Podcasts
No legacy, no cry – Greg Hendrick, CEO, Vantage
New carriers face big challenges of scale and competition, but they have major advantages by being free of legacy – whether that’s unwanted business or ageing technology.
This startup advantage was a focus of conversation on the latest Voice of Insurance podcast episode, with industry veteran Greg Hendrick, CEO of Vantage Risk Companies, as guest.
Hendrick is still only in the fourth year of building out a major specialty insurer, reinsurer and capital markets partner, completely from scratch.
The reinsurance bit is the most developed in scale along this journey, he reflected, but technology is a major tool in boosting the firm’s growth trajectory across classes of business.
Relying on a workbench strategy, combining internal and external data insights, has been the consistent vision of the firm, he explained.
“From the get go, it was meant to be a tech enabled company… “We have clean technology, we’re cloud native,” Hendrick said.
“We rent the things – policy issuance, claims, general ledger – that others have already built well,” he said. “We build in-house studios around that, some people call them work benches, with the ability to harness information that either we generate ourselves through our own analytical models, or that we harness from outside from others and bring that to the underwriters desktop.”
Host Mark Geoghegan highlighted Vantage’s construction surety product for subcontractor default risk, as one area of product innovation, using data and artificial intelligence (AI) to accelerate sophisticated product design and distribution that a generation ago would have been the domain of only much larger insurers and insureds.
“It’s an example in using the technology to do something on distribution. We feel there’s work that will give us great lift around operational efficiency, particularly as you get into AI and being able to take large amounts of information and synthesize them down to the critical facts to get that in front of the underwriter, or decision maker, it could be anybody,” Hendrick said.
“We’ve built our ‘Vantage GPT’, if you will, inside our own firewalls, so that people can go in and use that to explore,” he continued.
The next aspect is answering the question of how to build analytical models to make better risk selection or risk evaluation decisions, he emphasised.
“We’ve done work around assessing the riskiness of legal venues, work around commercial auto, and work around the lawyers’ E&O [errors and omissions] space as well,” Hendrick said.
“Because we’re a new company, we have the advantage of being nimble and lean, with a modern technology stack. We don’t have the historical data, but what has become apparent over the last five to 10 years, is that there’s so much more data available from the public domain.
“We were able to build each of those three models, with a little bit of our own proprietary [data], but almost all built from publicly available datasets. Our teams are using them and matching them up with exposure and building models,” he continued.
AI in its many applications can mean “different things to different people”, he noted.
“This current wave of tools that have come online are going to let us radically change the way people operate, and let the best talent, get the best insights, to make the best decisions. That’s what we’re wired towards and built towards advantage,” Hendrick said.
Caution is required, to avoid data biases and stay in line with regulatory expectations, he acknowledged, but the area is exciting for the future. The focus has been on building a safe version in-house for teams to experiment with, particularly for language recognition, to produce properly cited insights. About 40% of the staff have been using these AI tools, which he aims to increase.
The pace of change has been dramatic, he emphasised. Personal auto and homeowners’ insurance business, with “large numbers of smaller transactions”, have already been largely automated in recent years, he noted, replacing or almost replacing human underwriting teams.
For large commercial risks, there is not the same level of trust in fully automating transactions, he emphasised, such as for lower volume, higher value business, such as directors’ and officers’ liability (D&O).
“I don’t think we’re going to get to a [comparable] spot where the public D&O market in the US with a few thousand risks. It’s not credible enough, so what we’re looking for are insights, not answers, so the underwriter can synthesise that and make the decision,” Hendrick added.