Name: PromptToPower Cut
Uploaded: 2026-08-05T13:10:05.883Z
Duration: 48 min 29 s
Description: PromptToPower Cut

Transcript for "PromptToPower Cut": Hey, everyone. Welcome to our webinar today where we are going to be diving a little bit deeper into Quickbase AI features with the Quickbase Intelligence package, really specifically going, deeper into AI agents and AI actions. So thank you all for joining us here today. Super excited to, cover this topic and share a little bit more about the Intel pack and even a little bit about what's, coming next to the Intel pack. So, before we dive in and get started, a few housekeeping items. First and foremost, all of you are dropping into the chat where you're joining us from, which is great. We always love to see, where all of you are viewing from. Pretty early this morning here in Minneapolis, Minnesota, which is where I am joining from. And I know, even earlier for a few people that I'm seeing dropped in the chat. So hopefully, you've got your coffee and you're ready to go, with us this morning as we dive deep into AI. But, a few other items. One, this webinar is being recorded. So we will send that recording to all of you via email tomorrow. So if you need to, you know, hop off the webinar early or want to revisit any other content, share this with a team member, you will be able to do so. Again, that's usually done about twenty four hours after the webinar ends, so be on the lookout for that in your email inbox tomorrow. We do also have time at the end for q and a. You'll notice sort of a q and a tab by the chat, where you can submit questions. Please feel free to submit those at any time throughout the presentation. We have about fifteen minutes at the end for those questions. We'll try and answer any throughout the presentation, as we can. If we don't get to your question live today, we will make sure that someone follows up with you directly to get that question answered. Alright. And with that, let's get going here. So, first and foremost, I am Charlie Feiner. I am a customer marketing manager here at Quickbase. So, you know, my main responsibility is really to help all of you better understand the different features of Quickbase that you all can take advantage of. And so, again, today's topic is obviously the intelligence package, which launched earlier this year. So, very excited to dive deeper into that. And with me, to talk about that are two of our product experts. We have, Alex and Georgi, and I'm gonna bring them on stage now to introduce themselves. So, Alex, I will let you start and intro yourself. Thanks for joining us today. Yeah. Yeah. Thank you. Really excited to be here today. My name is Alex Pedersen. I'm a, for Deploy engineer here here at Quickbase, and I'll be walking you through the demo. Perfect. Thanks, Alex. And Georgi, I'll let you say hi to everyone as well. Hey, everybody. My name is Georgi. Folks call me g g. I'm super excited to be working on the vast majority of the features in our air pack. Very excited to walk you through what we have, been building all for the past months. Perfect. Thank you both for joining us. Here is a brief look at kind of the flow of what we're gonna talk through today. So we're gonna start off and go through just kind of an overview of Quickbase AI, really dive into the intelligence package, the different features, and Jorgi is gonna show you all a little sneak peek at what's coming next. Then Alex is gonna dive deeper into a demo, really going into AI agent and actions. Then we'll share some additional resources for getting started with AI and the Quickbase Intelligence package. And then like I mentioned, we'll have time at the end there for q and a as well. So let's dive in with that first and foremost, you know, Quickbase AI. So I think when we started, you know, AI has been a hot topic for the last few years. And when we started thinking at Quickbase of how is this gonna work within our platform, we really wanted to make sure that we were kind of differentiating from other AI tools. Right? We wanted to make sure that our AI tools were really beneficial for all Quickbase users regardless of the role that they had within Quickbase, and also that our AI tools have that same level of security and governance that you all are used to with Quickbase. I think we've all heard various horror stories about kind of the free and open AI platforms, that people are using for business purposes. So we really pride ourselves on these AI tools being built in that same click based platform that you all, know and trust and have those really in-depth security measures and controls that you all need. And like I said, we really wanted to make sure that the tools were useful for everyone, not just in specific roles within Quickbase. So the tools in the intel pack are really designed for the builders and super powerful for end users as well. So, you know, on the screen, you can see some high level, items, kind of those those high level things around, the benefits for both builders and users. And then, again, like, obviously, a lot of the security and governance is great for our realm admins who are kind of overseeing the entire platform as well. So, you know, with these tools, our builders are really able to kind of build those smarter systems faster. Right? They can accelerate their workflow creation and app creation with a lot of those AI powered automation tools as well as insights that go across, their different apps. You can also add the intelligence into the operations