I'd like to welcome John Joseph. He is the vice president of solutions, which is a great title, by the way, from Tray. He will be sharing about unlocking the potential of AI and modernizing digital workflows.

Thank you for joining us, John. I know we're in for a great session. And just before I leave it to you, I wanna remind the audience, please keep the questions coming via the live chat and our q and a. And if we don't get to the questions in this session, we will get back to you after the fact. John, I leave it to you.

Alright. Thanks, Fran, and thanks everyone for attending my session.

So what I'm gonna walk you through is is not so much a a deep understanding of Tray as a product, but more about how you should be thinking of, you know, setting up your your workflows, your low code automations, your integrations so that you can better position yourself to unlock the value of AI in your organization.

Alright. So let's get into the presentation.

So, if we take a look at some of the challenges that organizations are facing in today's digital landscape, if we take a look and sort of rewind the clock over the past decade, we've witnessed an unprecedented growth and now a consolidation in the SaaS market.

And as a result of that, businesses are continuing to reshape how they're operating because they're heavily reliant on the SaaS, products that make up their ecosystem of applications, business processes, and data.

So you couple that with a noticeable disparity in how organizations approach automation and integration.

You know, businesses are facing the challenge of connecting data processes and experiences in a landscape where each one of those solutions provides a unique approach.

And often what happens is it leads to operational silos and inefficiencies.

So you couple those two things with now us entering into this transformative phase in tech, marked by the rising impact of AI. And AI is not just reshaping individual applications, but it's fundamentally altering how enterprises are approaching decision making, customer experiences, and automating processes.

So we're going to double

click into each of all three of these here in a little bit.

So if we take a look at, you know, what, going back along the lines of of the growing SaaS stack, and the compressing SaaS stack, if we take a look at some of the, you know, the the data that we've received from analysts, from SaaSter and from Gartner, to kinda summarize here, the long pole in the tent in any transformation project, in any business initiative is integration.

You know, if we take a look at some of the metrics that are on the slide here, the average public SaaS company has about three hundred and fifty integrations.

75% of respondents of this survey from SaaSter say that it takes three months or longer to deliver or an integration for a project.

And according to Gartner, application integration still is a major challenge, for many, for many projects, and it's a result of failure projects.

So, again, it's not so much a technology problem. I think this is more of the approach that many organizations take towards automation and integration. And we'll touch on this a little bit more on how to resolve it.

So you compound that with, okay. Well, why is this the case? Well, if we take a look at some of the iPaaS architectures that many of these companies are using, it's based on rigid architectures that have been around for about twenty years.

They can't support automation at scale, without huge operational overhead and heavy developer lift.

They only produce static integrations that kind of fall out of alignment with the business as the business is changing, and their builder experiences are somewhat outdated.

And it's not easy to infuse AI.

And then IT teams have to battle with throttle delivery models, fragmented code bases, and frameworks.

And, again, that's making it impossible for continuous improvement and also embracing AI.

And we all know that AI is coming. Right? And according to Gartner, AI augmented data and process integration is nearing us, and so we've gotta prepare for this. We've got, you know, this ever changing SaaS stack.

We're approaching automation and integration in a sort of a fragmented way, and we know we've got this storm of AI that's sort of emerging upon us in our organization.

So how should we be rethinking this problem so that it better positions us, not only to deal with that fragmentation, but position us to take advantage of GenAI?

And so by addressing these challenges, what I'm effectively saying is that it's not just about adopting a new technology or new set of technologies.

It's about a paradigm shift in our approach to integration.

And it calls out for a robust and adaptable architecture that can serve as this foundational platform to tackle these challenges head on.

And so this approach involves sort of rethinking how we integrate and leverage technology to create seamless, efficient, and future proof business processes.

Okay. So what is that architecture? I'm calling it the composability architecture. I wouldn't necessarily say that this is novel.

Many organizations have kind of followed maybe a couple of these layers here and there.

But I'm going to walk you through what I mean by this composability architecture, and I'm also gonna cover a little bit of ownership of these layers.

So this composability architecture puts organizations on what I say is an improved path for AI, and we're gonna see this a little bit later in the presentation.

At the core of this architecture lies what are called our data centric services. And this layer is designed to handle diverse data types and data sources, ensuring robust data collection, storage, and management.

It's the bedrock that supports seamless data integration across multiple SaaS platforms, and it's turning data fragmentation into unified intelligence.

Building on top of the data centric foundation, the next layer is our process centric workflows, and this is where the magic of automation and efficiency takes place.

