Posted on

Small bank, big moves: Why a bottom-up strategy beats rip-and-replace in Gen AI

In our last article, we covered why Gen AI can help small banks remain competitive and improve six specific use cases: fraud detection, customer service automation, personalized marketing, document and data processing, knowledge bases, and operational analytics.

In this article, we will explore what implementation strategies work the best for these use cases, how to align the culture piece with the technological adoption to render plans into reality and build for competitiveness and efficiency. 

The implementation playbook

Complete overhauls are difficult to undertake and don’t work well for Gen AI implementation. Gen AI is new, and the regulations, technological infrastructure, and providers around it are still developing, making a rip-and-replace approach  risky. 

Therefore, institutions with less than $10 billion in assets would be better served if they think bottom-up. That means identifying point solutions and specific use cases that are likely to have a sizable impact. Learnings from these roll-outs can then be used and workshopped to build a firm-wide strategy. 

Ryan Lockard, Principal at Deloitte recommends the following six-step plan:

1) Identify the right use cases: Undertake an organization-wide review of which processes are likely to deliver the highest ROI post-Gen AI implementation. 

2) Find the right partner: Choose consulting partners, hyper scalers, and/or fintechs that have banking and compliance expertise. 

3) Go API-first: Undertake an API-first integration strategy because it aids with shorter deployment times and significantly reduces disruption to legacy infrastructure.

4) Dip your toes first:  Start out with a pilot project and ensure that you’re measuring and tracking ROI, so your organization can learn as much from these tests as possible. 

5) Stay on top of regulations: The regulatory climate around Gen AI is likely to change considerably as the technology evolves. Small FIs can benefit from working with partners and technology providers that have responsiveness to regulations built into their systems. 

Apart from what Lockard recommends, it is important for FIs big and small to recognize that Gen AI is no longer just a technological tool. It’s changing how people think of and structure their world. While the biggest players in the industry are busy touting how great the tool will be for their bottom lines, many fear job loss. In such a climate, effective, informative, and considerate communication with employees is key to ensuring that Gen AI is actually adopted and intelligently leveraged in the office. 

6) Communicate learnings: Executive adoption and ownership of a Gen AI strategy is key to getting buy-in from more junior managers and the rest of employees. 

Effective communication across the organization can help with adoption as well as provide opportunities for teams to identify which of their own processes could use a boost from Gen AI integrations. 

How to tackle change management 

Gen AI is not just changing how employees work, but is likely to reconfigure how work is done altogether. So, it’s important that employees see executives champion and lead the way forward with Gen AI usage, according to Lockard. This should involve sharing their personal observations and experiences of using the tech as well as clear communication around the importance of the vision that led to integrating Gen AI.

Firms also should recognize that Gen AI implementation may have to come with an educational component, helping employees to train and understand how the technology works. Enabling employees to overcome this learning barrier can bolster trust in the organization while also upskilling existing staff, rather than looking for new talent with AI expertise, where competition is already fierce, according to Lockard. 

How to re-work hiring requirements

As Gen AI solutions become a part of the technology stack, small banks and FIs will have to rework their job specifications to include digital and partner management skills, along with vendor management, data literacy and cloud integrations for more technical roles, says Lockard. 

However, unlike big banks that have started to build large AI teams in-house, small banks would be better served with maintaining a select group of internal stakeholders that can “effectively orchestrate and govern AI solutions delivered by trusted partners. Once hired, commitment to continuous learning is essential, ensuring existing staff are regularly upskilled on the latest AI tools and workflows to stay agile and competitive in a rapidly changing landscape,” he shared. 

Micro case study: How Bangor Savings Bank built an employee-centric AI-strategy

Through its partnership with Northeastern University’s Roux Institute, Bangor Savings Bank announced a two-year “Accelerating Insights” program which will help build data fluency and skills in ethical AI usage among its 1100 employees. Bangor staff will get access to bespoke learning modules built by The Roux Institute, a research center affiliated with the university. 

The program’s curriculum was developed through an analysis conducted by The Roux Institute, which examined organizational documents like employee role specifications and performance assessment standards, according to Liz Kohler, Managing Director of Strategy, Operations and Growth at The Roux Institute. Through the educational program, Bangor’s employees are building data fluency skills along with learning to use AI ethically. 

