Posted on

Joanne Richardson, former Health Director at AXA, on how insurers can overcome legacy systems, unstructured data, delays and inflation

Joanne Richardson spent over 20 years working in health insurance across multiple countries for AXA, including Mexico, France and the UK. She spent five years as a Health Director in the Group head office in Paris, where she developed group strategic initiatives such as Care Coordination. She then worked in Group Risk Management, leading in-depth reviews of the major AXA health entities, helping to ensure long term profitable growth through International best practice.

She spoke to us about the key challenges facing insurers today, and how Sprout.ai helps solve them.

Sprout.ai: What relationship do health insurers want to have with customers today?

Joanne Richardson: Health insurers today have a clear desire to be more customer-centric. They aim to help people be healthy and stay healthy.

They want seamless quick claims processing, and then they want to be able to concentrate on adding additional value services to customers. That means giving them access to the latest medical innovations, providing them with support to manage chronic illnesses, such as apps for diabetes management, understanding their needs and having the time to provide dedicated support for complex cases. 

This enables customers to live healthy lives and to return to health as quickly as possible.

Sprout.ai: And what relationship do health insurers have with medical providers?

JR: Many insurance health insurers have a wide network of medical providers. These providers want the right money to be promptly paid for the work that they’ve done. 

Insurers want to be able issue timely, accurate payments. They want to establish a high quality medical network at an optimal price point that is more a partnership than a transaction. 

Shareholders need to be considered here too. Health insurance is a small margin product, so profitable growth has to be done through efficiency, innovation, and a good reputation in the market. 

Sprout.ai: What challenges are health insurers facing today?

  1. Legacy Systems

JR: As mentioned, providers want to be paid accurately and quickly. Insurers want claims to happen seamlessly and without problems. However, because of the way that many health insurance products are designed to fit state provision, they are managed on old bespoke systems. These are slow to change. They’re incredibly expensive to replace, and have an awful lot of manual intervention. 

Health claims are as prone to fraud, waste, abuse as all insurance claims. It’s estimated that 5-6% of all global claims payments are paid inaccurately. 

Furthermore, it’s a slow process. It’s often not automated, and particularly complex claims take a long time to get through. 

How does Sprout.ai help solve the problem of legacy systems?

JR: One of the things that attracted us to Sprout.ai in the first place was the fact that its claims system sits on top of your old bespoke systems, so you don’t need to replace them. As anybody who’s ever tried to replace the claim system in health will know, this is expensive, takes a long time, and often does not result in success. 

  1. Unstructured data

JR: We then go forward to look at data that’s linked to claims. Old claims systems were written to capture the minimum amount of data required in order to pay claims, because data capture takes time and multiple attempts. 

Lots of data is provided during the health claims process, but an awful lot of it is not structured. When it is on claims forms for medical information, it’s often not captured. This means that there isn’t enough structured data within the systems to be able to understand customers’ needs, to analyse fraud, waste and abuse, and assess quality of care.

How does Sprout.ai help solve the problem of unstructured data?

JR: One of the things we liked about Sprout.ai was its ability to capture much more data, capture much richer data, and then format it so it is structured. This then allows analysis into needs, leakage, and care quality to happen. And as Sprout.ai sits outside of the legacy systems, you don’t need to adjust all your data structures.

  1. Delays and inflation 

JR: First, we’ve still got the backlog of treatments waiting to happen because of COVID within state systems, and also to an extent within private systems. This is an opportunity for health insurers. People are turning to private insurance as they realise how long they may have to wait to get treatment within the state. However, those treatments are also delayed in the private system. Delaying treatment often means that the final interventions are more complex, and consequently more expensive. 

Second, there’s the challenge of inflation.  We can see what’s happening with general inflation with the war in Ukraine, as well as the supply chain problems we have post COVID. We understand from historical analysis that medical inflation tends to be higher than general inflation. It isn’t clear yet exactly what the impact is going to be, but health insurers can’t wait a year to find out what the inflation results in. They have to price for it now. 