I would say here at Tray, we facilitate the design and execution of complex workflows, ensuring that every process from the simplest to the most intricate is streamlined and optimized. And so this layer is crucial for eliminating those operational bottom bottlenecks and overall, efficiency enhancing overall efficiencies.

And at the top of this architecture, we have experience, centric automation.

And this layer or this tier is focused on that end user or data ingestion experiences, and it's harnessing the power of both those data and process layers to deliver tailored and intuitive interactions for those experiences.

So from customer facing applications to internal tools, this layer ensures that every interaction is kind of optimized to fit the form and function of that consuming application and or tool.

And from an ownership perspective, the foundational tier of our architecture, that data centric services layer, that's primarily owned by, the steward these, stewards of integration specialists, and these are experts that ensure the, integrity, availability, and and so what I like to say is that that data centric layer is typically owned by, those data centric services are owned by the teams that own the the SaaS applications or those legacy applications, or the homegrown applications because they understand the data model and security model of those those existing applications.

And so they would be responsible for building those data centric services.

Process centric workflows represents kind of a collaborative domain.

This is where integration specialists ensure data flows efficiently or effectively through these processes.

And business technologies play a critical role in this layer as well. They kind of bridge the gap between technical capabilities and business needs, and they tailor workflows to optimize operational efficiency and align with business objectives.

So these process centric workflows are effectively owned by business technologists that understand the business process. Say, maybe it's its lead life cycle management or lead processing that would own by a rev ops team. And we'll walk through an example of how do you implement this architecture here in a little bit.

So that that process centric layer is an example of, going into the lead life cycle use case.

That would be owned by somebody in rev ops that understands lead processing, and then understands sort of the ecosystem of capabilities that are needed to produce lead processing.

Working with the integration specialist to handle, alright, how do I integrate with, my Salesforce or my CRM, to enrich contact information within those platforms.

The experience centric layers are led by the or owned by the business technologists, and they harness the power of both, data and process layers to personalize experiences, ensuring that every automated touchpoint, resonates with end users and delivers tangible value. So this experience layer is effectively owned by the business technologist.

You can think of them as tasks, automation tasks that will improve the experience of users that are trying to automate, you know, processing, like, in the lead life cycle use case. And, again, I will walk through an example here in a bit.

And at Tray, our approach is we, it's not just these layers of technology. We provide a well orchestrated system with clear ownership at each level.

So integration techno integration specialists and business technologists can work in tandem, ensuring that every aspect of our platform is expertly managed.

And so this is more of a collaborative approach and not only enhances the efficiency of operations, but also drives innovation and strategic alignment with business goals.

Okay.

So the foundation or the building block of this composability architecture is something that we call the modern digital workflow. And the modern digital workflow is effectively a logical construct that is made up of logic and connectivity.

So this is where you implement your conditional logic and your decision making capabilities that represent your business process or data centric service or or your experience centric, automation task.

And so it's a company and that includes logic and connectivity, ensuring seamless integration between different systems for smooth data flow. It's API enabled. So what that means is once you create one of these workflows, you can seamlessly publish this as an API for, for consumption, which leads to it becoming composable.

So taking more of a modular, based approach.

It's scalable, so it ensures that these digital workflows can scale based on the demands of the business process.

And then it's also there's governance baked in, which means that you can implement policies to ensure compliance with organizational standards and legal requirements such as data privacy laws.

There's security baked in, which ensures that we're securing sensitive data through encryption, access controls, and also, regular security audits. And, finally, it's AI enhanced, meaning that we can provide adaptive decision automation coupled with predictive analytics, to deliver intelligent automation within the workflow itself.

Alright.

So let's take a look at an example of, alright, how do you build to this architecture? And we're gonna use a lead life cycle example or a lead management, a lead processing example. So think of it as, you know, your company is involved in many marketing events, and you wanna be able to collect leads across those events, so that you can engage potentially with those leads and sell them on your own products and services.

And so there's many ways that you can collect lead information. You can collect them via files.

You can collect them via forms through your website, or you can expose the collection of leads to partners via an API.

But regardless of how that's being collected from an experience standpoint, and that's kind of what the top layer is representing, is, you know, there's sort of three different ways or experiences in which you can collect leads.