The bank hopes that in the future, through the upskilling achieved by the program, it’ll be able to undertake much more ambitious Gen AI programs, having improved the baseline skills and performance of its staff. 

Posted on

Small bank, big moves: Six Gen AI use cases that move the needle for small banks (and three that don’t)

Doing more with less is the small bank mantra but burdened with legacy tech and consumers’ preference for digital experiences, small banks and FIs have their work cut out for them. 

Especially when it comes to Gen AI. 

On the one hand, the potential Gen AI offers for increasing competitiveness, CX, and efficiency demands action, on the other, are the constraints of time, money, and resources, bogging down any meaningful discussion on AI strategy. 

But small banks must find a way to break free of these chains of legacy and size, especially when the biggest players are making billion dollar investments in Gen AI and capturing thousands in the talent pool.

This article is the first in a series of two focused on how small banks can capture the competitive advantage of Gen AI. In this edition, we will discuss what use cases may supercharge small banks’ efficiency and customer experience and which use cases aren’t likely worth the investment for firms under $10 billion. 

Use cases that work

With limited resource availability, small banks and FIs need to identify areas where they can get the most bank for their buck. 

Six use cases emerge here, according to Ryan Lockard, Principal at Deloitte:

i) Fraud Detection: Bad actors are already using Gen AI to game financial systems for their gain. Criminals were able to defraud Americans out of $21 million between 2021 and 2024 using voice cloning technology the efficacy of which has been significantly improved by  modern LLMs. Most big organizations are fighting this fire with fire, and using Gen AI tools to identify new types of fraud tactics as well as improve their fraud detection systems. Small banks can use these same systems to adopt real-time fraud monitoring with automated alerts to stay ahead of criminal activity.

ii) Customer Service Automation: Bank of America is about to enhance its award-winning digital assistant Erica with Gen AI. Small banks and FIs can learn from this playbook, and integrate Gen AI technology in their digital customer service agents to drive better response times and enable their staff to focus on higher-value processes.

iii) Personalized Marketing: Smaller FIs often operate with a limited marketing budget and team. Here, Gen AI tools can help better audience analytics by improving segmentation and targeting as well as help FIs create more with less through content generation. 

For example, Duke University Federal Credit Union (DUFCU) recently integrated Vertice AI’s copywriting tool called COMPOSE. “The marketing team can prioritize delivering high-quality content that drives new member growth. COMPOSE is equipping us to elevate our standards of excellence, while streamlining our efforts, ensuring our acquisition campaigns are highly personalized, on-brand and efficient,” said DUFCU’s Director of Marketing Jennifer Sider. 

Industry-specific tools offered by financial technology providers like Vertice are a critical differentiator here. They are trained to be compliant with financial services regulations, and can keep up with evolving regulations with relatively low-lift from FIs, and also learn from internal material to ensure messaging is in-line with the tone of the firm.

iv) Document and data processing: There is a misunderstanding in the market that just because data is available, lenders and FIs have the capabilities in place to be able to utilize it effectively.  61% of lenders report being overwhelmed by the volume of data available. Gen AI can prove to be of significant value here. 

For example, the $2.5 billion, Kentucky-based Commonwealth Credit Union integrated a tool by Zest AI called LuLu Pulse, which uses Gen AI to consolidate multiple data sources like NCUA Call Reports, HMDA, and economic data, allowing the firm to gain insight into how their products and services compare to their peers. 

Additionally, Gen AI-driven KYC processing can also accelerate onboarding, improving customer experience and minimizing the chance of errors. 

v) Operational Analytics: Small FIs can also use Gen AI to take a closer look at the health and efficiency inside their organization. Gen AI powered operational analytics can help firms identify process bottlenecks and improve resource allocation, allowing these firms to build as much efficiency into their lean workforce as possible. 

vi) Knowledge bases: Access to Gen AI-powered knowledge bases can prove to be useful for small teams, offering them quick access to information around internal policies, simplifying employees’ workflow. Banks like Citizens are already implementing such tools allowing everyone in the management chain to access information about topics like employee benefits.

Use cases that DON’T work

Given limited resources and time, small FIs need to make sure that their approach to Gen AI integrations focuses on use cases that are meaningful. 

Highly complex and infrequent: Processes like complex lending decisions, where human expertise and understanding play a big role, aren’t suitable for Gen AI implementations. Low-volume, high-complexity tasks are also unlikely to yield ROI for small FIs, according to Lockard. 