that you have directly. So, you know, using our intelligence tools to really create predictive models, on the operational data that you have to be able to help anticipate risk and any performance issues that, we might see kind of forthcoming. With that risk aspect as well, it does help reduce app risk and complexity. You know, the AI tools help you really quickly understand app structure and dependencies and logic to really simplify any sort of app maintenance and onboarding as well. And then, you know, deploy AI safely at scale, I think, is huge. Like I said, a lot of that security and governance aspect that we have, we really wanted to make sure that you are able to set really granular AI permissions, monitor people's usage, and really maintain that, level of audit ready oversight that, again, Quickbase has had for years and years. And then finally, we wanted to make sure that you could actually tailor AI to your business as well. You know, we didn't want the tools to just be generic. We wanted to make sure that it was, being used for your business best. And so, you know, Intellpac really allows you to upload, you know, SOPs, business terminology, other sorts of documentation to really help standardize the outputs that the Quickbase AI is giving you, across your teams, which is great. And then, you know, for those users as well of Quickbase, a lot of benefits of the intel pack. So, people can get work done a lot easier. AI really allows them to ask questions to the agent and generate reports, make various updates across apps in seconds, really with no learning curve required, which is great. People can pretty easily kind of hop in and use, those tools and make those updates across apps. Also really allows them to move from reactive to to proactive. So, you know, you can do things like track cost overruns, SLA risks, and any bottlenecks before they actually impact the business, and then you're able to kind of make those tweaks and improvements to, avoid those things. We help eliminate busy work a lot with AI. I know this is something that we talk about with Quickbase as a whole. Right? Getting rid of busy work, getting rid of, manual work, gray work, we've called it in the past. AI really kind of helps add an extra layer to that, lets you automatically extract data and create records and summarize any reports without having to do manual entry there. And then similar to, you know, the the AI controls, I keep coming back to security, and I think we will a lot throughout this presentation and when we talk about intelligence package in general. But it really allows users to trust the AI that they use. So you all can work really confidently knowing that our AI tools and our actions are transparent and also, like, secured, governed, and compliant with your organization and with the security standards that you all need. And then finally, again, using AI that just really understands you. So, kind of like we talked about on that builder side of tailoring AI to your business, The AI is really gonna understand you as a user as well, give you those insights and recommendations that are really aligned to how your company actually works and the day to day tasks that you're doing, which is great. So super high level there, lots of value. And with that, I'm gonna pass it over to Georgi, and he's actually gonna dive deep into what these features are that really lead to all of these, areas of value from the intel pack. So, Georgi, I'm gonna let you take away. Thanks so much, Charlie. I was so passionate to jump in and, like, to build on on top of what you just said, I want to start with their actions. And, as we mentioned, the busy work, gray work this is a great feature which is taking care of all the all this stuff. It pretty much handles exactly this type of task that slows down, like digging the details of an invoice, for example. A action is able to extract information from documents and create the records that you need. And it's all happening in the background. Once you set it, you'll pretty much just, read the benefits out of it. Very good use case for your actions that they often, come back to is the invoice case. So for instance, if we have an app which is taking care of customer relationship and we need to take, keep track of all the invoices that we we receive, and we need to know the amount, due dates, who do we need to pay or who need to pay us. Like, this is this is all information that somebody needs to manually dig in. But if we if we set up a pipeline and now the AI action is a step, we can actually get taken care of that and never never go back. Like, it can be fed inside a a text field, for example, and then we forever keep the keep the benefit. This is a sample set of screen, but you can now access it via the customizing step in pipelines. It's fairly simple to use. It has several steps. And if you click around, it's it's very easy to to hit the ground running. I personally started with some simple use cases, but it can also support some even more complicated ones. So we have good documentation and and, of course, like, you can, fine tune it for the cases that you need to the most. Next on the list, is the data analyzer. And what's specifically cool about this feature is that it's based on machine learning algorithms, and we invested a lot in, the mathematics. It's pretty much based on buckets of mathematics. It's spanning it's using data from the whole app in order to make predictions about the likelihood of an outcome. And it's not one size fits