Aired May 1, 2025

Intelligent Search and the GenAI advantage

Hosted by:

The first installment of our Futureproof webinar series saw Bud Product Manager Thomas Purton demonstrate our Intelligent Search feature and the profound improvement it can provide to the digital banking experience.

Why users open their app

The main reason for a customer to log into a digital banking app or platform is to check their transactions and account balance. In fact, one survey showed that 91% of consumers polled checked transactions on their banking app at least once per week, with 47% checking at least daily.

The problem: a clunky experience

Unfortunately, these consumers are often met with a disappointing digital banking experience and they may struggle to complete simple tasks because of limited or missing opportunities to interact with their transaction data in a conversational way.

The fix: Intelligent Search

‍Intelligent Search changes this. It helps customers to take control of their financial data, using natural language, so they can understand it fully and make the best decisions.

“How much did I save and invest so far this year?”

“Do I spend more on exercise or eating out?”

Personalized, insightful and actionable

With personalized suggested queries and smart keywords, Intelligent Search delivers clear answers about customers’ transactions, spending habits and regularity. It drives financial wellbeing for consumers and deeper digital engagement for the institutions that serve them.

What you'll learn in this webinar

Thomas explained how this AI-powered transaction search feature can help you to:

capitalize on GenAI banking that operates safely and at scale
deliver a superior user experience that’s guided by data
grow customer financial wellbeing and digital banking engagement
minimize transaction disputes.

[Iain: 00:34.1]
Hi everyone. Just waiting for a few more people to join and then we'll kick off.

[Iain: 01:00.5]
All right, I think we're ready to go. Good morning or good afternoon to you, depending on where you are in the world. Welcome to today's webinar. Really excited to have you here. I'm Iain and I'll be acting as a host and moderator today. My colleague Thomas, a product manager at Bud, is joining me to walk you through a game changing innovation called Intelligent Search.

[Iain: 01:22.2]
This is a product that's transformed how customers will interact with their finances online. Whether you're a financial institution looking for a more engaging solution for your customers or someone interested in the future of banking, you're in the right place.

[Thomas: 01:36.8]
Nice. Thanks very much, Iain, for the introduction. So, over the next 20 minutes or so, I'll be walking you through exactly what Intelligent Search is, is why we built it, how it works, and why we believe that it’s the future of banking.

[Iain: 01:54.7]
Great. If you've got any questions for Thomas, you can type them in the ‘questions’ field at any time at the top of your screen. We're going to take some time in the middle of the session after our demo to answer questions and then we'll answer as many as we can at the end of the webinar as well. Over to you, Thomas.

[Thomas: 02:10.1]
Right, so to kick things off, let's talk about what Intelligent Search actually is. At its core, Intelligent Search is a next generation tool that completely transforms how customers interact with their transactions in online banking, their mobile app or any other digital interface.

[Thomas: 02:30.2]
It's powered by Bud's core transactional intelligence and builds upon our AI models that we've developed over a number of years and trained on billions of transactions. Intelligent Search then takes this a step further by combining it with cutting edge technology like text embedding models and large language models.

[Thomas: 02:50.5]
This combination allows customers to dynamically search through their transactions and surface exactly what they're looking for. Whether it's searching for a specific merchant, tracking spending habits, or even asking direct questions, the technology is designed to deliver the answers to customers’ questions quickly and intuitively.

[Thomas: 03:12.0]
One of the standout features of Intelligent Search is its ability to provide insights. This means that beyond just presenting transactions back to a customer, it helps them to understand what their spending patterns say about them. As part of our consumer research, we found that 72% of customers wish their bank gave them more information about their spending habits.

[Thomas: 03:35.1]
For example, Intelligent Search can tell someone how much they've been spending on coffee or alert them when their Netflix subscription has increased. All of this happens organically based on the searches that your customers are making, no manual tracking required.

[Thomas: 03:53.3]
This means that, as a company, you don't need to be second guessing what is the most relevant thing to your customers. The insights are being personalised and driven based on what your customers want to know about their own finances. And because we didn't want to limit customers to searching for transactions by just a merchant or a category, we've also used LLMs, which means that customers can ask direct questions about their spending habits and transactions.

