A Product Manager’s Journey in AI /ML Space
My Journey at Ushur — A Conversational AI & Process Automation Platform
It has been an exciting ten-month journey as a product manager at Ushur, a Series A tech startup based in the Bay Area. We are a growing team of about 60 employees with two office locations — Santa Clara, California and Bangalore, India. I am privileged to work in the hottest AI technologies space along with some of the smartest minds. Ushur is led by an amazing leadership team — Simha Sadasiva(CEO) and Henry Peter (CTO), who have built a powerful Conversational AI and Process Automation platform that Enterprises can leverage to build custom workflows. Non-technical folks can build solutions using our no code builder to automate repetitive tasks and drastically improve their end-customer experience.
My role is very cross-functional, and I work closely with high performing and talented team members from Customer Success, Engineering, Data Science, Design, Solutions Engineering, Sales and Marketing teams to continually evolve our platform.
So what is Conversational AI? What problems does Ushur solve?
When I onboarded at Ushur, I spent quite some time trying to understand how our Enterprise customers are using our solutions to solve their business problems. I was amazed by how powerful the platform is, as Ushur can automate engagements for enterprises across many channels to improve operational efficiency & CSAT (customer satisfaction), while reducing costs. In fact, we use ushur as a verb to indicate the dynamism and real-time nature of the solution.
Conversational AI is simply a technology that has been empowered with language-processing skills to partake in urgent conversations that customers don’t want to be on-hold for — think smart chatbots and virtual assistants. With Conversational AI, we are able to process the end user’s utterances and identify their intent so we can deliver self-service over digital channels without having to escalate to a live agent.
Let’s say you are an enterprise and most of your customers are still reaching you or you are reaching out to them through traditional channels like phone or email for typical support use-cases such as filing a claim, making account updates, looking up a balance or even scheduling appointments. Customers often experience long wait-times or delays to hear back. On the other hand the call center agents are overwhelmed by processing repetitive requests from customers.
According to McKinsey, 75% of your online customers expect help within five minutes, and when demand surges, support centers can’t magically expand capacity.
Ushur delivers a digital self-service portal that can be deployed across any channel — SMS, Email, Social or Browser, powered by our proprietary conversational AI technology for customers to engage and fulfill their requests at their convenience, 24x7, from their digital device of choice. Customers don’t have to keep waiting on the phone for an agent or stay on-hold. Thereby, we improve CX and CSAT right off the bat. Enterprises can dramatically lower operational costs by reducing calls to expensive contact centers and re-assign their agents to handle more complex tasks.
With our no code builder, a business user can build and launch a fully tailored digital self-serve portal in a couple of hours. The Ushur platform integrates with any home-grown or third-party CRM to sync and update data automatically, reducing admin hassle and improving data integrity and audit compliance.
A interesting feature I have built and launched recently working with our talented team
Agent Co-Pilot — Feature Overview
I know many of you may be interested in the day-to-day of product management and development, so I’ll share some insight into the process at an AI startup.
Our Enterprise customers were seeking a solution to support customer support reps, adjusters or specialists to review or update customer data at any point during a completely automated journey.
My first step was researching and white-boarding the problem statements and use-cases for our enterprise customers and their end-users with Henry Peter and Michael Fisher (VP Product & Marketing). Once we arrived at a better understanding of personas, pain-points and related use-cases, I started drafting product requirements with wireframes to define the feature. The next step was to colloborate with our amazing engineering team, to define technical requirements to architect a solution and build this feature.
Agent Co-Pilot feature is a powerful capability as it provides an interface for Enterprise Agents to get looped in to review and modify data at any point during the workflow automation.
Here is an example of a couple of use-cases our Agent Co-pilot feature can support seamlessly:
Use-case 1: If a customer has sent an email to an enterprise requesting a price quote, we can process customer intent and automatically extract key data points from that email. By syncing with the enterprise backend systems, we can pull together a quote seamlessly (in mere seconds, which normally takes hours or even days). Just prior to sending the automated quote to the customer, we can have an agent to review the quote our platform has created.
Use-case 2: If a customer is filing a first notice of loss claim (FNOL) or submitting documentation for loan or mortgage application processing, agents can not only review or update submitted information but also request for missing information using the Agent Co-pilot portal. Thus, the platform reduces a lot of the friction and back & forth phone calls
Thanks for joining me today — it’s an exciting time to be a PM at a cutting-edge AI startup. I hope to cover some of the other features we have been working on and the fun stuff we do at Ushur in my future posts. Stay tuned!
