InsTech London Podcast with Hemant Shah: Moving Upstream - Digitizing and Standardizing Customer Data at Source
InsTech London identifies and promotes the use of the best technology, data, and analytics within insurance and risk-management around the world. The company provides insights through events, podcasts, reports, and through a powerful network of industry professionals.
On the podcast, Hemant discusses replacing spreadsheets and emails with technology, the benefits of the network effect and collaborating with partners across the full value chain, and launching a company remotely.
Read the InsTech London Podcast Episode 145 transcript or listen along below.
Moving upstream - digitizing and standardizing customer data at source - Episode 145 highlights
Matthew: Hemant, I really enjoy every podcast I host. But given the history we've got together and your own experiences I am really looking forward to this one.
How can the insurance industry benefit from Archipelago’s data platform?
Hemant: The market has been consuming information and modeling risk for years. The fundamental currency for these decisions is quality property data about customers, particularly large commercial accounts.
Our concept at Archipelago has been to go upstream, connect the data at the source and create insight. That allows customers to share data with the markets digitally.
Matthew: How do you find organizations willing and able to share data?
Hemant: We extract property data from the corporate customer itself, rather than observing from the outside. We didn’t want to publicly launch until we could validate that this would work so we went through a placement cycle on several high-profile corporate renewals.
We identified leading risk managers who wanted to take more proactive action to understand their own risk. We got through to them often not through the insurance channel, but through the digital initiatives these large companies do across the organization.
The risk manager is an important stakeholder in this, but they are part of an ecosystem of people in these companies that we are educating on their own data. There are powerful data sources, which may not be thought of as insurance data, that we can use for insurance.
One of our first customers is a real estate investment trust called Prologis, a large owner of industrial logistics and warehousing space. The risk manager Jeff Bray saw an opportunity not only to create a better insurance experience, but also to add value internally for resilience.
[Editor’s note: read customer case study with Prologis.]
Matthew: Do you work with brokers or just with insurers?
Hemant: All our customers have an insurance broker. We work with the brokers to ensure that the broker can communicate the quality of our customers’ data effectively to the insurance market. Brokers are at the table when we onboard and extract datasets. We work with brokers when staging the data and the broker sends it to the underwriter.
We connect the customer, broker and underwriter in a more efficient way with benefits across the value chain. We have been working with over a dozen of the principal brokerage houses on mutual customers.
Matthew: How are you using machine learning for your data analytics?
Hemant: The crux of the Archipelago proposition is the ingestion and transformation and enrichment of high-quality, accurate and trustworthy data. All data on the Archipelago platform has lineage to the data sources.
We extract information from source documents using machine learning and create a link back to the source document. An underwriter can see not only, for example, that a roof age has been provided, but also the roof report where the data came from.
Matthew: How do underwriters use the data? I assume some want to look at the data in detail whilst others just want to import it into their models and systems?
Hemant: We understand this marketplace. For each part of the value chain, we understand how the different parties want to provide and consume information, and what is too much or not enough information for them. For example, when an underwriting team gets a link, they can explore years of datasets about their customer, or they can just pull the data in and process it.
Some underwriters will spend a lot of time on the Archipelago platform to understand what changed from the last submission, what are the trends, and they will link in technical colleagues too. Others will just extract the model-ready files.
Matthew: How is the data updated?
Hemant: The data lives on-stream. The characteristics of building and the property portfolio changes all the time. As the property goes through its life, we identify what is happening. We don’t overwrite data, we create a record of the lifecycle of these assets: their exposures, incidents and claims.
The data can be interrogated, analyzed, audited and consumed in ways that it is difficult to do with spreadsheets. We are making the whole data experience more efficient.
Matthew: Who are your clients: the building owners, the insurers, the brokers or all three?
Hemant: Our foundational customers are the buyers of insurance. That’s where the data comes from and it’s the heart of our proposition. But we are starting to do business across the value chain.
