couldn't accomplish before.

Now as we all know that new large language models are emerging all the time. So what you integrate with this year might not necessarily be the language models or speech models or vision text, predictive models you want to integrate with next year. So you really want to make sure that the iPaaS you use can plug and play easily at the API level with all those different services so you can keep your integrations flowing with the latest innovations in AI.

So looking forward, you can see really the next area here, which is continuous process improvement. And this is the really the data exhaust your process is throwing off combined with predictive and prescriptive AI, so recommendations on where and how to improve processes.

So for example, identifying why your order process is slower in a particular region, such as slow and average approvals, and provide a recommendation on how to reduce approval latency, such as adding a reminder. So it's really about moving from static processes, what we're using today, to much more dynamic processes, tomorrow. You know, Gartner calls this autonomous operations, and it's really moving much more toward a self-adapting organization. Right? Static to much more active processes using data, intelligence to continuously improve processes using machine learning.

One more thing when thinking about incorporating AI, whether it's new development processes or new business processes, it's really critical to think about enterprise foundation, trust, elasticity, and flexibility.

So it's really important to have, you know, strong centralized controls around how and who can incorporate AI, controls around the flow of data to and from AI services for security and privacy, have strong SDLC processes, audit trails, and observability. So as you grow your velocity and you grow your deployment, you have all the controls you need.

At Tray, we call this our enterprise core. So, you know, all of the integration, automation, and AI integration is all built on basically a foundation of governance, instrumentation, service safe scalability, and, as well as security.

So just gonna, like, kind of summarize really how AI changes integration. So if you think old way, the new way, traditional iPaaS, and now integration there together, lots of avenues to change to how we build using AI, augmented development, who can build using no code, the ease of which we can add AI services and intelligence to our processes, and really shifting towards processes that are much more adaptive and responsive to the business.

You have some keys to success, really about providing role-based experiences to your team so they can build integration automation how they want to, code first, low code, no code, enabling a team to build a unified set of capabilities, process automation, data integration, API management, connectivity to your stack, as well as connectivity to AI services, ensuring that you have augmented development, you know, some of the areas I covered with Merlin in here, and, obviously, ensuring it's all on a strong governance and manageable foundation.

So with that, I'm gonna hand it over to Luke, who is gonna jump into how you can use AI and integration together for both augmenting your development and infusing AI into your processes. So over to you, Luke.

Thanks, Paul. So let's jump in and take a look at how we can bring this to life in a live demonstration.

So I'll run through a couple of use cases and how we can make use of AI and what we at Tray refer to as our Merlin AI layer in some of those use cases to help. So the first one that we're gonna start with is gonna be all around that augmented development where I've got, let's say, quite a complex integration that's been created on the platform. And I'm interested to just get a high-level summary of effectively what that workflow and integration is doing. Maybe I'm someone new to the team, or perhaps I just wanna generate a nice summary that I could use in something like some documentation or just sharing with a colleague of mine.

So what I can do is you can see that I've got a workflow here that's got a fair few steps going through it. I'm interested in just getting a summary of effectively what this is achieving. I'm gonna go ahead and click on Merlin over on the left-hand side, and I get a prompt request that will pop up on the left. So I'm gonna ask it quite simply to summarize this workflow.

And so what will happen here is Merlin will be put into kind of a read-only mode. There are circumstances and prompts that we'll touch on a little bit later where we are gonna actually be adjusting the workflow directly. And so that just means that if we do need to make any changes to the workflow, there's not gonna be any conflicts from you changing something to Merlin trying to change it as well while it is going through and processing that prompt. But we'll give it a few seconds here for Merlin to kind of take that prompt, analyze the workflow, and come back with a suitable response for us.

Amazing. So we can see that Merlin's come back here with the response for what this workflow is doing, and we can see that it's giving us a nice breakdown. So we can see that we've got, you know, this workflow appears to be designed to process a large amount of data from a CSV file, which is great. And then it's gone down and break down the main steps that come into play here for each of those different steps that we've got with a brief description of what that operation is doing, which is really, really handy.

We can see it also gives us a nice overview at the bottom here about getting the downloading the file from Google Drive and processing that CSV data, but also includes the mention of the pagination that we have built into this to handle the large amount of data that we're getting. So really nice to be able to kind of get that summary overview of what this workflow is doing, particularly helpful when you've got a large amount of complexity. Really just helps when you are building in that experience, when you are on the builder, to help just effectively help you understand different parts of that and generate good descriptions of that as well.

But let's now switch across and look at a secondary example. So in this case, I've got what we refer to as a scheduled trigger. Now this allows me to run an integration on a period of time or schedule that I'm interested in. But the general idea here is that I have the ability as part of my scheduled trigger to set up what is referred to as a cron expression.

