It’s now been almost 2 months since I joined Posh Technologies. It’s been fun so far and I expect that will continue. I joined Posh for many reasons including the people, the culture, and their product-focus. But there’s another reason: AI is revolutionizing our world and I wanted to move from the AI tools business to the front lines to build applications that more directly improve the lives of end users.
I’ve been in the AI tools business for over a decade. MathWorks was long the quintessential tool provider for AI researchers and practitioners. I was at MathWorks when Google’s TensorFlow first disrupted that dominance. While we missed the initial wave, the Deep Learning Toolbox from MathWorks today helps countless engineers train models and deploy code to a wide range of popular products, some of which you likely use. Later, I personally initiated the Text Analytics Toolbox which made NLP tools like word2vec and glove, predecessors of Sentence Transformers, more accessible to engineers. Last year at this time, I led product at Pinecone, a startup delivering a vector database to power next generation AI applications including chatbots, semantic search, and extreme classification (see for example this workshop Nils Reimers and I did on how to combine a large language model with a vector database for search). I similarly worked closely with both OpenAI and Cohere and scores of our shared customers to help them solve problems with our respective tools.
I take great pride in the AI tools I helped deliver to engineers and the many exciting new products those engineers brought to market. Yet, this past year, seeing first hand the exciting new things Pinecone customers were building, I felt the desire to jump in myself, leave the tools side of the business, and join a team building these new AI-powered products.
At Posh, we're using conversational AI to revolutionize the banking experience for both our clients and their customers. It's not easy work - we must balance accuracy, brand uniqueness, and user experience. ChatGPT is no panacea to these challenges. But there is no shortage of clever and creative ways to leverage the AI advances behind it. For example, see this workshop we at Pinecone did together with OpenAI. We use OpenAI-trained embeddings to search over specific content, apply business logic, and then merge results using GPT. Creative approaches like this are needed to move from generative AI as a novelty trick to something that creates honest value for users.
Our goals at Posh are to let humans be human and to empower our Financial Institutions (FIs) to serve their customers better. This means we automate the routine and mundane so that humans can focus on helping other humans when only they know how. This means helping FI Experience Managers and Brand Managers focus their efforts on where to add their special touch to IVRs and chatbots. It means delivering the right insights to FI leadership based on their customers’ actions without all the clutter so they can serve their customers better. And it means educating the right banking customer at the right moment about a line of credit to help them avoid accruing the wrong type of debt. That’s the kind of value I joined Posh to create.
Dave Bergstein is Director of Product at Posh. He’s been in the technology space for over 20 years and spent the last 10 in AI-related Product Management roles. He holds a PhD in Electrical Engineering from Boston University. In his spare time enjoys being with his family, crossfit, and walking his dog Zeus.
(note: chatGPT helped me write this post and saved me a bit of time, which was some honest value right there)