If a large corporation uses our platform for its data, it can share its information via Archipelago with its brokers and markets. But we are now getting more traction with insurers who see opportunities to use the platform.
800 underwriters have used the platform across over 100 insurance companies across our customers' placements. And, we are now signing up insurance companies as licensed clients.
We also have brokers as clients, some of which are now signing up their own customers directly onto the platform, because it is a better way for them to execute placement renewals.
Matthew: Do you experience a network effect?
Hemant: When large corporations buy insurance, they syndicate risk, they have multiple programs, often with many insurance companies sharing in the risk. Up until now, each insurer has had to go through a redundant process via emails with attached spreadsheets and PDFs.
When an Archipelago customer uses our platform to share its data in support of its renewal, all the underwriters get links, with secure, authenticated access. Each insurer, without needing a contract with Archipelago, can access trusted and standardized data.
Every placement facilitates a powerful network effect. As underwriters get more submissions on the Archipelago platform, they want to work with us more systematically. Underwriters may tell their brokers to bring more submissions on Archipelago, and the brokers come to us.
A key part of our go-to-market strategy has been to understand how the parties work with each other and address the use cases that each one cares about.
Matthew: Are you seeing examples of how sensor data can benefit insurer and insured alike?
Hemant: We recently completed a proof of concept with a customer around how sensor data might inform its view of risk, and how it could collect and share that information with the markets. Archipelago worked with the risk management and digital teams at the insured corporation and one of its lead insurers to explore how it might work.
It is challenging, but the opportunities are significant. As buildings are digitized, new data sources can inform risk and be shared with the markets for underwriting. Those sensors can also be used to reduce risk: predict and prevent. However, it’s not going to happen overnight.
In the early days of RMS, we learnt that making the modeling data actionable at the point of underwriting by the underwriters themselves results in a greater focus on getting better data.
The further upstream one goes back up the original source of data, the better the quality of the data that can be shared with the insurance industry.
A challenge of being an early-stage company is exploring future horizon use cases while also focusing on creating value today. For every conversation I have about sensor data, I have many more conversations with customers struggling just to keep up to date on the property they are acquiring and disposing. Keeping that balance of focus is part of the art of this.
Matthew: What was your experience of starting up a company remotely?
Hemant: It has taken hard work to build and manage the business, ensure the culture and deliver for customers. Since COVID, Archipelago has grown from 30 to over 100 staff.
Before COVID, my co-founders and I decided anyway that we would build a company that was natively distributed; we were going to recruit the best talent wherever we found it. Most of the team works from home offices, all over the US and Europe.
We have an office in San Francisco that I am starting to enjoy going back to. It is hard to replicate two hours on a whiteboard with colleagues over a Zoom call.
Matthew: Is there anything else you’d like to share?
Hemant: The industry spends too much time talking to itself. At Archipelago, we’ve gone upstream to engage directly with the customers of the insurance industry. That grounds our culture and our understanding of the problem to be solved: it’s not just an insurance problem per se, but the customer’s problems. By helping the large corporations solve their problems, you can help the insurers solve their problems.
I encourage others to catalyze innovation with that mindset. We need innovation to scale capacity and deliver more resilience to the economy, given the changing climate.
Originally published on InsTech London. Original transcript edited for standard American spelling.
Don’t have an invite? Simply ask your broker, underwriter, or colleague if they have a code to share with you. Or, request one here.
More from Thought Leaders
In this Q&A, Christine discusses the use of drones in the insurance industry, the need for more focus on large-scale catastrophes, and her favorite hiking app.
In this Q&A, Frank Nutter, President of the Reinsurance Association of America (RAA) shares his thoughts on underwriting, climate risk, and remote work.
"This requires an appreciation for the entire value-chain, with a healthy respect for why and how work currently gets done; not naively, as in 'I have an algorithm...' but how technology and data can slip comfortably into how the market currently works, and then drive change from the inside out."
All rights reserved. Trademarks are the property of their respective owners.