all. It's not something that was taken off from somewhere else. It's it's a model that we fine tune for exactly the data that's being stored inside the app. And the insights that are being eventually created, they are, they're allowing for proactive decisions. So for instance, we are managing managing deals. We can pick base up. You see, like, we put a big closing. You can see, like, we put a delays, like, we put a few incidents. Like, it's it's really helpful. And the best part is that we've also have the visuals to make a representation of how how often within the data we see certain events occurring. We also see factors, which influence an outcome. So for instance, if you are tracking deals and we can see if, for example, if you are a sales rep, maybe the deal is more likely to close or to not close, which is really cool because you can point your attention to exactly what matters most when you're making decisions. And data analyzer also provides for very good, cross team alignment, alignments with leadership, prioritizing with what matters because it's based on data and based on what we we need as a basis to make decisions. And, last but not least, I want to emphasize that the data analyzer, he has its own UI in app settings. But, this Thursday, we should be dropping this availability to consume data analysis bits directly inside the AI agent, which I'll tell you more in the next slide. After I alluded in my previous points, the AI agent is is not a simple feature. It's not a flaw feature. It's actually a brain of feature. It's a conglomerate. And the biggest value added that the agent provides is that pretty much it's a it's a jackhammer of everything to do inside Quickbase. If you previously need to do a manual app building, now we just have to, utilize a couple of clicks inside the agent or even a query will do, an English language query. If you have to manually look through the data, make and export, feed it inside the AI in order to consume any bits of, of a graph or a bar chart or, like, say how how is the average moving over time. This is this all can now happen inside of the structure. Again, only language queries are necessary. And the best part is that gives results of of four fingertips. We don't need to click to variety of interfaces anymore. It's it's all it's all integrated. We and and later in the slides, I'll tell you more about how we're making it even even more capable, in in as we speak. But in general, if I want to make a comparison, and now use myself as an example, I joined Quickbase two years ago, and they get to do a bunch of stuff, manually. And I got Sentry certified, and I build my first ops. And then I was just, like, almost, like, on my own. But now with the agents, we have new colleagues joining in, and their onboarding is it's much easier because they kept everything connected with their fingers. And we also have a skill which provides, product help, fine tune to Quickbase, help center articles so you can stay inside the product while doing, like, everything you need. Like, you don't need to leave quick base in order to get help. It's it's all there on the tip of fingers. One of the, other elements available via the agent is AI app intelligence, and that's also one of the one of the features that really allowed us to cut down to insight. So imagine you inherit an app, and that that that's actually something that also happens to me. I inherit an app every now and then. And sometimes, apps are simple, but, oftentimes, they're not. They are big, gnarly apps. We have apps being built for, like, five, ten, fifteen years, and I need to take care of a lot of things in my day job and onboarding in a in a new job takes a while, sometimes takes days, sometimes maybe weeks. Even that is too early. And we realized that's a very common experience when an ops change ends. So we collected, bits of feedback and realized that there is a huge demand for a feature which provides very quick on demand app documentation, which will really help you with time to insight, understand what an app is about, relationships, roles, what pipelines are there, information about the pipelines, pretty much all of the core information that you need to make sense of an app, and it's now available with app intelligence. I click up a button about a minute of waiting, then it's and it's being generated. And the best part is that you can also export it and share with, with your team. So, yeah, it's getting time to insight, and we are also here folks in the team. They are really happy about the fact that they don't need to click through hundreds something sometimes of tables and piling through a bunch of relationships to make sense of an app itself. Super fast. Speaking about speed, I also want to mention, couple of notes about control because control is equally important. Air control center is something that we released, a bit earlier in 2026. And, previously, we only had one very big master AI toggle. It was either on or off. And this didn't really allow for granular control on which feature is avoidable in the rail, which one. This was it was one. It was all or nothing. Like, it's it wasn't ideal. We realized that that's not, that's not a good way to present AI. So we listened to our feedback, and we we wanted to make sure that now, AI is being able to be adopted at a rate and at the amount that your specific use cases finds reasonable. So now it's perfectly possible to enable the AI agent, but not to utilize stuff intelligence or the opposite or enable and