Poor data quality: Firms must also keep in mind that Gen AI’s output is only as good as its data. So any use cases that hinge on poorly structured legacy data aren’t a good fit for Gen AI implementation, shares Lockard. When assessing whether a  use case will benefit from Gen AI implementation look for well-labelled and annotated because it allows Gen AI models to learn more quickly and produce better outcomes. 

Code generation: Additionally, while people are getting excited about Gen AI’s ability to write code, only firms with in-house development teams may be able to fully leverage such features. Without in-house technical expertise to properly vet, customize, and maintain AI-generated code, organizations may face security risks, integration failures, and compliance issues that far outweigh any potential benefits.

Gen AI use case suitability checklist: 

  • Does the use case occur frequently enough to justify investment? 
  • Is there a clear goal and KPI that would help measure the impact of Gen AI integration? 
  • Would the lack of human intervention in this use case severely negatively impact results?
  • What mechanisms are in place to take corrective action in case something goes wrong?
  • Who will be held accountable for mishaps? 
  • What regulations impact Gen AI usage in this use case and is there tolerance for regulatory action against the organization?
  • Is the underlying data infrastructure and data ready for Gen AI integration? 

Gen AI implementation holds massive potential for those small FIs that can rally their C-suite and employees to adopt the tech.  

“By automating manual tasks, Gen AI drives operational efficiency, reducing both costs and error rates. It also transforms the customer experience, delivering personalized, always-on service that can rival what the largest institutions offer. But perhaps most importantly, cloud-based AI solutions empower small banks to bring new products and services to market at a speed closer to that of Universal Banks and GSIBs, closing the innovation gap and leveling the playing field.” 

— Ryan Lockard, Principal at Deloitte

Identification of the right use cases is the first step in building a Gen AI strategy, stay tuned for the second part of our series to learn how to get started once the feasibility studies are over.

We will cover: 

– Implementation strategies
– Change management
– Role of technology providers
– Talent acquisition 

Posted on

“Take a hard look at your current ecosystem. If you were to double the assets under your management today, would your current ecosystem sustain that growth?” Finastra’s Kristen Lista, on what FIs need to do to compete in SME lending

Non-bank financial institutions (NBFIs) are capturing more and more market share in SME lending by leveraging technology to offer quicker lending solutions. This puts pressure on FIs to evolve their approaches while managing costs and improving service quality.

Finastra’s Principal Product Manager Kristen Lista joins the Tearsheet podcast today to discuss the most critical areas where FIs need to focus: consolidating technology to improve efficiency, decreasing the time between application and access to funding, enhancing back-office operations, and creating more client-centric experiences. 

Lista offers a valuable look inside the complex web of challenges that FIs are facing when trying to improve the SME lending products. From technology integration strategies to practical advice on process improvement, Lista offers an actionable blueprint that can help FIs better compete in the SME lending space, driving growth and customer loyalty.

Watch the full episode

 

Listen to the full episode

Barriers SMEs face in accessing funding

Limited credit history, lack of collateral, and sitting outside FI’s credit box prevent SMEs from tapping much-needed financial resources.

Addressing these barriers requires financial institutions to embrace digital transformation: “To overcome these barriers, FIs have to innovate and embrace digital transformation, and that helps to provide faster, more efficient lending decisions, ultimately getting to that time to say yes much quicker to provide for the SMEs funding needs.”

The threat of new entrants

Traditional financial institutions face significant challenges competing with more agile non-bank lenders in the SME space. Lista points out that NBFIs have gained market share by utilizing technology to streamline the lending process.

“Traditional FIs really are struggling to compete with non-bank financial institutions in the SME lending space, because NBFIs leverage technology to offer faster, more flexible lending solutions,” Lista noted. She added that financial institutions also face a tighter cost basis in the SME lending space, which means they need to find ways to reduce costs without compromising service quality, which is very tough to do.

But beyond technology and cost challenges, FIs may be overlooking a chance to diversify their offerings: “There’s really an opportunity and a need for FIs to grow into different asset classes as well, such as moving from SME lending into commercial and syndicated lending, which can help FIs diversify their risk and open new revenue streams.”

How technology can help FIs build better CX

CX and bank office efficiency play a critical role in the relationship SMEs have with their FIs –  technology in validating financial statements, automating underwriting, and enhancing loan servicing capabilities all contribute to the quality of service SMEs receive from their FIs. 