Now these cron expressions can allow you to get really granular with when this runs. So for example, I might want it to run at 2 PM on Tuesdays, for example, and that is the schedule that I would like to run. And so a cron expression will allow me to actually have that structure. But let's say I'm someone who isn't familiar with cron expressions and I'm not sure how to write it.

I could also make use of Merlin here to help with that. So again, I'm just gonna go ahead and click on Merlin here. And so what I'm gonna do is I'm gonna ask Merlin here to write me the expression that will run a scheduled workflow every Tuesday at 2 PM.

And so, hopefully, Merlin will go away and return back that cron expression that you can use. But this is a really common scenario where you might be dealing with quite intricate or complex parts of your integration design. And in this case, it wasn't like a cron expression for the schedule, but you might be, for example, interacting with a database and you're not quite sure how to enable or write the SQL query that you would like. And so this having this extra layer directly within the builder really helps with that development experience.

You can see that here it has hopefully provided back to me the cron expression that I can use. So I could just go ahead and copy that and pop that into the cron expression portion of the builder on the right-hand side. What's also really cool is that Merlin actually then broke down exactly what this was doing as well as what each of these numbers do or mean, so not only am I just getting the outcome that I'm interested in, but also, you know, a breakdown of effectively what is happening so that I can start to understand that in a little bit more detail as well.

So let's next think about and have a look at another example. So in this case, we took a look at how it can help us with some of those more intricate parts. But let's say I'm interested in connecting different portions. So in this case, I have a workflow that is effectively grabbing a file from an API. But at this point, I'm interested in putting this into an S3 bucket, for example.

So I'm gonna go ahead and ask Merlin, can you add to the workflow the ability to write the file to an S3 bucket?

And, again, this is gonna be doing slightly more than the other one in just in terms of not just providing your prompt response, but actually adding and making adjustments to the workflow within the builder here as well. And, again, this is really helpful because it means that when you are developing, you might wanna get a good starting point or a helping hand in what you need to do. But you can see that Merlin has come back here, and it's added in, figured out, okay, what we're gonna be, of course, using the S3 connector, but also the operation that we're gonna be leveraging here as well.

So in this case, the put object file. Because this is a service that we're dealing with, Merlin is also asking me for the authentication that I want to use. So I can go ahead and select mine there. Click confirm, and it will just go away and finish configuring that for us.

And we can see if we bring up the actual configuration here, it's given us an even link to the file from this previous step in our workflow in this operation that has been added on by Merlin. So really, really cool. We've taken it from just providing that prompt response to actually being able to interact with that workflow as well and add on that connect that we're interested in for this particular use case that we are kind of exploring effectively.

So, again, really powerful.

Finally, one of the other ones is take that a step further and be able to... So if we go ahead and clear the chat here, what we can do is we can then take a look at adding in some extra sections that we would be kind of interested in. So I'm gonna bring up Merlin one more time. I'm gonna ask you to do a little bit more now in terms of what I would like it to build. So I'm gonna go ahead and ask Merlin to basically create a new Google Sheet, get the name and email from all Salesforce leads, write each lead to the Google Sheet.

So we're gonna ask it to do a little bit more. So the previous one was looking at it, you know, adding in the S3 bucket operation, but now we're asking it to do a little bit more with a couple of extra systems as well. And, again, just as before, this is really powerful because it gives me a really great starting point when I am building out these integrations, that sort of lending and helping hand to really improve that building experience and ultimately the velocity at which I can build as well. And if there are certain things that I'm not sure on or certain connector operations, this can give you a really good skeleton starting point that you can start to work from as well. So we'll give Merlin a second to go through and process that.

And you can see that it's going through one by one and adding in the necessary steps. So we've got the Google Sheets, we've got the Salesforce leads, even got the loop step that we need to do to loop through. And just before, it's asking us for the authentication information. So I'm gonna go ahead and add in my Google Sheets authentication as well as my Salesforce authentication, and it's gonna go through and configure those steps for us one by one.

Just fill it in in the blanks now that we're providing it with that extra information. So we now see we have slightly more steps in our workflow, slightly more intricate example there, but Merlin was there to help us give that skeleton starting point to that building experience for us. They were some examples of how we can make use of AI when we are directly building kind of within the platform, but let's also look at how we can make use of it as part of some infused processes. So I've got an example here to take you through, which is effectively going to be enabling us to extract information from unstructured data by making use of AI effectively.

And so what we got with this workflow is that this one is in a process where we can basically classify some information. So we're classifying some unstructured data based on whether it's something like an inquiry, whether we want it to be an unsubscribe message and use AI as part of that or as part of an integration to be able to determine, okay, is someone looking for more information, or looking to unsubscribe, etc.? So this workflow is relatively straightforward. It's going to be grabbing a particular message just as, for testing purposes, from my Gmail account, and it's gonna be passing that into this cool workflow.