allowing calling for data analyzers. We've we've made sure that this permutation of use cases can work in regardless of what's on and off as long as AI is generally on inside the realm. And the best part is that with the permissions page, you can also utilize different, users and groups, and you can also be fine tuning, which user group has access or doesn't have access to to a specific feature. So you can also get even more granular than that. And right now, it's it's in a stage where we are we are better positioned to support the the various use cases that may come with with AI usage. I'm I'm closing with it for the existing, Intel part features, but then equally as excited about what's coming next. And that's the first bit is the knowledge layer. You might have heard me say in previous webinars that that's something that's really, really powerful, and it's because I really hate going through files to files in order to find the information I need. This type of feature will make it possible to find SOP information, onboarding manuals, repair procedures, checklists, compliance documentation, like, how do I request vacation days? What does several cost $3.05 5 mean? Directly, case you want to wire the AI agent. I wouldn't need to rely on tribal knowledge, historical Slack messages, going to conference, going to files of folders of SharePoint. Like, all will be available, the tip of my fingers, language query inside orchestrator. Like, all will be consumable and it will be really, really cool because you you wouldn't need to have to spend and waste so much time trying to figure out where information is hiding for you, and you'll be able to find it really fast when you need it the most. And that's why it's something that we we are prioritizing, and we're really looking forward to this release. Likely, first version is gonna come, for the end of the year, So stay tuned and also more than welcome to share feedback with us this drops. Second on the coming solution, the schema management. And if you if I really want you to remember something about it is that you'll be able to build in the shape up structure to simple conversation once this drops. Currently, the agent is not, able to update and delete reports, articles, and forms. And with this up to a feedback that these are exactly the three elements that you need support the most. So we've realized that it's worth enhancing the agent, and the use cases of renaming, updating the relationship, removing the table, cleaning up on use forms, they will be supportable by this initiative. So it's also it's one one more muscle in the already, carry category agent that that we have. Last but not least on the list is the workflow agent, where you'll be able to review, activate, and stay in control of a completely ready to run generated workflow, which is also named for pipeline. You it wouldn't be a shell. It wouldn't be a scaffolding. It will be asking you the right questions before building. It will run a simulation and validate whether the information is being touched the right way. We need to understand your app. So you'll have your own assistant, which will build your pipeline for you, and everything will be taken care of. And that, we believe will also position the whole AI experience in an even more convenient matter where we'll be taking care of all the nitty gritty type of fine tuning things, where you'll be able to take care of what matters most and spend spending time not necessarily, building things on your own, but being supported and being carrying a partner in all this journey of maintaining your applications and keeping track of your processes and fine tuning whenever necessary. In summary, we have five really cool features already available in the AI part. We see, good adoption numbers and we see that, faults are finding pilot of those. We made an outlining of the general use cases, but as everything is quick based, we support the multitude of those cases. We cannot outline in a single session. So feel free to send us your feedback, what you like, especially I'm open to things that you maybe not like so much, and we'll be able to take care and prioritize and shape these features based on what can serve you best. I really want to take out for the presentation and stay here for the call session, but I'm happy to to pass the mic to to Alex to show you more about some hands on implementation as well about what I just told you. Awesome. Thank you. Alright. I'm gonna go ahead and pop open my screen here, and I'll hop into the demo for us. Every operations team has some version of this. Requests come in from customers, from the field, from suppliers, and somebody has to read each one, figure out what it is, decide how urgent it is, route it to the right person, and then remember how they solved the last one like it. Now multiply that by a few 100 a week. That's where the time goes, not in the hard problems, in the reading and the sorting and the remembering. So the question we get asked constantly is this, where does AI actually help here? And where is it just hype dressed up as automation? That's what I wanna show you over the next twenty minutes. I'm going to split this into two kinds of AI because they feel completely different in practice. The first is background AI. This runs inside your workflows unattended. It does the work the moment something arrives. You never open a chat window. It just happens. In Quickbase, that's AI actions running inside pipelines. The second is foreground AI. This is AI you