That means that, unless they can find ways of reducing costs in the short term, they are going to end up putting their prices up to a point where for some people health insurance will become unaffordable. And let’s face it, the people who stop paying health insurance are not the ones having treatment. They’re the ones who aren’t. That means that the subsidy that comes from those people to the people who are claiming is reduced. Therefore, prices go up again. It’s a vicious circle. 

So, it is important that insurers can react to both the opportunities of new health insurance customers, and also price health insurance in a way that they can retain their existing customer base. 

Sprout.ai: How does Sprout.ai help solve the problems of delays and inflation?

Our AI powered technology delivers fast and accurate claims decisions, enabling you to better serve your customers.

It extracts all of the data in the claims documents from providers, customers, doctors. Then, the data is structured and put into the insurer’s system. Natural language processing (NLP) interprets the documents and unstructured data, then triages the claim, and checks for fraud, anomalies, and coverage. Then, using machine learning, our technology checks the claim against historical claims and recommends the next best step. The claim can either be settled right away, or, if it’s more complex, passed on to a claim handler.

Some answers have been lightly edited for clarity.

Watch the full Webinar here.

To learn more about how Sprout.ai can boost efficiency and enhance customer experience, book a call today.

Posted on Leave a comment

The insurance industry in 2023 and beyond

The global insurance market faces many challenges this year, from ever higher customer expectations, to soaring inflation, climate change, cyber security risks and pandemics. Few insurance lines will be unaffected by these issues. Insurers operating in every jurisdiction and across every line will need to think about how to deal with them.

Here is how Sprout.ai envisions the year ahead playing out, as well as some thoughts on what might happen further ahead.

Inflation will drive up insurance premiums

Inflation is significantly impacting the insurance industry as the rising cost of goods, repairs, and services drives up claims payouts for insurers. As a result, premiums may also increase as insurers look to offset their losses. Traditional insurers may have an advantage in this market as they have a more established brand and deeper coffers, which customers may find more reassuring during uncertain times.

However, traditional insurers also have higher operational costs compared to technology-first insurance providers. To remain competitive, traditional insurers will need to optimise their operations and invest in technology that allows them to provide better service at a lower cost. The insurers that can do this quickly and efficiently will come out on top.

The insurance industry will also need to meet cost-of-living pressures

Despite the higher costs caused by inflation, insurers will need to adapt by offering flexible and competitively priced products to their customers as cost-of-living pressures continue to bite. To stay competitive, insurers should invest in product innovation and offer flexibility and choice. In 2023, customers will look for lower cost policies that are tailored to their specific needs.

To help prevent or mitigate potential losses or risks, many insurance companies will provide additional services to their customers. These will reduce exposure, reduce overall claims volumes, and improve profitability for the company. For example, offering health insurance customers periodic health checks or gym memberships can improve customers’ health and well-being, which reduces the likelihood of making a claim, while generating additional revenues if the right partnerships are created. This can be seen as a win-win situation for both the insurer and the customer.

Insurtech startups will face funding challenges

The decline in tech funding in 2022 has affected the insurtech sector, with companies experiencing a drop in later-stage growth funding and reductions in valuations. Investors are becoming more cautious and focusing on growth metrics, which poses a risk for insurtech startups that rely on funding to fuel their growth. However, this presents opportunities for well-financed companies with deep pockets.

 Traditional insurers may be able to regain market share or drive consolidation in the market as they partner with new and traditional insurance companies. The future of the industry is uncertain and the funding situation may change, but for now, venture capital and private equity funds still have money to invest and insurtech is still considered a “hot” area.

Artificial intelligence and the Internet of Things will continue to disrupt

Artificial intelligence (AI) is overhauling the insurance industry by automating complex tasks that were once difficult to perform with traditional software , as well as the simpler, repetitive ones. With the use of deep learning and other AI techniques, many decision-making processes in insurance can be automated or assisted by software. For example, identifying the circumstances that led to a customer’s insurance claim and cross-referencing it with the terms of the current policy is a complex task that can now be automated. More and more insurers are leveraging this type of technology.