Now this cool workflow is gonna call on another Tray workflow that's actually making use of one of our AI connectors for OpenAI as part of the actual integration processing. So you can see I've given it a few pieces of information that it's made use of in that course. So figure it's the text that we want to kind of classify as well as what we want those classes to be based on the information that we need and as well as some examples of, you know, for example, this text is an example of an inquiry. This one is an unsubscribe and different things.

So we have a couple of script steps that are just gonna format those classes and those examples for us into a nice piece of text as well as the actual prompt that we're gonna provide as well. So you can see what we're doing here is we're building out the prompt that we wanna use in this call to this OpenAI service, and we're saying that we wanna classify the text as one of the following classes, returning one of the listed ones as well. You can see that we then have our OpenAI connector here, which has been authenticated in tune as making use of their API, that we can then pass in that information that we've configured.

Once we've got all of that information back, then we're gonna provide that into a callable response. And, effectively, to keep things super simple, if we go ahead and run this and click on the logging portion, we can see that we've got a run at the very bottom here, and we should, in just a moment, start to see that populating and updating with that extra information. So you can see that we received the email contents. And just to show you what that was, was effectively someone saying, hey. I would like some information on the Tray platform in order to be able to implement a point-to-point integration.

I then got the cool workflow here, and you can see that the response that came back was inquiry. So you can see that what's happened is we've taken that text from Gmail, in this case, with one of our connectors. We then made use of one of our AI connectors, in this case, OpenAI, passed in the prompt and the information that we need to do some text classification, and that's, of course, returned back that classification criteria. And now I can make use of that on the rest of my integration as well.

So now that I know that this is an inquiry, for example, perhaps I want to automatically create something in my CRM system. If it's someone asking to be unsubscribed, perhaps I want to create a ticket or perhaps even automatically unsubscribe them as part of this automation as well. So, really, what you can start doing here as part of the integrations that you're building is making use of these various AI connectors that are out there for all these different services and start infusing your existing processes with those. So you might already have existing integrations that you're working with, but how can you start to bring AI into that?

This connector here and these connectors that we have provide a way to do that so you can start having some of those infused processes as well.

Hope that's again helped bring some of that to life about how I can be using a couple of different ways there from the building experience to being able to infuse those processes. But I'm gonna pass back to you, Paul.

Wonderful. Thank you so much, Luke, for that demo there. So I hope everybody put it at home, right? The understanding the ways you can use AI to augment your integration and automation development, but also, you know, the ways you can start to bring in AI services into your integrations, new automations to upgrade the level of automation you can do and add much more intelligence to your processes too. And so with that, I'm gonna hand it back to you, Vance.

Paul, Luke, awesome. Great session. Great overview of AI's potential.

And, wow, who doesn't love a demo? And you really gave a lot of good flavor to how Merlin AI has really given a lot of very close attention to some of the intricacies of integration. So overall, a really great show.

Thanks, Vance.

Thank you, Vance.

And luckily, you both left us some time for questions. So with your permission, let's start with some questions.

Sure. That'd be a great idea.

Paul, you gave a great overview of not only the technology components and the implementation mindset that companies should have. And one of the questions that we've been dealing with all day in certain sessions is something like this. Question says, does Tray have suggestions for how we could build an AI team across our IT and knowledge worker groups?

I think what we have found is that it's really important just to start prototyping and really understand what you can build. I mean, our focus at Tray is that you don't have to be a PhD, right, in AI to get value. And we're taking the same kind of thinking with our low code about making it easy to build integration automation. So we're taking the same point of view with AI as well.

And so what you'll find, for example, is that it's pretty straightforward to start to call out to OpenAI or Anthropic or Vertex AI and those kinds of things and incorporate those into your workflow using API integration. And the other thing that we have focused on is that we're providing templates as well, you know, thought starters. And so even, you know, today, right, in-app, if you sign up for a trial with Tray, you can start to deploy templates that have some infused AI service. You know? But it's really a partnership between the development team and the business team as well.

Right? Because, ultimately, infusing AI into your processes really comes down to automating work, making those processes more intelligent with things such as understanding customer risk or customer sentiment or those kinds of things or maybe automatically extracting text from an order. It's all about driving business value. One of the use cases we've seen come up over and over again from our customers using modern AI today is just explainability.

When you think about automation at scale, right, when you think of all of these workflows running, and we have some customers that are running a hundred automations across their business. This is pre-AI. Even things like asking modern AI to say, you know, explain this workflow to me. What does it do step by step?

It's really helpful when you have a new engineer on your team to help them get up to speed. So what we're finding is there are lots of unexpected ways that you can use AI to augment the development.

