SaaStr AI Day: Using Product Integrations to Enhance your AI Capabilities with Prismatic's VP of …

Hello, everybody. Welcome. Gonna give it a few more seconds to have some more people join. Welcome. Okay, I think we can go ahead and get started. Like I said, welcome everyone to SASter's inaugural AI Day. We appreciate you joining us for the session. We're super excited to have Marcus Edgington with us this morning, who is the VP of Product at Prismatic. And he's going to be discussing with us how to use product integrations to enhance your AI capabilities. So a quick note for everyone tuning in, we would love to make this session as interactive as possible. So if you feel comfortable, you can turn on, uh, your cameras and then drop in the chat where you are tuning in from. And then if you have any questions that come up during his presentation, feel free to also drop in any questions into the chat. And he will make some time at the end to answer as many as he can. So please join me in welcoming Marcus to AI Day.

Awesome. Thank you very much, Jesse. So I just said I'm the VP of product at Prismatic. And so during this session, we'll be talking all about integrations. So before coming to Prismatic, I was a product manager for several B2B software companies that relied heavily on integrations to bolster the capabilities of their core product offerings. And through those past experiences as a product manager, I've seen firsthand the value that these integrations can provide to SaaS companies and also the challenges associated with building those integrations. And those challenges are a big reason that I wanted to join Prismatic. Here at Prismatic, our goal is to make it easier for B2B SaaS companies to connect their product to the other products that their customers are using. We believe that the rise of AI-powered applications have made these integrations even more important. All right. So even before the rise of AI, B2B companies have been expecting their SaaS vendors to provide integrations for them. And it's really not all that surprising given that most companies have over 100 software applications. Your customers want integrations because they create— these integrations can create data consistency across their large software ecosystems. It eliminates dual entry. and it can also decrease the amount of time that they're, you know, they need to spend across, across multiple applications. So for many years, offering up like a robust API was going to be the go-to answer for vendors that wanted to help their customers connect their application to the other applications that their customers are using. But that's just really no longer enough. Customers are now expecting their SaaS vendors to provide many integrations as essentially out-of-the-box functionality. Rather than something that they have to build on their own. And if you're providing AI-enabled applications, you know, these integrations become even more important because not only do they provide the user conveniences that I, that I mentioned before, but they're also going to feed third-party data to the AI models that are driving the features or functionality within, within your product. All right. So building and managing integrations has become a giant headache for most software companies. So many B2B SaaS companies have found themselves spending way too much time or way too much of their dev resources' time on these integrations rather than their core product. As they scale up the number of integrations that they're building, they start to push the limits of their integration infrastructure. And this oftentimes leads to support issues or even worse integration downtime. And, you know, the list of challenges associated with building integrations goes on and on. Many of you watching this session have either experienced those yourselves or, you know, you anticipate, you know, experiencing those at some point in the future. And this is, this is where Prismatic comes in. It's going to help you alleviate these challenges by providing tools that enable you to build the integrations faster, a white-labeled embedded marketplace that allows your customers to set up and enable these integrations themselves, and an infrastructure to run and manage those integrations at scale. Alright, so let's see Prismatic in action so we can see a little bit more about how, how Prismatic makes it easier to build those integrations. So right now I'm logged into Prismatic as a member of a fictitious company. Let's call it Acme AI. Acme builds software for customer support teams of manufacturing companies. Their secret sauce is an AI model that makes it easy for its users to automatically triage support requests, offer real-time suggestions to their support team as they're helping customers, and also predict future staffing needs and so on and so on. So Acme AI, what I've described here with Acme AI is very similar to many of the AI B2B SaaS companies that are using our product today. Their customers are using their AI product alongside any number of other software applications that contain data that's going to help make their AI functionality in their product even better. So in Acme AI situation, for example, their customers typically have a CRM that's going to be the source of truth for their customers. It's got information about like where your customer's located, how long they've been a customer, products that they've purchased, and so on. They'll likely have an accounts receivable application that knows things about open invoices. Maybe they also have supply chain management tools that's going to have information about things like inventory backlog and shipping times, things like that. So this is just a few examples of the applications that a company like or a customer of Acme AI may have. And they want— Acme AI wants the data that's in these applications to be fed into Acme AI because it's going to provide better recommendations within the AI functionality that the support team members are using. So in order to get this data into their product, Acme needs to create integrations with these products. And