The way in which you process leads is going to be exactly the same. Right? So in a lead processing work flow, it involves validating the lead, handling consent management, enriching that lead with, additional data, recording that lead in your CRM system, providing attribution, in terms of, like, where did that lead come from, scoring the lead in terms of that potential, you know, business opportunity with that lead, routing it internally to teams so that they can start then doing gauge and outreach, to that particular lead. So this is an example of how you can build to this architecture so that you have a uniform way in which you're processing leads.

So you can quickly add additional experiences in terms of how you're intaking leads, and then process those leads effectively.

What this architecture also does is it future proofs your investment in SaaS.

As I mentioned before, during the beginning of this presentation, what many organizations are kinda facing today is a rationalization of their application portfolio.

And by providing these data centric services that sit on top of your systems of systems of record, make up your technology stack, what you're essentially doing is providing asset installation.

So in the event that, you know, I wanna change my CRM from Salesforce to Zoho or HubSpot, I don't have to go back and rework this lead processing workflow. All I have to do is change my record, and maybe my attribute, data centric services to point it to the new CRM and my lead processing system or workflow remains full. So it's a way of future proofing your investment, without having to rework your business processes, all over again.

And, again, by providing this data centric process, this process centric lead processing workflow, I can quickly add additional experiences on here as well.

Maybe I wanna take, form intake from or or lead intake from another data source that isn't via API or files or from forms. I can quickly add that additional experience on top of here, and I can take advantage of that lead processing workflow for that new experience.

So this is just an example of how you could implement this architecture, to support your integration initiatives for your projects. And my recommendation is that you don't build everything.

You don't, you don't start by saying, okay. I wanna adopt this composability architecture, and I'm gonna start building a bunch of data centric services on top of my systems of record.

What I recommend is the way that you start going on this is, start with that first project to identify what are some data centric services that I can put in place on top of my systems of record, so that I can use those data centric services to orchestrate a workflow. So in this, lead life cycle example, the way that I would approach this with, you know, with my stakeholder in building out this architecture is, let's just say the, the first iteration of the lead life cycle management involved ingesting a CSV file with leads. So what I would say is, okay. We're gonna build out the lead processing workflow, build out the data centric services that validate, do consent management, enrich, record, attribute, and maybe not necessarily score, but maybe engage.

So I would build some of these data centric services in that first project to do that file intake of leads.

And then, when the next project comes up and says, okay. We now wanna make that capability available via an API so that we can expose it to our part or to other marketing agencies that we're partnering with. Then then what we would do is we'd say, okay. We have our lead processing workflow that's sort of experience agnostic or ingest agnostic.

We're gonna go ahead and build an API that sits on top of it that allows that it's purpose built for those marketing agencies, to receive leads via an API, invocation.

And so now what I've done is I've cut down on the integration time on that next project by investing a little bit upfront in the first project. So that's sort of the methodology and approach on how you can build to this architecture in the context of projects rather than building data centric services and process centric workflows in isolation. You're building based on the requirements of projects. And you may have to go in and enhance them over time, which is okay, because you're gonna get more requirements to improve the robustness of those services over time.

Now one of the added benefits of using this approach is that what you're effectively doing is putting together a taxonomy around your automations and integrations.

And a taxonomy is just a fancy way of saying that I'm sort of carving up the way I handle data integration. I'm carving up the way I'm handling business processes. I'm carving up the way I'm handling user automations and tasks into categories.

Right? And what does AI really need in order for it to flourish? It needs well indexed data. Right? So, typically, when you train a model, you've gotta you've gotta enrich a vector database of some kind with with, index data that represents, contextual information that LLMs can use to get educated on so that, they can perform, you know, their their their tasks based on prompts that are provided by, you know, business technologists and or prompt or or, prompt engineers.

And so when you build to this architecture, one of the things that emerges is something that I call an automation network, which is basically a fabric of automations and integrations that represents how you're orchestrating your data, your processes, and your user tasks and automations across your enterprise.

And it's effectively a well indexed dataset of those tasks, processes, and data integrations that you can feed into a vector database and educate your LLM on how your organization is leveraging applications, data, and experiences, in a of your stack in a unique way to support your business partners, your internal users, and your customers. Right?

Every organization kinda uses a similar set of applications, but what makes organizations unique is how they're orchestrating those applications and data to deliver experiences.

And so by building to this composability architecture, what you're basically doing is you're creating a network of how that represents how you've orchestrated all of those applications, data, and experiences together. And as I mentioned before, it's now well indexed so that you can feed it into a vector database. Say it's a pine cone or a llama index or chroma, so that you can start to educate these LLMs on how your organization operates. And the more that you can educate these LLMs on how you or your organization operates, you're effectively getting towards that path of becoming an autonomous enterprise.

