Most companies’ digital journeys begin by migrating legacy applications to the cloud — the theory being that lifting and shifting workloads can provide a fast onramp to making good on services and capabilities unique to the cloud.
But home and automobile insurance company Allstate is taking a different approach. Zulfi Jeevanjee, EVP and CIO, believes the best way to build and align next-generation business processes and modern IT platforms is to build anew, and so he is taking a cloud-first approach to digital transformation, dumping out all legacy infrastructure along the way.
The result, Jeevanjee says, is a technology-driven business strategy “that’s a very empowering thing.”
To fuel its transformation, the Northbrook, Ill.-based insurer has rebuilt its core application for claims processing, sales, and support, and plans to overhaul its entire portfolio of business processes, all with the aim to enhance and accelerate the customer experience. Nearly 40 per cent of the company’s business processes have now been digitised, and filing claim time — a key measure of customer satisfaction — has been reduced from four minutes to 43 seconds, according to the company.
Aside from digitally rebuilding its processes, Allstate has also methodically adopted a multi-cloud architecture based primarily on AWS for containers and development, Google BigQuery and Vertex and Microsoft Azure GenAI for specialised AI workloads.
Many companies are taking this same approach — using BigQuery and Vertex for generative AI pilot applications to gain new insights and produce better business results. At Google’s recent Next conference, for instance, L’Oréal and Shopify announced they are using BigQuery in gen AI pilots to accelerate and optimise business processes.
Allstate’s Jeevanjee credits the company’s top brass for grasping the importance of aligning its IT infrastructure with new business processes, which had to be rearchitected to maximise the rewards and minimise the risks of its switch to digital business. And it has all been undertaken with a cloud-first approach.
“It was built and designed to run on the cloud,” Jeevanjee says. “It wasn’t built to run on-premise.”
Re-engineering how claims processing gets done
In many ways, Allstate is still at the beginning of its digital journey — only 3 per cent to 4 per cent of claims “on the book” are currently processed on the cloud and much data remains in on-prem XML databases common to the insurance industry — though the advanced technology and business modernisation blueprint is rock-solid, the CIO says.
Allstate’s global IT team, with staff based in Illinois, Northern Ireland, and India, developed the infrastructure and new processes, deploying them first in Allstate’s home state of Illinois for nine months to determine how customers would react to the new digital experience before rolling it out in Tennessee. Allstate expects to be up and running in 10 states this year for automobile policies and 19 states for rental insurance.
“We learned a ton about how customers react to us, and this is what I mean about changing our organisation,” Jeevanjee says. “Getting a fast focus on that experience was extremely helpful in making the experience the best we possibly could.”
Allstate began moving to the cloud in 2019 but paused during the pandemic and began the deployment of its curated multi-cloud blueprint upon Jeevanjee’s return to the company in 2022.
As the company re-engineers and tests every aspect of its auto and home insurance applications in each state, it is building it anew on AWS as the core workhorse, while making use of Google and Microsoft cloud services for specialised applications such as AI.
The methodical approach does not mean Allstate is a technology novice. The company deployed automation in business processes long ago to remove manual steps and speed up the transactions.
In fact, the company has deployed several machine learning models for key applications, including claims predictions — such as determining whether an automobile in an accident is a total loss — and more evolved homegrown-trained machine learning models that make these recommendations independently.
Enhancing customer experience with gen AI
Allstate has also developed a generative AI application based on ChatGPT 3.3 informally dubbed MyStory, which vastly reduces the time for customers to report a claim after an accident or incident. Rather than go over and over an account of an accident to various clerks and adjustors, customers recount the incident once and it is summarised in a document and delivered to all necessary parties.
When a human representative calls the customer, they are fully informed and ready to move to the next stages. “We put people in the flow when it matters most,” says Jeevanjee, noting this important change in the process has vastly improved customer satisfaction.
Arun Chandrasekaran, distinguished VP and analyst at Gartner, says he is seeing similar generative AI pilots being undertaken at financial services and other insurance companies, as well as technology, media, and entertainment organisations.
According to Chandrasekaran, insurance companies have been leading the way in implementing technologies such as speech-to-text and transcription to speed up claims processing and customer satisfaction, and use of generative AI, which, although presently low in the insurance industry, is projected to grow significantly in the next 12 months, is just the latest stab at achieving these outcomes.
“These use cases have been around in the insurance sector actually for a while now,” he says. “Where language models can add value essentially is in terms of better cognition, because they’re able to craft responses” and bring in more information to re-create events in an effort to more accurately handle claims. “As models evolve in the future to become fully multi-modal, they [will be] able to traverse different types of data.”
One of the four largest automobile and home insurers in the US, Allstate manages roughly 190 million policies protecting cars, homes, motorcycles, health, disability, lives, personal devices, and identities, according to a company spokesperson.
The organisation employs roughly 54,500 and supports 10,100 exclusive Allstate agents and 51,900 independent agents. About 8,000 employees are dedicated to IT throughout the company and roughly 1,000 of these employees report directly to Jeevanjee.
The CIO’s biggest challenge in its cloud-native approach from the start is “changing the culture to become a digital company,” Jeevanjee says, noting that getting the commitment from the entire C-suite to embrace the plan has made his job far easier. “What we’ve done is look at all of our processes and make them digitally ready.”