In 2023 and beyond, insurers will begin to use the Internet of Things (IoT) and wearable devices to track customers’ behaviour to identify and mitigate the risk of claiming on insurance. This technology can be helpful in tracking patterns and behaviours, or monitoring health and activity levels for health insurance customers. IoT and wearables can provide insurers with valuable data that can help them better understand the risks their customers face, and allow them to offer more tailored insurance products.

Customer-centric approaches to claims will become the norm

Today, the majority of insurers structure their operations by insurance lines. For customers, this is highly inefficient. One customer holds multiple policies for different insurance lines. It becomes even more inefficient and problematic if the customer needs to claim on more than one type of insurance. 

Here’s what that problem looks like:

Customer ‘A’ owns a house where a fire breaks out. The fire destroys the entire contents of the home and causes the customer severe respiratory issues. Customer ‘A’ will need to claim on up to three insurance policies: buildings, contents and health insurance. This are likely to be held with different companies.

Customer are often emotionally, financially, or physically vulnerable when filing an insurance claim. Improving efficiencies in the claims process will help reduce that stress and enhance the overall customer experience. 

Flexible workforces and embedded technology will enable resilience to global crises

The Covid-19 pandemic had a far-reaching impact on health insurers. Claims volumes increased by 40-60% and have not fallen since. This dramatic increase has been driven not just by Covid infections, but by the knock-on impact of the pandemic on other long-term health conditions. Insurance companies have had to retrain claims handlers from other areas such as motor insurance to manage health claims. This has been a long and expensive process as it can take up to 9 months to train claims handlers.

This highlighted the challenge of dealing with a crisis and a spike in claims using legacy processes and technology. In the years ahead, global warming will present enormous challenges for insurers in the form of wildfires, rising sea levels, and floods. Insurers will prepare for these events by automating much of the claims process and increasing the flexibility and elasticity of their workforce.

To learn more about how Sprout.ai can boost efficiency and enhance customer experience, book a call today.

Posted on

Customers’ expectations vs reality

Insurance customers anticipate a far faster claims process than the one they experience. How can insurers catch up?

More than 1 in 5 (21%) insurance customers expect claims to be resolved within hours. A total of 100% of 18 to 24 year olds expect a resolution on an insurance claim within one week, according to our recent research into customer expectations of the claims process.

Download the report: Responding to rising customer expectations in insurance

We surveyed 1,000 consumers about their attitudes towards insurance claims. Of that number, 99% were either solely responsible for choosing and purchasing insurance products, or made the decisions with a member of their household. 

They told us:

  • 43% of customers across multiple insurance lines waited over two weeks for a claim to be resolved
  • 62% of claimants with a “good” or “very good” customer experience said that they stayed with their existing insurance provider
  • 31% of claimants with a “good” or “very good” customer experience said that they would use the same provider at some point in the future
  • 19% of those with a “bad” or “very bad” experience said they are still a customer of their existing insurance provider
  • 89% with a “bad” or “very bad” experience said they would not purchase a policy from the same insurer in the future

Customer service matters most during claims

In today’s highly competitive insurance markets, margins are tight. It is easy for customers to switch between different providers. Delivering excellent customer experience is, therefore, vital for customer retention.

After a customer purchases an insurance policy, the next ‘touchpoint’ is typically when a claim is filed. The customer is likely to be vulnerable or distressed. The insurer needs to deliver and prove that the customer’s purchase was a wise investment.

The time it takes for a claim to be processed and the ease of speaking to a claims handler have a significant impact on the overall customer experience. This can boost customer experience, and in turn, the insurer’s Transactional Net Promoter Scores (TNPS).