[Thomas: 04:19.3]
For example: “Am I spending more at Starbucks or Dunkin?”. Or: “How much was I last paid?” They can receive accurate context aware answers in return. So, why did we build Intelligent Search? What problems were we trying to solve?

[Thomas: 04:36.0]
If you've ever used a banking app, you've probably used it to check your balance or review your recent transactions. And you wouldn't be alone. In fact, 91% of customers polled say they check transactions on their banking app at least once per week, with 47% checking at least daily.

[Thomas: 04:54.0]
It's a task that is relatively simple, but it's also incredibly important to a customer in order for them to know and manage their finances. Furthermore, upon reviewing their transactions, 17% of customers reported that, within the last month, they have disputed a transaction with their bank that they didn't recognise, costing their institutions time and money in support costs.

[Thomas: 05:17.1]
Despite the vast majority of these disputes being for legitimate transactions, here is where we saw an opportunity to simultaneously improve the user experience for every single one of your customers during one of the most common touch points that you have with them, whilst at the same time reducing your operational costs for a reduced number of disputes.

[Thomas: 05:39.3]
As hopefully many of you know, Bud has long been a leader in transactional intelligence. Our AI models can take raw transactional data, clean it and turn it into something far more valuable – structured, insightful metadata.

[Thomas: 05:57.9]
Whether that be the coffee shop you frequent each morning, how much you spend on groceries, or the location of the restaurant that you went to last week, Bud's models can identify it all. This allows for analysis of spending behaviours, categories, merchants, locations and much more.

[Thomas: 06:14.0]
But we wanted to do something bigger. As the field of large language models continues to evolve, we saw an opportunity to move beyond just returning a list of enriched transactions to the customer. We also didn't want to create just another chatbot, as I'm sure you've all seen in the market.

[Thomas: 06:29.4]
Instead, we asked ourselves: “How could generative AI complement our existing capabilities? How could it elevate the way customers interact with their financial data?”. By combining Bud's advanced enrichment and the power of generative AI, we've built a best in class solution that takes the most common banking interactions (like checking transactions) and turns them into meaningful, insightful and actionable experiences for customers, while at the same time organically educating them about their spending habits.

[Thomas: 07:01.4]
Ultimately, we're helping customers make smarter decisions about their finances.

[Iain: 07:14.0]
Now let's dive into the fun part; some real life use cases for Intelligent Search and a demo. Just a reminder while Thomas is setting up the demo that you can ask questions at any time using the questions box at the top of your screen. We'll answer some right after this demo and we'll leave some time at the end to try to tackle some more.

[Thomas: 07:35.5]
Great. Cheers again. So of course, as you'd expect, customers can use Intelligent Search to review their transactions. So first and foremost it acts as a central list for your customers to view their most recent transactions. Here they can, of course, click into a transaction and see the various bits of metadata that Bud has applied to it.

[Thomas: 07:56.5]
In this case of Starbucks, you can see the amount, merchant name and categories that have been assigned to the transaction, as well as date, type, status, and any other information that you may want to display about a transaction. But let's say someone doesn't just want to scroll through a list of their transactions.

[Thomas: 08:14.1]
Instead, they want to see all of their spending. So let's imagine that they've been hitting up Starbucks every morning for their coffee fix. With Intelligent Search, what they can do is just start by typing ‘Starbucks’ into the search bar and instantly see all the transactions related to that merchant.

[Thomas: 08:31.5]
You may have also noticed that I didn't actually type in ‘Starbucks’, I just typed in ‘Starbuck’. Through our use of text embedding models, we don't need customers to be specific in what they're exactly searching for, whether they've shortened the word or they maybe used an acronym of what that company may be.

[Thomas: 08:52.5]
Our text embedding models can still understand throughout a customer's transactions what they're looking for and know what to filter on to display the relevant transactions to their customer. But here's where it gets more interesting. Intelligent Search doesn't just limit you to searching by merchant names; it can use any of Bud's metadata.

[Thomas: 09:12.1]
For example, consider coffee, something that I'm sure most of us drink. Now, it may be that you're not just buying your daily cup at Starbucks. You could also be buying coffee from Dunkin’, from local cafes or even grocery stores. By simply searching the word ‘coffee’, Intelligent Search will return all of your coffee related transactions across all of your merchants. And that's not all. We wanted people to search for transactions the way that they think about their finances, not just tied to merchants or Bud's list of predefined categories.