Lista also highlighted a frequently overlooked area, loan servicing. It’s here that FIs have a particularly important investment to make: “Loan servicing is kind of an afterthought in digital transformation, but it really can’t be an afterthought anymore. Loan servicing capabilities have to have that digital transformation, as well, and provide for better communication and transparency for bank clients.”

The importance of modernizing back-office operations

One reason why FIs have struggled to keep up with the pace of their competitors is their historical underinvestment in back-office operations.

“FIs really haven’t focused on the back-office servicing operations of the SME lending process,” Lista said. “There’s a lot of reasons for that. Primarily, they’ve been underfunded because banks thought that the back-office was not revenue generating, and so they would really focus their resources and their budget going towards client-facing systems.”

Blueprint for customer retention and growth

At a time when FIs have to play catch up with their NBFI counterparts, the key to their success may lie in focusing on customer needs, expectations, and experience. 

On the revenue front, Lista noted that while SME loans are typically lower in value, they’re higher in volume, creating a unique opportunity: “In order to grow their revenue, they have to prioritize the customer experience to get that retention and loyalty, and if they do that, they may see an approximate two and a half times increase in their revenue growth compared to those who do not prioritize customer centricity.”

To-do list: What FIs can do to better serve their SME customers

FIs that want to improve their SME customer experience can take the following steps:

– Gain a full system view: Undertake a detailed overview of the systems involved from loan initiation through servicing and termination. “Key systems that you should look at are customer portals, your borrower portals. You should be looking across the ecosystem at KYC, AML systems, loan origination systems, loan documentation systems, and, of course, your loan servicing system.”

– Evaluate potential for automation: Integration between these systems that contribute to the loan servicing process is crucial: “It’s important to look at what they’re doing, what the purpose is, but also, how do you automate and integrate these systems together and piece them together in an automated way throughout the ecosystem?”

– Don’t overlook KYC and AML – they impact CX: KYC and AML processes have a significant impact on turnaround time and customer experience. When considering modernization strategies, this is a critical area of evaluation with a CX lens. Complex KYC and AML processes can impact onboarding success and impact client retention. “If they don’t have smooth transitions from a portal to a KYC system, for example, the borrower is going to feel the delay and the impact,” said Lista. 

– Find technology partners that can technological lift in automation: Tools like Finastra’s offerings can help FIs create a consolidated and streamlined lending ecosystem.

“Finastra offers a lot of solutions to enable this consolidated and streamlined end-to-end ecosystem,” Lista said. “Our products help FIs to digitally transform their ecosystems to provide those right SME offerings and mirror up with their demands and expectations.”

– Integrate teams into SME strategy: “FIs must not only look at the technology, but they must also integrate their people and processes within the technology as well,” Lista advised. “Enhancing collaboration across business segments—the front office, the back office, the middle office—and also working with their IT departments and their technology departments internally. That collaboration is key.”

To read more about what FIs can do to build competitive lending experiences and products for SMEs and find partners that can accelerate modernization efforts and positively impact bottom lines, please visit Finastra’s website.

Posted on

Why e-commerce brands have the processes but lack the resources to execute personalization programs

Focusing on personalization can drive real topline growth and customer loyalty for e-commerce brands. In a world chock-full of brands competing for customers’ attention and spend, AI-driven personalization is helping mature brands capture more market share. For e-commerce firms, the right personalization strategy can make all the difference, turning browsers into buyers by delivering experiences that truly resonate.

Over the next five years, businesses that excel at creating tailored experiences and communications stand to capture $2 trillion in revenue, according to BCG research.

But despite the global opportunity personalization presents for retail and e-commerce brands, firms are struggling to build cohesive, organization-wide strategies – significantly hampering future revenue growth that can come from personalization.

E-commerce brands can use a 4 dimensional framework to assess their personalization maturity and here’s how Dynamic Yield by Mastercard found the industry shakes out:

  • Culture: While 44% of brands report that they are on track in terms of establishing a personalization program, more than half have yet to build a holistic strategy encompassing dedicated resources/support and accurate ROI measurement.
  • Resources: Only 50% of organizations have established a personalization team with business, technical and creative expertise, while the other half relies on ad-hoc support or goes without it completely.
  • Processes: Only 36% of businesses report using insights to run additional tests to understand their audience more deeply and optimize their personalization programs.
  • Effectiveness: Although 48% of e-commerce firms report that an audience strategy is critical to a firm’s ideation and planning for personalization projects, only 36% report that it’s occasionally applied to hypothesis setting and testing analytics. 