they want their customers to be able to stand up these integrations themselves without any, any kind of development effort. All right. So let's walk through the project a little bit more. So from the integration page that I'm on now, I can create new integrations or view those that I've already built or in progress. As I mentioned, Acme AI is interested in setting up, getting up-to-date CRM data. So let's take a look at the HubSpot integration. So right now I'm looking at the low-code integration builder. Prismatic's unique in that it allows its users to build integrations in one of two ways, either with our low-code builder that you see here, or via what we call a code-native SDK. The code-native SDK is just a toolkit that allows you to define an integration via a TypeScript project rather, rather than low-code. So many of our more technical builders favor this code-native approach because it's going to allow you to define integrations in a way that's that's more aligned with what you're used to doing day to day. But the low-code builder still has a lot of advantages, especially if you're looking to offload integrations from your dev team to maybe others in your company, such as your professional services team. I'm going to stay in low-code builder because it provides a, you know, a great visual for how Prismatic works. But our website has a lot of great documentation and videos on, on the code native if that's, if that's going to be more your thing. All right, so let's talk a little bit more about what we're seeing here. So in Prismatic, an integration is one or more flows that execute a series of steps to perform some business logic. So in this integration, I've got two flows, one that's writing data to HubSpot from Acme and another one that's receiving data from HubSpot and just, sorry, the other way, but that's pushing that data from Acme to HubSpot. And within these flows, there's four main kinds of steps. There's triggers, that are going to identify when a particular flow should run. This is most commonly going to be like a predefined schedule, say once a day, but it could be a third-party webhook or a webhook that's invoked by your own application. We also have steps that interact with the APIs of other products. Usually these steps are going to take some action within the application such as writing or reading data. So in a couple examples here, you see I've got a find contact or an update record or create a contact in HubSpot. We've also got utility or logic steps like this mapper that's going to map fields from your data to the data in your— in HubSpot. We have branch steps that are also going to do things like determine which steps to perform based off of what kind of data you have. And so together these steps are going to define the logic for integration. Once we define the integrations, I'm able to test this integration, what we call the test runner. The test runner is a way to simulate what would happen when this integration is enabled by your customers. I'm seeing past integrations as well as one that I just ran. They can see the status of the steps in that integration, the outputs of those steps, and logs associated with those steps. So once I decide that after building and running a few tests that the integration is doing what I want it to do, now I want to deploy it to my customers. Pragmatic gives you a lot of flexibility in the way that you're going to do this. If you want, you can deploy these integrations directly to a customer. But better yet, I can put this in an integration marketplace within my product and allowing my customers to enable these integrations themselves. So let's take a quick look at Acme AI's fictitious website. Within their website, they've added a page that lists all the integrations that are available to their customers. Prismatic makes it super easy to do with our embedded SDK. Our customers use the embedded SDK to set up this embedded marketplace in 30 minutes or less. Within this marketplace, I'm able to see which integrations are available to me as a customer and any integrations that I've already activated. So in this example, I've already enabled the HubSpot integration, but let's pretend I didn't. What I would see when I open this for the first time is a config wizard. That's going to tell me things about the integration. It's going to allow me to put in my connection information for HubSpot, apply field mapping, and any other configuration that's going to be unique to my integration. This is super important because it allows me to, as Acme, build integrations once and make them reusable for a lot of customers rather than a bunch of bespoke integrations that are unique to all my customers. All right, so after an integration's enabled, you can now manage these integrations via some robust monitoring tools within Prismatic. It's going to allow you to see all the executions that are happening and the logs that are being generated by those executions. You can access these logs within Prismatic, or if you want, you can stream them to Datadog or New Relic or whatever logging system that your, that your company's using. And these monitoring tools are essentially going to allow you to to troubleshoot quicker if anybody calls in and say, hey, I'm having, having trouble with my, with my integrations. All right, that was a super fast overview of Prismatic and how we make integrations easier. So to recap, Prismatic allows you to build integrations faster. It makes it easy to deploy them directly to your customers or through a marketplace, as well as tools that help you monitor and troubleshoot these integrations. The Prismatic platform's in use by customers across all verticals and sizes, including a number of AI-focused companies. So I think we're going to take a few questions here, but I do want to call out that you can visit prismatic.io to learn a lot more about our product. All of our documentation's available on the website. You can also try a free trial. And if it's something that you're interested in, you can reach out directly to us via, via our website to set up a demo or a proof of concept.