[Thomas: 09:47.0]
Take something like a car, for example. Owning a car is often one of the biggest financial commitments a person makes after their home, but the costs don't just show up as a single type of spending. There's your car payment, your insurance, repairs, gas, and more.

[Thomas: 10:02.9]
With Intelligent Search, a customer can just search for the word ‘car’ and once it loads, what we can start to see is an array of transactions relating to their car being shown – regardless of whether that's a payment to the dealership, an insurance company, a gas station, a mechanic, or whatever else that may be.

[Thomas: 10:26.2]
This helps to provide clarity on just how much your car is really costing you over time. And sometimes those hidden costs, like rising insurance rates or frequent repairs, can be eye-opening.

[Thomas: 10:43.8]
The retrieval of the transactions themselves is just one aspect of intelligent search. Let's take the example of eating out. We also provide summaries and insights into how much you're spending based on the query that you've performed. In this example, you can see that Intelligent Search has pulled transactions from various merchants in order to provide a holistic view of your eating out related expenses for the customer.

[Thomas: 11:03.8]
It's not just limited to totals either. It can also call out whether you've experienced an increase for a recurring transaction or, in this case, your most common merchant for your query. In addition to these insights, we also return smart filters.

[Thomas: 11:21.1]
Here is what you can see at the top. These provide intelligent suggestions for further refinement based on the various criteria which have been triggered for a transaction to be returned. This helps customers to very easily refine their query and explore their transactions without unnecessary friction to the journey, as you can see with this example.

[Thomas: 11:41.7]
Again, by searching, customers can then filter down to whether they want to look at all eating out and takeaways, they want to look at food and drink, or they want to look at anything else. Now, some of you may be wondering at this point about guardrails, what happens if someone asks a question unrelated to finances or unrelated to a customer's transactions?

[Thomas: 12:04.2]
Because we've limited the scope of the LLM within Intelligent Search, the results are always tied to our database of a customer's transactions. If a customer were to search for something such as ‘Should I get a credit card?’, you will see that no transactions are returned and the LLM does not give a response, ensuring you can have peace of mind in its interactions with your customers. Furthermore, any insights that are generated, if we just go back to ‘eating out’, are always linked to the underlying transactions that have been used to generate them.

[Thomas: 12:44.5]
This creates an element of explainability for you, so you can always understand how your customers have been shown the relevant insights for them. In short, Intelligent Search brings transparency to your customers’ spending and makes it easier for them to understand their financial picture.

[Thomas: 13:03.4]
That's the end of the demo, but I'll pause here in case you have any questions in the chat.

[Iain: 13:10.1]
Yes, we've got a few practical questions, Thomas. The first one is: “Can you white label Intelligent Search?”.

[Thomas: 13:18.8]
Yep, absolutely. That’s something that I'll actually touch on a bit more later on. But we offer Intelligent Search both as a backend API and also as a front end widget which can be embedded within your existing digital experiences.

[Thomas: 13:35.1]
Now what you're seeing here is an example of the widget that can be embedded within your digital experience. This can obviously be white-labelled and can be adjusted to the exact branding that you need.

[Iain: 13:50.8]
Okay, another one from Alejandro. Hopefully I'm pronouncing that right. “Does Intelligent Search work at a corporate level? For example, if I'm a company owner and I want to know the spending habits of my employees using our corporate cards.”

[Thomas: 14:08.0]
Yeah, sure, absolutely. So Intelligent Search initially was preconceived to be consumer-facing, but it can also be internally facing as well. We're in a number of discussions at the moment with clients who are also wanting to put Intelligent Search in the hands of their support centre staff.

[Thomas: 14:27.8]
So anytime that they're having a discussion with a customer, if they want to look through their transactions and find out information about that customer quickly, they can very easily do that using Intelligent Search. The same for your use case, Alejandro. If you want to look at transactions on a corporate card, that can also be done. Bud is ultimately flexible in terms of the method and format of the data that we receive.