Culture is the soul of personalization programs 

Personalized experiences are quickly becoming a mainstay of e-commerce brands’ targeting strategy, with 67% of brands currently planning to invest further in their personalization programs.

However, gaps remain for brands looking to align their entire organization around personalization: 40% of brands still rely on measuring only conceptual KPIs rather than fixed, quantitative ones like increased revenue and improved add-to-cart rate that can be linked back to generated value.

Even more damaging: 30% of e-commerce firms report making spur-of-the-moment decisions without establishing a plan based on data and clear KPIs. 


“Establishing clear ownership, mandates and accountability, from the C-suite down to analysts, is key to creating a culture of personalization across an organization. Once those are in place, cross-department collaboration and alignment naturally follows, driving quantitative impact and value to the wider brand mission,” said Ben Malki, Vice President of Customer Success, Americas, Dynamic Yield by Mastercard. 

Resource allocation: the lifeblood of personalization

Although most firms recognize the value personalization can bring to both their business and customers, there is a disconnect between understanding the importance of personalization and making more resources available to these programs and teams. Firms should take a cross-functional approach to personalization, bringing on board business, technical and creative expertise to ensure the success of their programs. 

Currently, 38% of firms have a singular team that works with other departments to implement web-based personalization, while 28% of the firms have multiple teams that operate without a holistic approach.

Processes are the backbone for ROI

Processes help teams establish a clear workflow and pipeline for actions and analysis, and brands that ignore this step are often unable to accurately maximize the impact of their personalization efforts. 

26% of firms are currently failing to share detailed insights from their ongoing personalization campaigns with the wider organization, and 25% rarely share insights if at all. The result of this communication failure is an inability to showcase personalization’s value to the business. This can significantly impede executive prioritization and greater cross-functional teamwork to improve future campaigns.

Mind over matter: Measuring effectiveness

Data from Mastercard suggests that e-commerce brands that were late to adopt personalization have failed to keep up with growing customer expectations and are falling even further behind those that have gone on to develop sophisticated strategies. 

67% of brands have yet to build internal alignment around their audience strategies, and only 41% have identified different data sources, like CRM and offline data, that can help inform personalization programs, but have not yet put these data sources into action. 

“Looking around corners and adapting to online customers’ expectations on-the-fly calls for a well-crafted personalization roadmap. By crafting a single source of truth for customer data and fostering strategic alignment, brands can clinch a multi-trillion-dollar opportunity in personalization,” said Donovan Yong, Principal, Advisors Business Development, Dynamic Yield by Mastercard. 

If you want to read a more detailed analysis of how personalization efforts are performing across the four signals mentioned above and explore the global landscape of e-commerce programs across regions like America, APAC and EMEA, download this report from Dynamic Yield by Mastercard. 

Posted on

How Citizens Bank is building GenAI with a five-year vision, not just quick fixes

Investment in data is the hallmark of successful Gen AI implementations, according to Citizens’ Chief Data and Analytics Officer, Krish Swamy. 

Giving us a system wide view of how Citizens is leveraging Gen AI, Swamy joins the podcast to talk about harnessing the power of data to drive decision-making, enhance customer experiences, and navigate the complexities of digital transformation in the banking sector. 

Our conversation delves into the challenges and opportunities of building a data-driven culture within a traditional banking environment, and how Citizens is positioning itself at the forefront of financial innovation through strategic analytics initiatives.

Swamy, who also heads the firm’s Generative AI Council, shares his vision for the future of data in banking and the tangible ways Citizens is turning data insights into meaningful actions that benefit both the institution and its customers.

Watch the full episode

Listen to the full episode

Subscribe: Apple Podcasts | SoundCloud | Spotify

Long term view of Gen AI implementations

Citizens’ approach to Gen AI is best described as cautious and optimistic. While the firm is not rushing into any use case and is instead taking a methodical approach to evaluating every time a process or task can be improved by Gen AI, it is also sketching out what role the technology could play in the future for its employees and customer experience. 

“We’re not just taking a process, or a component within a process, and applying Generative AI there. While that might be the starting point, the end game is always going to be: How does this function three or five years from now? How do we work towards that end game?” said Swamy. 