talk to on demand to explore and shape your data in the moment. That's QB chat. One works while you sleep, the other works while you think. And here's a promise I'll make upfront and keep proving as we go. At every step, I'm going to tell you which part is genuinely AI and which part is just plain automation. Because the fastest way to lose a technical audience is to wave your hands and call the whole thing AI. Most good systems are mostly deterministic plumbing with AI doing the two or three steps that actually need judgment. That's a feature, not a confession. You don't wanna pay an AI tax on the things regular automation already does perfectly. Let's get into it. The raw request. Here's a request that just landed. Look at what we have, a subject, a description, and that's basically it. No category, no priority, no summary, nobody assigned. This is a customer writing in plain English. Normally, somebody reads this and does the triage by hand. Read it with me. Three of the series 40 actuators from our last order are sticking around 60% travel on line two. Started yesterday, and it's holding up production. So a human looks at that and thinks, this is a product defect. It's urgent because production is down. It should go to field ops. That judgment is the work. Let's watch the AI do it. All I'm gonna do is flag it for triage, one checkbox. That kicks off a pipeline. While it runs, let me tell you exactly what's happening underneath because this is the honest part. Two things are happening here. Only one of them is AI. The AI step reads the description and returns four things as structured data. The category, the priority, a confidence score, and a one line summary. Then a second step. Pure automation, no AI, takes that result, sets the status, and routes it to the right owner. The AI does the judgment. The plumbing does the plumbing. And there it is. Category, product defect. Priority, critical because it caught the phrase, holding up production. Confidence, 95. And a clean summary it wrote itself, describing three sticking actuators on line two with a production impact. Status flipped to triaged. That took about eight seconds, and no person touched it. One thing I want you to notice, the priority isn't random. It read urgency and business impact right out of the text. If I'd sent in a routine billing question, this comes back low. Let me actually show you, because one example is a magic trick, and three is a pattern. The billing one comes back billing question, medium, and the summary names the exact invoice. The quote request comes back RFQ, same pipeline, same prompt, different judgment each time, driven entirely by what the text actually says. Classifying is useful, but the thing people really spend time on is figuring out how to resolve it. And I wanna be careful here because this is a different capability with a different risk. So I'll check a second box, draft resolution. Here's what this one does. First, a plain automation step searches our own history. It pulls the closed requests for this same product and category, the ones we've already solved. Then the AI takes those real past resolutions and drafts a recommendation grounded in them. A query can find history. Only the AI can synthesize it into a draft. That's the line. Read what it wrote. Replace the actuator seals under warranty. Issue an RMA. Investigate the root cause, likely a meridian seal batch variance. Now that specific phrase, seal batch variance, did not come from nowhere. It came out of our own closed records. And this is my favorite part because it kills the hallucination question before anyone even has to ask it. Look below the draft. These are the exact past records it reasoned from, every one of them. If you wanna fact check this recommendation, you don't have to trust the AI. You go read the source records yourself. The AI wrote the pros. The system stamped the receipts straight from the query, so it can't cite something that isn't there. In a regulated or high stakes environment, that's the difference between a demo and something you'd actually turn on. So before we leave this first half, let me keep my promise and name the parts. The AI did the classification, the priority judgment, the summary, and the resolution draft. The automation did the status change, the routing, and the history search, roughly half and half, and that's the right ratio. You're spending AI exactly where judgment lives and nowhere else. That was background AI, running your process while you're not looking. Now let's flip to the other kind, AI you sit down and talk to. Same app, same data, but now I'm going to have a conversation with it. And the thing to watch for isn't any single answer. It's whether it holds a thread, whether it feels like talking to someone who's actually following along. I'll start simple just to get oriented. I ask what tables are in the app, and there's the structure straight from a plain English question. Now something real. Show me all open requests with high or critical priority. Nine of them. And notice, it didn't just dump the table. It pulled the columns that matter. Priority, status, who's assigned. And by the way, one of these is the record we triaged sixty seconds ago in the pipeline. The background work and the foreground conversation are looking at the same live data. Here's the test. I ask, which of those don't have an owner assigned yet? I said those. I didn't repeat the query. I didn't respecify the filter. I'm relying on it to remember what we were just looking