Amazing. Thank you so much, Marcus. That was really great content and appreciate you making some time for questions. So it looks like we have one question from David in the chat and looks like he is curious about what are some of the newest demands, if there are any, that you are hearing from customers as a result of new AI needs?

Any new demands? So I think one thing that we hear more of is that, you know, if, if I'm going to have an effective AI application, I need the most up-to-date information. So being able to create what are essentially real-time or frequent, frequent updates to your data are super important. So we've done a lot within our product to do things like making webhook setup easier to do so that you can get on-demand data basically anytime. Say, for example, a contact's updated in, in, in your CRM, you want that immediately available in your product so that your AI tools can take advantage of that data. So we've got robust things like webhooks set up, but we also have fast execution times on our integrations, things that are gonna make sure that the data that's in your application is up to date as possible from these other third-party applications.

Amazing. One question we received on our livestream from a viewer, how does Prismatic combat customer churn?

How does Pragmatic combat customer churn? Yeah, absolutely. So I think that when you have integrations and your customers are using these integrations, they're making your product stickier because not only are they now using your product, they're leveraging your other products in other ways. So if they were to replace your product, now they're not only going to have to replace your product, but also the integrations associated with that, that have already been set up. So there's a stickiness factor that comes with having your customer set up a lot of integrations. So that's one element of it, and I think that by doing that, you can definitely reduce customer churn.

Amazing. Another question we received in the chat, this one's from Brian. Does, does Promatic support sales CRM and developer tools primarily, or do they, or do you also offer drop-in integrations for other platforms, PLM to design?

Oh, you bet. That's a great question. So at prismatic.io, you can go and see a list of all the products that we've already built components or connectors to. We have a super robust list of components there, and it's not specific to any one vertical. So we support CRMs, ERPs, PLMs, you name it. But if there is a product that you don't see in there, you can request it, and we're constantly adding to our library. If it's a more niche product that doesn't really make sense to add to our library, we offer a, a custom component SDK. It has the same tools that we use to build our publicly facing components. And so you can build your own, you can build your own components and they're gonna work just like any other component that Prismatic's already built.

Amazing. Super helpful. And then one last question from the chat. Akash says that they are integrated with Zapier. What would you say are the main differences?

I think the biggest difference with Zapier is that it's what we call more of a traditional integration platform. It's best if you're using internally to connect the products that you're using internally within your, your integration. Whereas with Prismatic, we are an embedded-first company. So that means that our goal and only goal is to make sure that B2B SaaS companies can offer these integrations to their customers. And in doing so, there's a lot of advantages, flexibility, and configuration that we can offer that products like Zapier can't because they're just focused on those internal integration needs rather than surfacing those integrations to your customers. Great.

Okay, well, we are at time, but thank you so much again, Marcus, for this presentation. And a reminder for everyone who is watching right now, this session was recorded, so it'll be able to be watched in the future if you need to reference it, and his slides will also be available. So thank you everyone who tuned in, and thank you again, Marcus.

Yeah, thank you, Jess, and thanks to Saastr for hosting me today. It was great.

Of course. Thanks everyone. We'll see you in future AI Day sessions.