[Thomas: 14:52.7]
Whether you have that data yourselves and you're a bank, or you're aggregating that data from a third party such as Plaid, that can all be ingested into the Bud platform and then you can use Intelligent Search on top of that.

[Iain: 15:08.2]
Cool. Just one more before we move on and that is from Emma. “Can we present offers based on searches?”

[Thomas: 15:18.4]
Yes, absolutely. I'll just answer briefly because it's something that we'll actually talk about later on in the webinar. But one of the things that we're focusing on on our roadmap is being able to create personalized and customizable insights.

[Thomas: 15:34.4]
So essentially the insights that you see at the top of the screen, depending on what a customer may search for, and depending on all of the insights that Bud knows about that customer, clients will be able to configure what insight is shown. So let's say for example, a customer is searching for a savings account with a competitor, but you have your own savings account.

[Thomas: 15:56.1]
That may be an opportune time to remind your customer of the savings accounts that you have available and you would be able to show that to them within Intelligent Search.

[Iain: 16:08.1]
All right, thanks Thomas. So maybe we can carry on. Just a reminder once again that you can use the questions function at the top of your screen to ask a question. It'll queue up and we'll cover as many as we can at the end of the call as well. Thank you, over to you Thomas.

[Thomas: 16:23.7]
Thanks Iain. So, how does Intelligent Search actually benefit financial institutions? Well, first and foremost, Intelligent Search drives improved customer engagement. When customers can quickly find the information that they need, whether it's a recent transaction or an insightful breakdown of their spending, they're more likely to be engaging with your platform regularly.

[Thomas: 16:47.8]
As we mentioned earlier, checking transactions within their banking app is something 46% of customers are doing daily. Therefore, it's something that's crucial and that you want to be getting right. By getting the fundamentals right, not only can financial institutions reassure customers that you deeply know them, you can also delight them.

[Thomas: 17:08.4]
Now, according to Accenture's 2025 report on consumer banking and advocacy, reassuring and delighting customers are two of the four drivers in creating customer advocates for your brand. And when a financial institution has advocates, they can expect to see between a 5 to 30% boost in those customers’ share of wallet and see their overall revenue grow 1.7 times faster than banks with the lowest customer advocacy scores.

[Thomas: 17:35.4]
Next up is a reduction of transaction disputes and the associated costs around them. When polled, 17% of customers reported disputing a transaction that they did not recognise within their banking app in the last month. Through intelligent search, banks can easily show Bud’s enriched transactions and, according to our research, showing a merchant name and logo increases customers’ ability to recognise their transactions by 47% – meaning less confused customers and less disputes that your support teams need to handle.

[Thomas: 18:09.3]
Finally, Intelligent Search provides data-driven insights that financial institutions can use for product development. By analysing search trends, popular queries and customer behaviour, banks can get a better sense of what their customers are looking for and what they care about. This data can inform the development of new products and services, helping institutions to stay ahead of the competition and offer better value to their customers.

[Thomas: 18:49.8]
So as we start to look into the future for Intelligent Search – and actually linking back to one of the questions that we had at the end of the demo – the future of Intelligent Search is incredibly exciting. We're just scratching the surface of what's possible in the near future. Clients will have the ability to define specific insights to be delivered dynamically based on customer searches and other criteria. For example, if a customer is searching for transactions related to savings, or maybe a credit card with a competitor, Intelligent Search can ensure that these customers are shown relevant offers for your own products, giving financial institutions a chance to capture a larger share of their customers’ wallet at the right time.

[Thomas: 19:21.1]
Or imagine this: If a customer has a personal account but is searching for business related transactions, you can use Intelligent Search to guide them into opening a business account. You can see an example here with QuickBooks. Ultimately, what is shown to your customer is completely up to you.

[Thomas: 19:37.7]
With our personalised insights, you will have the ability to define what is shown to who and when to show it, all underpinned by Bud's extensive knowledge and understanding of your customers and their individual financial situations. The possibilities here are vast and we believe this dynamic delivery method will become an integral part of the digital banking experience, allowing banks to personalise their in-app communications to their customers, not just on the specific financial situation of the customer, but also their in-app activity.