at, and it did. It took the nine from a second ago and narrowed to the two with no owner. That's the whole point of foreground AI. It isn't answering isolated questions. It's in a conversation with you. That's how a real person works through a problem, and it's genuinely useful because you can think out loud and it keeps up. Let me ask something an analyst would normally build a report for. Which supplier has the most product defect requests? And it went and counted and drew the chart. Meridian Components, 14 product defects, way out ahead of everyone else. If I'm running this operation, that's a supplier conversation I need to be having, and I got there in one sentence. One more. What's the average days to resolution by category? Look at the spread. Billing questions close in about a day. Warranty claims take three weeks. Product defects sit in the middle, around fourteen days. That's a performance picture I can act on. And, again, I just asked for it in plain English. Remember, in the first half, the pipeline drafted a resolution from history. Here's that same history, but now I'm pulling it up myself, live. I ask for the last five closed requests for the Meridian Series 40 actuator and their resolution notes. Five closed records, same product, and there are the notes. See how consistent they are? Everyone traces back to that same seal batch variance. This is institutional memory that was sitting in your closed records the whole time, and now anyone can surface it just by asking. I'm not going to ask it a question. I'm going to ask it to build something. Create a formula field called days open that returns the number of days a request has been open, and it just builds it. Field created, formula written, live in the app. That's normally a five minute trip into the builder done from one sentence. So it can build a field from one sentence. Let me show you where that actually goes. Here's the problem I want to solve. Back at the start when we triaged that actuator request, it got a category and a priority, but it never got assigned to anyone. Assignment is its own decision, and it's a harder one than it looks. You need to know what kind of work this is, which team handles that, and who on that team actually has room right now. So I need two things, a switch to kick it off and something behind the switch that makes the decision. Let me build the switch. One field, one sentence, same as before. But a checkbox is just a switch. The interesting part is what it's wired to. Four steps. The trigger fires when that checkbox gets checked, then a search pulls my entire team roster with each person's current open workload attached. Then, and this is the step that matters, an AI action looks at the request and looks at the roster and picks the owner. Then a plain update writes the assignment back to the record. Four steps. One of them is AI. Same ratio I promised you at the top. The trigger is plumbing. The roster search is plumbing. The right is plumbing. The AI does exactly one thing, the judgment call, and nothing else. Let me read you what I ask it to weigh. Given the request and the roster, which team handles this kind of work and who on that team has the most capacity? Prefer the least loaded person on the right team. That's a judgment call. It's not a sort. A sort would just find the smallest number and hand me whoever that is regardless of whether they do this kind of work. And notice what I did not do. I didn't write a routing table. I didn't map categories to teams. I didn't hard code a single name. Let's run it. Here's our actuator request. Still unassigned. Same record we triaged at the top of this demo. And here's the roster it's about to look at. Seven people, spread across a few different teams with wildly different workloads. Nothing in this table says who should get an actuator problem. Save. Marcus Hale. And here's the part I actually care about. It wrote down why. Read that. Support team, which handles product defects, and the lightest workload on it. So it did the two things I asked for in order. It figured out what kind of problem this is and which team owns that kind of problem. Then it looked at capacity inside that team. That's the difference between judgment and a sort. A sort finds the smallest number in the column. It has no idea what a series 40 actuator is or who fixes one, and the reasoning is in the record permanently, next to the decision it drove. Same move you saw in the first half. The AI makes the call. The system writes down how it got there. Six months from now, when somebody asks why this ticket went to Marcus, the answer is right there. Let me bring it home. Background. AI ran your process the moment work arrived. It classified. It drafted a grounded answer. It showed its sources. Foreground. AI let you turn around and interrogate that same data in plain English and then build on it. Same platform, same data. And as you just saw, the two halves can build each other. And I kept the seams visible on purpose. The honest version of AI in your operations is not a magic box. It's your existing automation doing what it already does well with AI dropped in at the exact points where judgment matters. Reading a request, weighing urgency, drafting from precedent, knowing who should pick up the phone. That's the version that actually survives contact with a real business. So where do you start? If you're already building on Quickbase, this is a pattern you can apply to a real workflow in your own app this week. If you're earlier in the