How Sprout.ai helps insurers meet their customers’ expectations

Customers want: Quick resolutions

Sprout.ai empowers insurers to settle many claims in real time, and speed up the time it takes to process more complex claims.

Customers want: To be able to speak to a handler

Sprout.ai performs many of the repetitive data entry and checking tasks that take up claim handlers’ time, freeing them up to speak to customers.

Customers want: Confidence that their claim has been processed fairly

Sprout.ai is free from bias and 97% accurate. 

How it works

Data extraction

Documents submitted for a claim can include PDFs, handwritten reports, images and freeform notes, as well as structured text and digital content. Manually processing these documents is time consuming and open to inaccuracy, fraud and wastage. 

Our NLP and patented OCR technology can extract all relevant information from any type of document submitted as part of the claims process. As a result, it can be used to automate and provide insights. 

Data enrichment 

We refine and improve the data we have captured by up to 300% by attaching external data points such as fraud checks, replacement prices, claims history and much more. This helps us validate the claim, checking for fraud, reduce waste and abuse and identify outliers. 

Policy checking

Our technology enables fast, accurate and superior policy checking and claim validation. It takes all relevant information and validates it against the policy documents to check whether the claim is covered under the customer’s specific policy. 

Our patented NLP solution can automatically check for coverage a moment a claim is made due to its deep understanding of claims and insurance related language, including synonyms for the same word (e.g. waste, garbage, rubbish). 

As a result, claims can be processed in real time, or far faster than before, freeing up handlers to focus on customer service.

To learn more about how Sprout.ai can help you process claims in real time, book a call with one of our claims experts

Posted on

How can insurers attract Gen Z?

Two young women looking at a phone

Real time claims processing gives insurers an edge that Gen Z care about

Young people have the highest consumer expectations of any age group according to multiple surveys and reports. Whether they are buying an insurance policy, opening a bank account or sending a parcel, they look for and expect exceptional customer service. Digitally literate, they are the group most likely to access and purchase products and services online. If providers fail to meet their high expectations, they are happy to vote with their feet and change providers. 

Our own recent research into insurance customers’ expectations around the claims process confirms this. Of the 1,000 insurance customers we spoke to, 100% of those aged 18 to 24 said that they expect a claim to be settled within a week. 

They are likely to be disappointed. Our research found that 43% of customers across multiple insurance lines waited two weeks or more for their claim to be resolved.

Download the report: Responding to rising customer expectations in insurance

We also discovered:

  • 18% of 18-24 year olds buy insurance solely from digital players. In contrast, no 55-64 year olds said they would consider buying insurance from a digital player only
  • 50% of respondents who buy solely from traditional insurers would consider buying from a challenger brand
  • 34% already buy policies from both traditional insurers, such as AXA, Aviva and Allianz, and challenger brands, such as Lemonade and WeFox

The allure of digital challengers

There is a significant adoption of digital insurance products Gen Z and millennials. They actively look for providers offering instant digital access, transactions and customer support, as well as fast settlement times.

The insurtechs providing these services are thriving. Lemonade, the US digital insurer, recently released its best ever results. It saw 76% growth in ‘in-force’ or active premiums, more than triple the number it had two years ago. The company has 1.7 million total customers, 30% more than a year ago. 

In Europe, Wefox is gaining traction by selling insurance products through in-house and external brokers, rather than taking the direct-to-consumer route. In July 2022, it raised $400 million in a Series D round at a valuation of $4.5 billion. 

How can legacy insurers keep up?

Legacy technology and processes are simply too slow to keep up with Gen Z’s expectations. Traditional insurers have focused on innovating and digitising the front end experience of insurance products, but less on the claims experience.

Legacy insurers are at risk of losing ground to these digital players if they do not invest in technologies that enable excellent customer service through the claims process. 

Although they increasingly offer the same, intuitive User Experience (UX) consumers have come to expect from their favourite grocery delivery or ride-hailing apps, back-end processes also need to be updated. This can drive growth and reduce costs for insurers operating on increasingly tight margins. 