[Thomas: 20:12.9]
Now, hopefully at this point some of you are thinking about how you can integrate Intelligent Search into your existing banking apps and digital experiences. Now, as we touched on earlier, we want to offer flexibility in how our clients can consume our products.

[Thomas: 20:29.0]
This is why we offer Intelligent Search as both a set of APIs that our clients can choose to consume, allowing them full creative control over what is shown to the end user, but we also provide an out of the box white labelled widget which can be seamlessly integrated into your existing banking applications.

[Thomas: 20:47.7]
With just a single API, you'll be able to generate a dynamic URL which allows you to display Intelligent Search to your customers, whether that be on mobile, desktop or through a native app. This is aimed at clients who care about time to market or those who have limited developer resources.

[Thomas: 21:04.9]
Ultimately, Bud can be flexible to your needs.

[Thomas: 21:11.6]
So as we wrap up, I just want to emphasise that Intelligent Search is truly a game changer. By combining powerful AI technology, intuitive search functionality and actionable insights, we've created a product that empowers customers to understand and manage their money with ease.

[Thomas: 21:30.5]
But this is just the beginning. As AI and machine learning continue to evolve, we can expect even more advanced features, from real time budgeting advice to investment recommendations, fraud alerts and beyond. So whether you're a financial institution looking to offer more value to your customers, or a customer who’s tired of manually tracking your spending, Intelligent Search is here to help.

[Iain: 21:55.7]
Thank you, Thomas. And thank you all for joining today's webinar. We're really excited by the potential of Intelligent Search to transform everyday banking. We'll now turn to another Q&A, but if you're tuning out now, we just want to say that we're always on hand to provide a personalised demo and talk more about Intelligent Search. Just get in touch and we'll be happy to help. Now let's get into the questions. So, Thomas, we've got one from Harry and it is: “Is anyone currently using Intelligent Search?”

[Thomas: 22:24.7]
Sure. So we've got clients integrating with Intelligent Search both in the UK and in the US. The public API docs will be live next week, so you'll be able to take a look at them then. These clients have already started integrating and we're expecting that we'll have customers live with Intelligent Search by the end of June [2025].

[Iain: 22:48.6]
Okay. And further from Jakob: “How does the widget version of Intelligent Search differ from the API version and which one would be better for which one?”

[Iain: 23:06.8]
Sorry, it's hard to read this one. I'll just paraphrase it here. Which one would be more suitable for different types of banks or financial institutions?

[Thomas: 23:14.5]
Sure. So the difference between the API and the widget versions of Intelligent Search are minimal. Ultimately, they're both designed to be in parity with each other and should a client wish, they should be able to recreate Bud's widget for Intelligent Search utilising just our API. It ultimately comes down to how much developer resource you have and how much control as a company you want over your end customer’s user experience. We can work flexibly with clients with other widgets.

[Thomas: 23:47.9]
In the past we've had examples where, for a POC or for a limited rollout, we see clients using the embeddable widget because that's a lower barrier to entry and quicker to get up to speed. And then when they roll out into a full delivery for customers, then they go down the API route, but ultimately they work very similarly.

[Iain: 24:16.2]
Okay, and I think we have time for one more. And from Anton we have: “What are some of the examples of insights that Intelligent Search returns?”

[Thomas: 24:34.0]
Sure. So the insights available within Intelligent Search are always growing, as demonstrated. Intelligent Search can handle spending insights about specific categories, groups of transactions and merchants. It can also identify things such as regular transactions, and if your subscription prices have increased or fallen – realistically, they're probably going to be increasing.

[Thomas: 24:58.2]
It can also identify popular merchants that you may have, or a new merchant in a specific area or category that you haven't seen before. But it's one of the things that's on our roadmap, and we're constantly adding to the bank of insights that can be shown within Intelligent Search.

[Iain: 25:16.6]
Great. Thanks, Thomas. That's all we have time for on the webinar today, but if you've asked a question we didn't get to you, we'll follow up via email and make sure it's answered. You'll also receive a recording of this webinar shortly so that you can refer back to it or share with colleagues.

[Iain: 25:32.9]
Thank you everyone for your time, and thank you Thomas for your insight.

[Thomas: 25:37.4]
No worries. Cheers everyone.