Strong data based foundation as a differentiator

Swamy is a firm believer in using a comprehensive data infrastructure as the scaffolding for new technological implementations. “When we invest in data, when we make data easily available, and when we teach people how to use data, I think they become a lot more effective at being able to self-serve. So creating that foundation is an area of differentiation,” he shared. 

One area where this focus helps the bank drive powerful results is fraud, which has seen a significant uptick since the pandemic, according to Swamy. “We’ve spent a lot of time overhauling the fraud infrastructure and the fraud platform itself. There are multiple sub components around fraud detection, claims processing, case management, which all are parts of the overall fraud value chain. And we made investments to improve the quality of those platforms,” he said. 

Helping the fraud team stay ahead of bad actors, is the firm’s move to the cloud, which should be completed by the end of this year. “We are almost 80% migrated to AWS, and it makes it easier to get access to data and we are able to bring better data when it comes to our fraud defenses,” he said. 

Having a centralized source for the data also ensures that fraud teams that include analysts and contact center employees are working from the same source of information. This allows these teams to be more effective and coordinated when trying to spot trends and undertake fraud mitigation strategies, he shared. 

Another area where the firm is applying data-led Gen AI strategies is the call center. “A lot of the customers’ questions tend to be fairly narrow, almost esoteric and [call enter employees] have to reference procedure documents to be able to give that answer,”  he said.

In the past, call center employees have used keyword search to access this information, but now the firm is using Gen AI and helping call center agents learn how to prompt more effectively to reach information,” he said. 

Similarly, the firm is also using the tech to help its developers take care of some of the most frustrating parts of coding: documentation and testing. “Those are areas where we’ve been able to find a lot of leverage from giving software development engineers the right tools to be able to do the testing, documentation, sometimes even writing code, and become more efficient at that,” he shared. 

Citizens’ partnership strategy 

When it comes to assembling the right technology partners, Swamy believes building consistency across the organization is the golden rule. “For instance, there are multiple teams that need the ability to have machine learning platforms, and it is conceivable that everybody then goes out and figures out their own thing. That would be a really bad outcome, because I think that would lead to proliferation of costs and would lead to loss of control,” he said.

“What you do need to do is make sure these solutions are all integrated with all of the other solutions, which is a lot of work for sure. The place where we have spent a lot of time on homegrown solutions is on managing our data. Those are critical assets which are unique to us, which we would not be comfortable leaving completely in the hands of a commercial solution or a bought out solution,” he said.

Posted on

The story of Erica, Bank of America’s homegrown digital assistant

Banking digital assistants may be common now, but in 2017, Bank of America was one of the first to be thinking about how they make the firm’s customer experience more powerful. The answer was an in-house build of a digital assistant that required the firm to hire PhDs in linguistics and build a collaboration structure that could facilitate teams from different departments. 

In 2024, BofA clients interacted with Erica 676 million times bringing its total interactions since its launch in 2018 to 2.5 billion. 

On the show today, Hari Gopalkrishnan, who leads Bank of America’s Consumer, Business & Wealth Management Technology team, joins us to tell the tale of  how the firm built its industry-leading digital assistant, Erica.

Hari shares how the firm has gradually expanded Erica’s remit beyond consumer banking to also include multiple lines of business and individual and corporate clients across the firm’s global footprint.

It’s a dive into what it takes to push the boundaries in this industry, how the firm thought about development, testing, expansion, and how Erica’s capabilities can be expanded with the recent innovations of Gen AI. 

Listen to full episode

Subscribe: Apple Podcasts | SoundCloud | Spotify

Source: Bank of America

CX as the impetus behind Bank of America’s investment into building Erica

Back in 2017, Hari Gopalkrishnan’s team realized that despite the expanding remit of the BofA app, customers were still flocking to traditional channels like banks and call centers before they tried self-serving. A dissection of this behavior served as the catalyst for the Erica build. 

“We were putting tons and tons of features into our app, a five inch screen, and our customers were still walking into the branch, calling into the call center … It’s hunting and pecking, and it was hard to navigate once you put more than 10 or 15 key features. And so our first insight was, you want to be able to have your customers interact with a platform in a way they choose, not the way you choose,” said Hari.