And, Paul, you bring up some really great points about the implementation, how people do early adoption. One of the impressions I got while Luke was working on the demo that we saw is the idea that you could easily have shoulder to shoulder an IT person or an AI expert alongside of just a knowledge worker who knew a lot about process but not a lot about integration, kind of guiding one another into completing a project or at least a POC for what that AI-driven integration should look like. Is that something you're seeing with customers?

Yeah. That is. Collaboration is key, and it really helps to have a single ID, single interface that everyone can kind of crowd around, you know, whether you have your integration expert and your marketing operations person or your sales operations person, everyone can crowd around and start to iteratively add AI to their business processes. So collaboration is really important. It's a big part of how to start this.

Excellent. Excellent. You know, let's turn to another topic, and that's the whole idea of models or LLMs. You brought that up a couple times. The question here says, the speaker mentioned LLMs can boost every step of the iPaaS development. Can the speaker talk a little bit more about why the models are important and what types help with integration and workflow projects?

So let's start with the infusion side of things. This is adding AI to your processes.

What we found is that you wanna try out different large language models. You might find, for example, the Anthropic Claude is providing better results for you than using OpenAI, or you might start to look at Llama 2, for example. Right? So what we've found is that different LLMs shine at different applications and different processes.

AI, you know, you just have to augment it with your data. Different LLMs may shine versus others. So you want to have the flexibility to plug and play and see what works best for you.

You know, Paul, you had an awesome, if maybe eye-bending chart that showed a ton of models going about that people might be using or heard of these days. Has Tray designed Merlin AI to keep up to date with all these models and maybe talk a little bit about what your interface or what your maintenance program is for letting companies take advantage of so many models?

The good news is that with Tray, we are AI agnostic. So we're an API-based integration platform. So you can easily kind of plug and play these different services within your business processes. And we also provide a connected development kit as well.

So you can also wrap these services as a connector as well. The service is gonna change over time. Right? So you want a platform that is flexible enough that the AI that you incorporate into your customer escalation process today, you wanna make sure that it doesn't end up as technical debt for tomorrow.

Right? You have the fluidity to plug and play. The model is du jour for 2025 and 2026.

You know, Paul, this has been a great session, Luke. Great demos. Just one or two more I think we can squeeze in. You know, you mentioned process several times in your session.

Process improvement's a big topic here, certainly too, at the AI Deployment Summit. Kind of an interesting question that caught my eye. So it's a bit futuristic but kind of fun. Does Tray envision that Merlin AI could actually monitor our existing processes and use AI to make suggestions for how we can improve them?

Yeah. I mean, I think that's kind of the ultimate vision going forward. Right? Is that if you think about automation at scale and if you think about the execution data that your processes are throwing off, if you think about the processes we've built today, they're pretty static. You know, when you build that integration, it's not gonna be changing. The business logic is unchanged.

As you start to infuse AI into the processes, they become more dynamic where the AI is making the sentiment or the risk-based decisions in the process. But as you go even further out, it's actually the process being much more adaptive based on the data as well, even all the way through to making recommendations, right, on which processes you can automate too. So it's a rising tide of AI for much smarter, independent, autonomous, self-improving processes.

You know, Paul, that was a great summary. One last note before we go. You mentioned the free trial. Let's see if we can give a quick summary of all the ways that our attendees can engage with Tray and on-ramp to you or get more information. Give us a couple of other next steps that our attendees can take to get a little more info and get closer to Tray.

Yeah. Thanks, Vance. So, yeah, there's a number of different ways. You can sign – I mean, we make it very easy to sign up for a trial.

So if you wanna sign up for a live trial of Tray and start prototyping and building, you can do that, you know, right after this event. We're also providing a whole set of materials. We have a Gartner integration capabilities white paper where you can start to design what your integration strategy looks like, as well as a number of other materials as well. And we also provide a personalized demo, so we're happy if you sign up.

We're happy to schedule a personalized demo with you and learn what you're looking for in terms of your business processes, and we're happy to suggest ways you can start to use integration automation within your business as well.

Awesome. Awesome. Paul Turner, integration strategist, Luke Smith, sales engineer at Tray.ai. We learned a lot about Tray’s technology, the universal automation cloud, and specifically the Merlin AI offering that really brings AI to bear for integration and ops and certainly process improvement. Really great session, guys. Really glad to have you here.

Thanks for the opportunity, Vance.

Thank you, Vance.

Totally our pleasure. And a quick note for our attendees before we go. Paul and Luke mentioned some great takeaway assets for you, including that great way to take a free trial of Merlin. If you wanna get to learn more, there's certainly a lot of terrific on-ramp assets, including the use cases and some great ebooks and certainly that Gartner report, which is really worth taking a look at, especially for those of you in the integration space. And let me just remind you that download today's slides and you'll be able to get even more great homework because all these links will be live directly to the Tray.ai website. So thanks again to our speakers this morning, and thanks again to the attendees for some really terrific questions made for a great session. Thanks, everyone.