Back to resources

Building a Next-Gen Product


16 August, 2021

Richard Li
Richard Li

Senior Software Engineer Manager at ASSURANCE IQ

Richard Li, Senior Software Engineer Manager at ASSURANCE IQ, describes how automating much of the processes in the industry that is still largely manual led him to build a next-gen product.


My company operates in the insurance industry, one of the industries that value stability and old ways. A number of processes critical for the industry are still manual, done the same way they were done 50 years ago. People still use fax machines and send their documents by mail. Not surprisingly, our own system has many loose ends, which cost us additional operating hours and workforce.

For example, to get an insurance license that will allow agents to sell a product, agents will have to submit a number of documents on paper and then wait for those documents to be verified. In the meantime, they won’t be able to do anything because they are not certified. The problem was rather obvious as well as the solution: we needed to automate the existing processes as much as possible, thus making the life of insurance agents easier.

Actions taken

We are a data-driven company, so the most natural course of action for us is to start by collecting and analyzing the data. So, we began collecting data on what agents’ needs and pain points were. We ran a couple of surveys but also held workshops to discuss with them their most urgent needs. We also invited agents to a design session to help them grasp the product in its entirety. We collected input from more than 100 agents, which we considered a sample large enough to create the basis for the idea pool. Then we tried to group common themes and mix them together to build a new product.

We also tried to collect and process information from carriers, which provided products for agents to sell. We wanted to learn more about the process that made agents licensed to sell those products. Then we combined all those available products together and created a simplified solution of the whole process. Once we had all the info, we started to build the product, iteration by iteration, following the best agile practices. We would build the most rudimentary version, launch it to agents immediately, and collect their feedback. We used their feedback to generate new ideas and continuously reiterate. The whole process didn’t last too long, not longer than four months. We went through three rounds of iterations plus finalizing the whole setup.

The product we created was nothing like what existed before. For many agents, it had a life-changing impact and helped them improve their operations significantly. It was a great leap forward into a more automated space that challenged the old ways that were slowing down our industry. The numbers were also stunning: we achieved over 50 percent increase in efficiency and over 40 percent in cost reduction. Not only were we the pioneers building a next-gen product, but we also built a hugely successful product.

Lessons learned

  • In retrospect, it doesn’t seem incredibly difficult to do what we had done. But the question remains: why didn’t our competitors or other companies give it a shot before we did. I believe two things are critical:
    a. No one wants to spend the time to innovate when people are adapted to a certain way of working, which works well. We are effectively forcing those people out of their comfort zone by being more innovative.

    b We went against the grain. We challenged the ways people were operating for more than 50 years. There was no reason not to automate before, but no one wanted to be the first to go against the grain.

  • We still maintained some amount of human operations in the product, and we think it is worthy because it saves significant effort to engineers. Good engineering is all about knowing when to innovate and when not. Sometimes, overengineering will bring little benefit. For example, people would try to add an auto-reply feature to a conversation with customers. While it takes significant engineering effort, it doesn’t deliver the necessary results. People want to connect with a real person and will seek to speak with a real person, not a chatbot.

  • Work with your potential customers to understand what they want. However, if they want to speed up some things, they will have to get outside of their comfort zone. Explain to them long-term benefits and how embracing a new solution will help them improve their operations.

Discover Plato

Scale your coaching effort for your engineering and product teams
Develop yourself to become a stronger engineering / product leader

Related stories

How I failed at my startup

14 October

There are nine specific building blocks and functional areas every org/company need to work to launch the product and provide services to customers. How effectively founders tackle them determine the destiny of the company.

Mission / Vision / Charter
Scaling Team
Building A Team
Praveen Cheruvu

Praveen Cheruvu

Senior Software Engineering Manager at Anaplan

Scaling a Team in Two Parts: The Product and Manager

2 August

Viswa Mani Kiran Peddinti, Sr Engineering Manager at Instacart, walks through his experience scaling a team, product and his skills as a leader.

Managing Expectations
Scaling Team
Viswa Mani Kiran Peddinti

Viswa Mani Kiran Peddinti

Sr Engineering Manager at Instacart

How Product Management Chose Me

23 June

My accidental journey into product management

Personal Growth
New PM
Career Path
Michael Castro

Michael Castro

Sr. Manager, Product Management at Capital One

How Product Marketing Can (and Should) Help Product Development

20 June

Pavel Safarik, Head of Product at ROI Hunter, discusses the frequently overlooked role of product marketing in getting high user adoption rates for your product.

Goal Setting
Product Team
Different Skillsets
Cross-Functional Collaboration
Pavel Safarik

Pavel Safarik

Head of Product at ROI Hunter

The Intermediate Engineer Syndrome

16 June

How not to stuck at the Intermediate Engineer level

Changing Company
Career Path
null null

null null

null at Unit21