Why Bank of America decided on a DIY approach to build Erica

Although the industry is pretty well-versed in building, deploying, and improving chatbots now, the landscape was quite different in 2017, when BofA had first thought of developing Erica. There was no blueprint for what a banking digital assistant looked like and technology providers were few and far between. The slim pickings in the marketplace, as well as the limited applicability most software had in the financial services, fueled an internal build. 

“There were a bunch of startups that were coming out, we spent a few weeks with them and either they disappeared or were going to get gobbled up. So it was a very volatile marketplace for acquiring software that did this. People were trying to play in this game, but weren’t quite getting it. The ones that understood some level of NLU (Natural Language Understanding) didn’t understand financial ontology,” he said.

“We did an assessment, and we looked at third parties, like we always do, but at the end of the day, it just wasn’t going to work out. At the same time, we found a couple of really good open source NLP engines. They were strong, solid, and very well regarded in the industry. We actually hired up a team. We always had good technical engineers. But we also actually hired people with PhDs in linguistics to work on this. Then we started to work with our teams to figure out: what is the digital experience going to look like?”.

How Bank of America structured teams across the org to build Erica

Building Erica required BofA to think across organizational silos, and really invest into creating a collaboration framework that would allow Erica to improve CX without compromising on risk tolerance. 

“We actually took the Agile construct to the next level. We had teams set up in [different] regions which were actually in the room. It was the engineering team, the UX team, the appropriate legal team, all opining day in, day out, on all aspects of the platform. The sprints were not just engineers running off and UX coming in weeks later. It had UX teams embedded. In fact, when I used to visit the teams, it was sometimes hard to tell who was in the design team and who was in the engineering team. That was actually the power of how this came together,” Hari shared. 

The fork in the road: User testing proved presumptions wrong

Designers and developers have a conceptual map of the software they are building and they also spend time trying to understand their users’ behaviors and pain points. But there is no silver bullet for getting everything right, right off the bat. Hari shares how testing motivated a major pivot in Erica before it launched. 

 “We thought we would have to invest just as much in voice as in text -– that people would half the time talk into the app, and half the time they would type into the app. Then we go to a small group of customers, and we get more feedback from this. And the feedback we got there was people were just typing in text 90% of the time. They were rarely using voice.”

Gen AI potentialities for Erica

Bank of America is not sleeping on Gen AI – it’s just chosen to stay quieter than most. “As we look at the emergence of Generative AI, we actually see that classification can actually get a lot better. You can actually talk even more naturally in a natural language. So that is just a natural sort of expansion of where we go with Erica. We have about 25 different proof of concepts right now, many of them are actually about to get into production, which use a Large Language Model in some way, shape or form, to continue to enhance the work that we’ve been doing,”. 

—

The following excerpts were edited for clarity

BofA’s blueprint for Erica’s expansion into multiple lines of business

Some of it actually is using fit for purpose language models that are pre-trained on certain things. So, for example, for employees there are available technologies that actually are trained on things like integration into your HR system and integrating into your help desk system. We don’t want to go build a whole bunch of things that actually have been built by somebody before. 

The reason we built what we built is because nobody was building that before. When it came to employees, we realized that we can leverage all the goodness we have on NLU (Natural Language Understanding) and User Experience. We also found that there are available models that actually do a really good job of NLU to service intent and to calling of existing HR platforms. 

We had to figure out case by case, do we start a native build? Do we integrate with existing models when it comes to Erica for business banking there? We had to go to a different set of data sources. You had to make sure those sources were clean. You had to make sure, in some cases, that there was an API available to make that interaction happen … in some cases, many of the services in the past may have been built for a specific User Experience or a specific application, we had to make sure that they get rebuilt or reimagined to be invoked by a chat bot, because sometimes you may need clarifying steps. You may have a multi-step process before you actually call an interaction. That also helped us become better in our core platforms, because that helps us now be ready for the future. 

How BofA is balancing ROI, risk, and innovation when it comes to Gen AI

We have an AI Council. Even though we’re obviously a very large company, we try to work in a very integrated fashion and look to learn from everything that’s going on across the company. There are lots of parts of this company, and we come together. You could say it slows us down, but we’re okay with that. We ask, what are the pilots and POCs you want to run? And why do we want to run those POCs and pilots? The people involved in that council involve senior business leaders, senior strategy leaders, senior risk leaders. We’re asking, does this thing align with our risk framework? We have a risk framework that has 16 points of risk. You can imagine bias, intellectual property, transparency, and explainability, in there.  

