Prior to this engagement, financial advisors executed 100% of their FundVisualizer searches by manually searching and sifting through results to find relevant funds for their clients. This approach takes time and introduces human error.
With a voice-enabled Intelligent Assistant, search times can be significantly reduced, and financial advisors are presented with a hands-free alternative to text-based searches. Further, when augmented by machine learning and AI, this search approach can yield more personalized results for Putnam’s clients. The potential impact: less time spent researching and more time for financial advisors to interact with their clients in meaningful ways.
Because of the importance of this initiative, Putnam was looking to move swiftly. They engaged Cantina to help design, build, and ship a beta version of the Intelligent Assistant in less than 5 months. To do this, Cantina quickly ramped up a team and got to work with Putnam.
We started by segmenting our work process into two distinct phases: phase I (technical design) and phase II (build). Along the way, we introduced Cantina experts in voice user interfaces, experience strategy, and machine learning to help shape the voice experience and accelerate delivery for Putnam.
Phase I began with a technical assessment and an evaluation of conversational flows. Cantina used insights derived from this assessment to build a conceptual prototype that informed the build phase of the engagement. Throughout this phase we collaborated closely with Putnam to review progress, collect feedback, and ensure that our technical plan and prototype were feasible and addressed Putnam’s core needs.
After completing phase I, we transitioned into the build phase. In this phase, Cantina and Putnam worked together to engineer an Intelligent Assistant that operated well on four critical dimensions:
-
Speech recognition: successfully converting speech into distinct words that the Intelligent Assistant could understand
-
Natural-language understanding: successfully extracting intended meanings so that the Intelligent Assistant could take appropriate action
-
Text-to-speech synthesis: successfully converting text into understandable machine-spoken words
-
Visual cues: successfully embedding visual cues into the Intelligent Assistant’s user experience to accompany and reinforce the system’s responses
Choosing the right tools and application programming interfaces (APIs) was critical for this phase. Each tool and API needed to work well within Putnam’s technical environment and address the core customer needs met with the Intelligent Assistant. For example, we implemented the Google Speech-to-Text API because it performs well with Google Chrome and other web browsers and provides greater accessibility and a better experience for financial advisors, while integrating well with the existing FundVisualizer application. Each of these decisions, among many others, was critical for building an effective and user-friendly solution for Putnam. In the end, we delivered what Putnam needed ⎼ a fully functional beta version of the Intelligent Assistant, and a critical first step toward implementing deeper machine learning and AI capabilities into FundVisualizer and the financial advisor experience.
As client expectations continue to grow, the tools to help financial advisors analyze investment ideas continue to leverage the power of new technology. FundVisualizer is leading the way with beta implementations of cutting-edge technology like natural-language interaction and machine learning.