journey, pick one process that passes three tests, high volume, judgment heavy, and you already have a pile of closed records for it. That third one matters more than people expect. The AI wasn't smart about seal batch variance. Your closed records were. The AI just read them. Find that process and prove it there. I'd love to help you scope the first one. Thanks for watching. Let's see. It's for are you hopping on? Yes. Sorry. It took me a minute to shift from the backstage there, but, thanks, Alex, for walking us through that. And, yeah, we could yeah, Alex. I know, again, wanna spend the last couple of minutes here just talking about some other resources that we have, to really help with Quickbase AI and the intelligence package. So I'm gonna, like, let you talk a little bit about, your new role here at Quickbase and kind of the new departments that, we've created really in response to, Quickbase AI and the intelligence package, the forward deployed engineer team. So, Alex, take it away from here and tell tell us what a forward deployed engineer is. Yeah. Absolutely. So, yeah, it's it's been so exciting, just working in this role so far as the forward deployed engineer, working with customers very closely on their AI use cases. Not every use case is, I would say, qualified where it kinda meets some of the requirements that we look for, on that end. We, as FTEs, we say that upfront. As an example, if there's, let's say, a workflow that could potentially have AI, and we'll investigate that. We'll kinda do a deep analysis. So we would look at, let's say your app, the schema in the application. Is it ready, like, data readiness for the app or for, like, AI specifically? And if it is not, we'll say that we'll say that, you know, here's a better route, to get exactly what you're looking for. But if it is AI ready, then where the for Deploy engineer comes into play is we want to really help and stand up our customers to understand the value of AI. So we're able to, you know, if if you're able to invite us to an app and we have a couple working sessions with you, we typically like to stand up a couple use cases within your application around, those AI workflows. So, we're really working kind of side by side as a partner with you. And we're as the slide here, you know, shows, we're very outcome driven. And so we, again, take a lot of time through that discovery process and understanding the envision and the goal, and then we build the entire pipeline for you. But we don't necessarily just wanna build it. I'm a huge fan of also coaching and enabling our customers, why I built the AI stuff the certain way that I did, why I used the prompts that I did, because I do believe that really enabling you all on how to use this properly and build more is is really where the power's at. So that that's where the for deploy comes in is to help our customers really excel with some of these new, AI features. And let's see if we can pop over to the next slide here, Charlie. Just looking for that button. Oh, cool. Okay. These these are just kind of some high level of the value of the what the FTE unlocks. Definitely, so, like, improving safety, where we can predict, prevent safety incidents by identifying leading indicators. We can do that with AI. As an example here, like a certification gate placement assistant blocks operators from unsafe equipment assignments before the shift starts. Now an example of this might be, let's say, there's a a construction worker there. They wanna use a certain, machine. And if, you know, the AI like, if you actually go through and let's say you're trying to, add certain people to certain areas, even in a warehouse as an example, and only certain folks can go to certain stations based on certification level or years of experience, the AI can very kinda quickly look at your personnel and then identify and then, you know, auto assign based on that information that you have. Another example, so tightened processes. So billing, happens faster, more accurately. So this is another good one that I've seen where an AI action flags unbilled work the moment a job closes. This way you don't you're not kinda waiting for, invoices or the invoices don't get missed as easy. You know, increase profitability. Yes. Win more work. Okay. Another big one. Yeah. Surfaces historical insights from your quick based apps to improve bid competitiveness. Another example. In this example, we have asked past projects in plain English. Questions like, what's our win rate on this job, or why? In interesting facts, you can start to see, like, a based on, you know, these certain people that are always on the projects, we seem to move those projects faster or close faster. And this is the big one here. The reduced manual worked. So you can automate a lot of those repetitive tasks. Just like in the demo that I was, trying to showcase there was the ability when a request comes in, I wanna I wanna streamline that process leveraging AI. Also, leveraging the past data and history that we have for those resolutions. So it just really increases that that manual work. Faster decisions? Absolutely. And and there was a good example in that demo too. And so instead of digging through, you know, a bunch of data or, different apps, you can, you know, either ask the AI, in plain English through the chat. Hey. You know, can you pull up the top five, requests that match this criteria that are already closed? Right? Or who would be the best personnel to add to this type of request and why? So you can