So that's that's sort of the the the value that this composability architecture brings is it sets you up, for taking advantage of generative AI, by having this well indexed representation of how you're orchestrating your experiences, your processes and data to support your internal consumers, your customers, your partners.

Now I would be remiss by not sharing a bit of a customer story.

So the example here is we were working with a digital marketing agency, and we helped them scale their business by infusing AI, using Tray.

And so this particular digital marketing agency was sort of resource constrained. They were struggling to scale, without significant cost increases, and they were struggling with handling things like lead management, client reporting, customer demand operations.

And so by infusing AI into these processes, leveraging that composability architecture that I shared earlier and that lead processing example that I showed earlier, they were able to save about six hundred thousand dollars. They increased their campaigns managed by, per per full time employee.

They were also able to reduce, many, hours spent in manual reporting, and they increased their margin per opportunity.

And this the biggest call out here is, they were able to quickly operationalize AI, compared to any other company that they were competing against.

And so it became a competitive advantage to them. And so, you know, as I would say to you, this is a way for you to gain a competitive advantage against your peers that you're competing against.

Alright?

So that sort of concludes the presentation.

We could go ahead and open it up to questions from the audience.

Sure. We have a couple of questions now. Thank you for a great presentation, John. So, the first question is actually something that is top of mind for me at this time of great change with AI. What ethical considerations must be taken into account when implementing AI driven decision making processes within these automated autonomous enterprises?

Yeah. So that's a great question.

I think, there's the way I like to say this is there's eight things that organizations need to be thinking about, around the ethical concerns related to AI gen AI.

The first is around biases and fairness.

You know, AI systems can inherit biases from their training data.

So, you know, leading to unfair decisions that are made through automated processes.

So what organizations need to strive to identify is mitigating these biases in their AI algorithms to ensure fairness in that decision making. And this goes back to, you know, making sure that you have well indexed data that you feed into these vector databases to help educate your LLMs so that they're making the right decisions. There also needs to be transparency in how these systems make decisions, because they impact individuals at the end of the day. So it's important for enterprises to disclose the role of AI in their operations.

There needs to be accountability. There's gotta be someone or some entity within an organization that's responsible for the decisions made by AI.

Obviously, privacy is a big concern.

So AI systems often are processing vast amounts of personal data. And what enterprises need to ensure is that, you know, the data is collected, used, and stored in ways that protect individuals' privacy and complies with data protection laws in various countries.

Mhmm. Obviously, security is a big concern.

There's a societal impact to it. So there's the broader impact of AI, on society such as job displacement and social inequality. Those could be concerns. Right? And so what organizations need to be actively working towards is, you know, mitigating some of the negative, societal impacts, such as, you know, providing, supporting, workforce transition to new roles.

And then the last two is around sustainability and consent and autonomy. So sustainability meaning that, you know, you're processing vast amounts of data. Right? So there's consumption concerns around power that more organizations need to be aware of.

And then in the case of consent and autonomy, when AI systems make decisions, it's crucial to consider the autonomy of those individuals.

And so enterprises need to ensure that individuals have consented to the use of their data, in a way that is informed and voluntary. So there are eight things that you need to be aware of.

I think those were really great. I actually pointed them out in our live chat.

Is there anything else you'd like to say to the audience before we head out? Because we have run out of time.

So the one thing I would say is that, when you're, when you're embarking on your approach to AI, what you may wanna do is go and rethink your strategy around integration.

Take a look at this composability architecture.

There's nothing in it that is tied to a specific vendor. It's an architectural pattern that you can apply to many techno to your existing technology investment.

I would say here at Tray, those three layers that I showed you earlier, we provide all of those capabilities in one platform, which makes us a bit unique. But should you have you should you have made investments in creating your own data services, or you've made some investments in that process layer, or you made some investment in that experience layer, that's okay.

Build that architecture because it's gonna set you up for greater success as you start to embark on your AI journey.

Great advice. Thank you so much for this presentation.

Again, audience, if you send any more questions and the ones that had come to us, I see Hugo has sent one now. We'll make sure that John gets a hold of this and follows up with you. Alright? And be sure to connect with all of us on LinkedIn. John, thank you so much for a great presentation.

You can exit by clicking on the backstage, at this time. Thank you. And we are going to get ready for our last session of the day and of the webinar series.

So I'm glad you're all staying with us.