How Sprout.ai helps insurers attract and retain Gen Z customers

Gen Z consumers want: Quick resolutions

Sprout.ai empowers insurers to settle many claims in real time, and speed up the time it takes to process more complex claims.

Customers want: Great customer service

Sprout.ai performs many of the repetitive data entry and checking tasks that take up claim handlers’ time, freeing them up to speak to customers.

Gen Z consumers want: Confidence that their claim has been processed fairly

Sprout.ai is free from bias and 97% accurate. 

Gain a reputation for efficiency and boost tNPS 

Sprout.ai’s intelligent claims automation engine empowers insurers to resolve claims in a time frame that meets customer expectations.

This can boost customer experience, and in turn, Transactional Net Promoter Scores (TNPSC), which indicates how likely a customer is to recommend that company’s product or service to others, based on the customer experience of the joining process, resolution of a support question, or interaction with their customer service representative. 

How it works

Data extraction

Our NLP and patented OCR technology can extract all relevant information from any type of document. That includes everything from PDFs to handwritten reports, images and freeform notes, as well as structured text and digital content. Manually processing these documents is time consuming and open to inaccuracy, fraud and wastage. 

By automating data extraction, insurers speed up processing times and gain a valuable bank of information that can be used to generate insights into customers and their claims.

Data enrichment 

By attaching external data points such as fraud checks, replacement prices, claims history and more, we can refine and improve customer data. This helps validate the claim, reduce waste and abuse, and identify outliers. 

Policy checking

Our technology takes all relevant information and validates it against the policy documents to check whether the claim is covered under the customer’s specific policy. This enables fast, accurate and superior policy checking and claim validation. 

Our patented NLP solution has a deep understanding of claims and insurance related language, including synonyms for the same word (e.g. waste, garbage, rubbish). As a result, it can automatically check for coverage a moment a claim is made due to its As a result, claims can be processed in real time, or far faster than before, freeing up handlers to focus on customer service.

To learn more about how Sprout.ai can help you process claims in real time, book a call with one of our claims experts

Posted on

AI Implementation in the Insurance Sector

A tablet with an insurance form using AI

Artificial Intelligence is touching many facets of the world around us, and is being put into use in almost any industry you can imagine. The insurance sector is no different. At Sprout.ai, we are using the ever evolving power of AI to benefit insurers and the millions of people they serve.

What is AI? 

AI is best defined as intelligence performed by machines. It is not the same as rule-driven automation, which we see in the robotic assembly lines that have underpinned manufacturing since the 1970s. AI operations are far more advanced. They don’t just carry out repetitive actions according to their programming, but also develop through self-learning, without the need for human intervention.

How can AI benefit insurers?

To an extent, AI is already embedded in both personal and commercial insurance lines. Consider a motor vehicle accident claim – a type of insurance claim that almost everyone makes at some point in their lives. Numerous steps in the claims process are already streamlined and processed much faster thanks to AI. Identity verification, checking coverage is in place, collating information about the accident from forms, and performing checks for evidence of fraud would all fall into this category.

Sprout.ai takes AI capabilities to the next level. Our patented technology can process claims forms and extract the meaning of a claim far, far, faster than a human, which is ever more useful in a world in which data and digital footprints are growing exponentially. Its optical character recognition (OCR) capabilities can read handwritten language at a level that surpasses human competence, in languages from Greek to Japanese. It can help eliminate loss making or incorrect decisions by automating the process and referring to thousands of data points. 

This reduces manual workloads, shrinks the time to resolution, eliminates human error, and provides value for policyholders during their time of need. A claims process that might previously have taken weeks can potentially be wrapped up in hours. At the same time, costs are reduced for the insurer, and enhanced transparency boosts confidence for the insurer and their customers.