get those faster decisions as well. Cool. Now one of the questions, you know, am I a good fit for an FTE? And very you know, most likely, yes. There are some certain qualifications. What I would always recommend is that you would actually reach out to, let's say, your account executives or your support agent that you have, tied to the realm, and you can actually go through that process and discuss with them. But I do have a couple key questions that I like to ask or if you were to think through on your own. One, is is yourself or anyone on your team retyping data from PDFs or forms? Right? So like Georgi mentioned, if you have a lot of invoices coming in or if you're needing to look through a lot of different PDFs and then put that data into a quick based record, then very much so, yes, you're probably a good fit just with that use case alone. Are there a lot of decisions waiting on one person's memory? So do you have a do you have someone with kind of that internal tribal knowledge that just, you know, wealth of knowledge inside? And if, let's say, they're out for a week, is it kinda chaos? Right? And so if if there's a lot of decisions typically depending on just that one person, then that's a use case that we could really dig into and and figure out where AI could fit there. The third one there, are teams assembling updates by hand before calls? So, you know, if they were looking through and they're trying to, again, summarize certain notes or information before, let's say, they had they speak to a customer or a client of yours, this is another great use case for AI to be able to bubble that up and and create a summary there. And then what are we currently using? Are we currently using AI, or where? Right? That's another one, I'll just dive into too. Are you already leaned into AI within an organization? How are you using AI internally? I always love to hear from customers on on what you're what you're doing. But, also, if you're a builder, then I think it's AI is another great use case just with how much, it can help you speed up the build process, and whatnot, like formulas and tables and fields. But, long story short, reach out to your account executive or your support team, and then they can talk you through how to get in touch with the FDE team. Awesome. Thanks, Alex. Very excited about, the FDEs, and it's been cool seeing, all the good work that you and the team have done already with that role being so new. A couple other resources for the Quickbase Intelligence Pack while we're here as well. On the screen, you'll see these two QR codes. One on top is the sixty day free trial. So if you are an existing Quickbase customer, you do have access to that sixty day free trial of the intelligence package. So we highly recommend, trying it out if you have not yet. Again, for Quickbase customers, it does include all of the AI features that Yorgi walked us through at the beginning of the hour. It's free for sixty days. If you are a Realm admin, you can actually, begin that trial directly. If you are not a Realm admin, please reach out to your Realm admin or your account team at Quickbase, to discuss that and help, with getting started with that. And then on the bottom of the screen, we have a crew meetup in a couple of weeks, so August 18. If you are not part of the crew or haven't been to any of the meetups, highly recommend. They're super beneficial. There's usually a few each month kind of going through different topics and, you know, sharing best practices, resources, all that good stuff. And, you know, since we did this webinar here today, our next crew meetup is actually really tailored around the quick based intelligence package, which is super exciting. So, you know, we'll be kind of discussing that a little bit more in-depth. We have a couple of our customers who are gonna share their stories of how they're using, AI agent, AI actions, and the rest of the intel pack kind of in their realms and in their real world day to day. So super excited for that. It's gonna be a great discussion. So we would love to see you all join us for that as well. With that, I know we are at time because we had so much to go through. So, unfortunately, we will not have time for a live q and a today. I know Vencey was answering a ton of the questions directly in the chat, so hopefully, most of those questions actually got answered directly to you all real time. What we are going to do though as well since we ran out of time for q and a, we will put together a summary of the questions that were asked and kind of those answers and make sure that gets sent out to all of you as well just in case you weren't able to go through that q and a and see the answers, to the questions that people were asking. So we will make sure that that is an available resource for you all. We will send out the recording of this webinar, tomorrow once that is all processed and everything. So be on the lookout for an email from us. Again, that email will include the recording, that q and a summary that I just mentioned, the demo video from Alex that he mentioned as well as well as some other resources around the intelligence package. So, keep your eyes out for that email tomorrow. And with that, that is all that we have today. So thank you all so much for joining us. Thank you, Yorgi and Alex, for joining us to present. Thanks to Vencey for answering a bunch of your guys' questions, backstage. And thanks, everyone. Have a great rest of your day, and we will see you on the next webinar. Thank you.