Is the work you’re going to do, going to abide by those risk frameworks? Is there an adequate human in the loop so that you can make sure that the thing doesn’t run away? How do you measure the performance? What guardrails are you going to implement? Those are the things we look at as we implement any of these proof of concepts and eventually take them to commercial use. 

The second part is we also look at, what is the mindset you have on the ROI generation of doing this work, because none of this stuff is cheap. This is something everybody is wrestling with. There’s so much hype out there that people are throwing out, I’m going to spend a billion dollars. I’m going to spend $5 billion and when you ask the question, tell me what your bottom line is going to be, what are you going to get in return? The answers are a little bit more diffused. 

So we’re taking an approach of saying, we want to understand how work gets done. We want to understand activities, jobs, tasks. We want to understand what part of those tasks cost, what money. And then, when you implement solutions like this, what’s the ROI? 

Posted on

Sell your investment strategies (without the cost and burden of creating a fund)

screener

I frequently meet people with really compelling investment strategies and ideas.

Hey, can you help me raise some money?

They’re looking for help putting together a fund to demonstrate exactly how good their strategy or stock picking really is.

Starting a fund is hard…and expensive

It’s not that I can’t really help them — it’s that starting a hedge fund or mutual fund is pretty complicated and expensive. You need to see significant growth in assets to be able to scale these things to profitability (once they achieve that, though, they’re pretty damn profitable).

Like any startup, the chances of these startup funds achieving escape velocity — getting enough traction to turn their good ideas into profitable ones  — is pretty slim.

But, there are other ways of putting your investing talent to work and make money while doing it — all without the headache and onerous infrastructure needed to manage a fund or a regulated investment advisory.

How to make money from your investment ideas (without starting a fund or having $$)

Here are 5 ways to get started selling your portfolio strategies:

Continue reading “Sell your investment strategies (without the cost and burden of creating a fund)”

Posted on

Winning strategies: Investing using policy as a guide – with James Juliano

policy based investing

A lot of investors claim they’re focused on the big picture.

Others give lip service to their focus on “the long term”.

James Juliano, partner and portfolio manager at Kairos Capital Advisors, walks the walk. He and founder of the firm, Russell Redenbaugh, use economic and government policy as a guide for their investment decisions.

James joins us on Tradestreaming Radio to talk about how he deciphers the policy tea leaves and how those big ideas get implemented tactically int their portfolio.

Listen to the FULL episode


Continue reading “Winning strategies: Investing using policy as a guide – with James Juliano”

Posted on

Using Google to forecast a stock’s reaction to earnings reports – with Darren Roulstone

Smart investors are looking at various data sets to help give them an edge with their investing. Some of this information is financial in nature — much of it isn’t.

Professor Darren Roulstone has studied how investors are using Google to search out financial information and what search volume may say about future stock prices.

Continue reading “Using Google to forecast a stock’s reaction to earnings reports – with Darren Roulstone”

Posted on

Using Google to forecast an earnings pop (or plunk)

Google’s my friend.

Not only do I rely upon it for email, video, and of course, search, but I’m using it to invest  better and smarter (the Tradestreaming way, right?).

Let me explain:

One of my first podcasts on Tradestreaming Radio was with finance professor, Joey Engelberg. In How to use Google search data to invest, I asked Engelberg about a paper he had recently published that showed how useful Google could be in forecasting stock prices.

Using Google Search Data to Invest by tradestreaming

Specifically, Engelberg noticed:

  1. Google search volume likely measures the attention of retail investors
  2. and does so in a more timely fashion that existing proxies of investor attention

And of course, stock prices tend to follow attention.

So, an increase in Google search frequency (SVI) predicts higher prices in the next two weeks and also contributes to a large first-day return (and long-run underperformance) of IPO stocks.

Awesome stuff and after we spoke, Joey kind of went underground (he did leave UNC and headed for UCSD), using his research to make coin at a hedge fund. I spent a whole chapter in Tradestreaming (my book) describing co-lateral research — stuff that’s inherently non-financial in nature (Google search, Amazon ratings, etc) to help us make better investing choices.

Now a new paper shines light on how Google search reflects investor information demand and what that means for earnings news.

Continue reading “Using Google to forecast an earnings pop (or plunk)”