AI offers significant ROI for insurers:

  • 86% of insurance companies say they have created better customer experiences using AI
  • 75% of of insurance companies say AI helped improved decision-making
  • 75% of insurance companies say AI helped innovate insurance products and services
  • 75% of insurance companies achieved cost-savings using AI
  • 64% of insurance companies increased productivity using AI

Statistics from PwC 2022 AI Business Survey

AI for complex claims

As we mentioned, it’s not only routine insurance lines that benefit from the use of AI. Sprout.ai’s Natural Language Processing (NLP) can assimilate almost inconceivable quantities of policyholder and claims data, which significantly cuts down the time and manpower it takes to process highly complex claims, such as those in marine or commercial property. Sprout.ai’s conclusions give skilled and experienced insurance professionals a highly reliable and robust basis to inform their decision making.

Changing times in the commercial insurance market

Some business customers are already experiencing fully automated insurance processes. In the food industry, for example, sensors identify when a refrigeration unit goes down, resulting in spoiled food, and triggers a fully automated insurance pay out to the policy holder. These systems are now appearing in other areas of insurance. As challenger insurance companies offering premiums from as little as £4 per month take a growing share of the market, incumbent insurers will need to ensure they offer the optimal claims experience.

A common worry is that Artificial Intelligence is replacing people and eliminating jobs. In fact, AI removes the need to waste time on repetitive, time consuming tasks, such as data entry, freeing people to dedicate time to valuable and rewarding tasks which optimise the customer experience. This is akin to the revolutions in farming in the late 19th century and manufacturing in the late 20th century. 

In the years to come, AI will become even more ubiquitous across industries and applications. As adoption becomes the norm, there’s no doubt it will create new types of insurance products and new classes of risk. Inevitably, some players in the global insurance market will fully embrace these new areas more quickly and enthusiastically than others. One thing is for sure: at Sprout.ai we are proud to be at the forefront of AI and working with its early adopters to bring about a new age of insurance handling.

Posted on

Meeting Insurance Challenges Across the World with Sprout.ai Claim Solutions

At Sprout.ai, we are proud to work alongside many of the world’s leading insurance providers, using our proprietary solutions to make real changes that benefit both our clients and their millions of customers. One thing we’ve noticed is that providers in diverse parts of the world face challenges that seem quite different at first glance. But often, they come down to the same root causes. Two recent projects, with clients in Chile and Japan, have brought this into sharp focus.

Tackling spiralling costs for a health insurance provider in Chile

The first project was with one of the leading health insurance providers in Latin America a business that processes more than a million claims every year. The company called us in to help them address two different but related challenges. The first was that claims processing costs were going through the roof. The second was that claims handlers were spending so much of their time on data entry that there was practically no opportunity for them to do anything else.

Our client was effectively treading water to process claims as quickly as possible. This constrained its ability to devote sufficient resources to activities like fraud detection and prevention. The result? Spiralling costs and ever-longer processing times.

Sprout.ai built a Straight Through Processing solution based on OCR and NLP technology to read claims data. We then trained it using historic claims data. This had the immediate effect of freeing up time for claims handlers to focus on the insights drawn out by the tool instead of being buried in repetitive data entry work. At the conclusion of the pilot, accuracy was at 97-99 percent (compared to the industry standard of 80 percent), while 70 percent more fraudulent claims were detected.

Reducing processing time for one of the largest insurers in Japan

Meanwhile, almost 11,000 miles away, we had a team working on a pilot project with a client in Japan. Here, the business’s new bespoke healthcare offering had resulted in a sharp increase in processing times due to its complexity – there were hundreds of different policy variations, exclusions and so on. The result was an average processing time of 30 days – 20 of which were being spent on data entry.

While the challenges here looked very different to those faced in Chile, the solution, once again, lay in automating a straight though processing solution to reduce the time being spent on data entry. Setting up the solution meant some late nights with teams from client and Sprout.ai working together to consolidate the necessary training data to get the pilot underway. But the results were certainly worth the effort.

Processing times were reduced by 90 percent to an average 2-3 days, while our client also observed significant cost reduction. All this came with no impact at all on accuracy, which at 96 percent was above average human levels.

