Customer Service in a Digital Age
The Covid-19 pandemic brought with it a sweeping shift to digital that permeated nearly every aspect of daily life. As we adjusted to Zoom work meetings and attempted Tik Tok dance trends, many of us also turned to online and mobile channels to conduct banking transactions.
Banks were forced to quickly stand up new digital offerings and expand existing ones. In some ways this was a good thing; banks had been trying for years to get consumers to adopt digital channels, and Covid was a massive accelerant of that trend.
But for many credit unions and community banks – institutions that have historically relied on a personalized, customer-centric style of banking – the shift to digital presented new challenges. In an increasingly digital-first world, maintaining connectivity with customers – and delivering customers a first-rate experience – suddenly became much harder.
AI-powered digital customer service tools provided a natural way for banks to engage with customers remotely. As digital banking neophytes flooded call centers and helplines with questions about balances, wire transfers and online check deposits, banks increasingly looked to intelligent text and voice-based platforms to serve as their digital front desk.
The world is beginning to return to normal, but there is no doubt that the pandemic will leave its mark. Going forward, banks will need to develop strategies that help them deliver exceptional customer experiences in a world that will be partly in-person and partly online. We believe that conversational AI platforms are uniquely poised to play a central role in the banking experience of the future.
Not Just Any Platform
The appeal of conversational AI “chatbots” for banking has always been clear; replacing costly customer service agents with software that can answer customers’ questions 24/7 – who wouldn’t want that?
Unfortunately, most chatbots have been less-than-ideal solutions for banks. Until recently, the technologies that power conversational AI were relatively nascent. Legacy banking chatbots created in the early 2010’s relied on basic triaging and flowcharts; they were rigid and
felt robotic and were more likely to frustrate a customer than track down the right answer.
Many early banking chatbots were not integrated into the broader banking infrastructure. This meant that not only could they not answer complicated questions, but they also couldn’t be passed off to other channels – if they couldn’t answer a question, they were basically a dead end.
Another shortcoming came from applying generalized AI systems to banking use cases. Conversational AI needs to be trained with copious amounts of relevant data before it can begin to reliably answer domain-specific questions. Although generalized solutions may be cheap and quick to deploy, they struggle mightily to understand nuanced banking-specific requests.
In recent years, some banks have been able to build great virtual customer service platforms – Bank of America’s “Erica” is a compelling example of this. But it takes resources to build and deploy a solution like Erica – resources that are out of reach for most credit unions and community banks.
Although conversational AI can be a powerful tool, we believe the best credit unions and community banks will need a very specific type of conversational AI platform. Specifically, credit unions and community banks will need a platform based on the latest natural language processing research that can understand and respond to customers in the same way a human would.
The platform would need to be trained on a robust set of banking-specific data and built on advanced machine learning that learns rapidly, empowering bank employees to easily augment its responses and models to accelerate its learning.
The platform would need to be integrated into core banking systems so that it could perform real banking commands such as surfacing an account balance or paying down a credit card.
The platform would need to be omnichannel and flow seamlessly between online, mobile and voice. The platform would also need to be personalized, greeting customers by name, prepared with relevant details, and, if the bank chose, ready to surface relevant services or products.
Crucially, the platform would need to be priced affordably and ideally run by a team that recognizes the importance of this technology for the future of credit unions and community banking.
Such a platform could be a force multiplier for credit unions and community banks, helping them to seamlessly bridge online, mobile and in-person banking services in a way that is personalized, scalable, always-on and ever-improving.
That platform, would, we think, make a profound difference for financial institutions and their customers.
We believe we’ve found that platform.
Powering More Natural Conversations
Posh’s cofounders – CEO Karan Kashyap and CTO Matt McEachern – met as freshman roommates at MIT, where they quickly discovered a shared passion for artificial intelligence. As undergrads and then graduate students, the pair launched a consulting company, building custom chatbot solutions for a wide range of businesses, often recruiting classmates to help them on projects. By the time they were ready to launch Posh, they had quite a few things going for them, not least of which were a deep understanding of the science, direct access to the brightest minds in the space (many of whom they’ve hired) and the benefit of significant recent technological advancements on which to build.
Today, Posh’s conversational AI platform helps credit unions and community banks automate contact center and help desk FAQs and workflows in a way that feels
natural. With true omnichannel coverage, Posh’s text-based bots and interactive voice response (IVR) provide seamless support across Web, Mobile, Digital Assistants and phone.
Unlike most one-size-fits-all platforms, Posh’s AI has been trained on banking-specific data, so it understands the nuances of customers’ needs. Posh offers easy-to-use content management tooling which helps banks and credit unions accelerate the AI’s learning and customize responses and workflows. Posh’s patent-pending IP for managing contextual memory helps Posh outperform 3
rd parties by 15%+ accuracy and deliver a natural, conversational experience.
In just a few short years in-market, Posh has built a customer base of more than fifty credit unions and community banks, even as they continue to iterate on product and improve their models. Importantly, we think this team is just getting started; always-on and ever-improving, remember?
That’s why we’re thrilled to be leading Posh’s Series A round, alongside friends and fellow investors at Human Capital, CMFG, Curql, JAM Fintop and Piedmont.
Here at Canapi, we believe the future of banking will in part be powered by conversational AI. While we recognize the challenges that our bank partners face in navigating the new digital normal, we’re encouraged by the vision of a future where banks leverage digital channels to get closer to their customers. We’re eager for the promise of banking as a dynamic, personal,
conversational experience powered by software that gets better every time we use it.
We’re incredibly excited to be joining Karan, Matt and the rest of the Posh team on this next step of the journey. We cannot wait to see what this talented team is able to achieve and are very much looking forward to continuing the conversation.