Diverse problems, a single solution

The global insurance market faces a whole host of challenges, especially in these unusual times. Sprout.ai’s automation processes can contribute significantly to solving a range of these problems, increasing efficiency, reducing costs and allowing human experts to step away from the repetitive admin task and use their knowledge and experience to add real value.

Posted on

Sprout.ai tech means Zurich can now resolve property claims within 24 hours

zurich property insurance logo

The insurance media is buzzing this month with the news that a trial of the Automated Policy Checking AI solution developed by Sprout.ai means Zurich Insurance can cut property claims settlement times to sub 24 hours. The announcement was made following a three-month pilot of the technology, run between December 2020 and February 2021.

What is Automated Policy Checking?

This AI engine uses what we call explainable AI – tools like Natural Language Processing and Knowledge Graphs add layers of context and aid decision making. To put the software’s capabilities into context, its speed and ability are the equivalent of a claims handler with 100 years’ experience and a reading speed of 10,000 words per microsecond. That equates to 600 million words per minute, compared to the most commonly cited average human reading speed of 300 words per minute.

The software is not, however, designed to replace claims handlers any more than a calculator is designed to replace an accountant. On the contrary, it is a tool that the handler uses and drives in order to aid decision-making and to triage and process claims better and faster. That was the exact outcome of the three-month pilot at Zurich.

Better accuracy, improved transparency

Checks found an accuracy rate of more than 98 percent during the pilot, which exceeds the average accuracy of human claims handlers. By automatically providing links to the insurance ombudsman data used to arrive at each recommendation, it also means there is greater transparency to support every decision.

The AI solution was developed using more than 20,000 historic claims. However, it is important to understand that those were just the beginning – they can be thought of as the training material used so that the software could learn its craft. Every new claim added to the database brings new knowledge and experience, and as with a human claims handler, this will help the AI to keep getting better at what it does.

A success for all

The pilot is part of a company-wide commitment to innovation at Zurich aimed at delivering a faster service with the best possible outcomes for every customer. Everyone involved in the pilot, from the claims handlers to the executive decision makers, felt that it was an unqualified success in helping the company achieve those goals.

One highly experienced claims handler who was directly involved in the pilot took the trouble to email us at Sprout.ai afterwards. Here’s what she said:

Just want to say thanks for the opportunity to participate in the pilot, it’s been great fun and I’ve certainly learned some things along the way! The Sprout.ai team have been second to none and fully invested/listened in our feedback to ensure it’s a success for all! I will certainly be singing your praises.”

Meanwhile, Amy Brettell, Zurich’s Head of Customer, UK Claims described this project as “a fantastic example of a collaboration which provides a true solution to some of the issues that often stand in the way of creating customer loyalty and helping rebuild trust in our sector.”

Posted on

The Sprout.ai Growth Game – Meet the Players

The team at Sprout.ai continues to grow. It only seems like yesterday that we were welcoming several new faces last summer. Since then, however, the team has grown still further and was joined by seven more new members in the fourth quarter of 2020. Now 2021 is well and truly underway, we thought it was high time we caught up with some of them to find out how they have settled in over their opening weeks and what they have planned for the months ahead.

Tom Batstone – Talent Acquisition Manager

Tom joined us in December, and has been the driving force behind our growth plans for 2021. With experience building high-calibre Tech and Product teams across a number of startups and scale-ups, Tom is perfectly placed to help lead our headcount expansion for the future.

There were a number of things that compelled me to join Sprout.ai at the start of such an exciting journey. Firstly, the mission: the idea of building software that has a meaningful impact on people’s lives, helping individuals when they’re at their most vulnerable, really resonated with me. Second was the ambition: Niels and Raphael have such passion for the product and what it could evolve into in years to come. That enthusiasm is incredibly infectious, and makes it easy to realise you’re building something truly game-changing. Finally, the people: there is such a welcoming, open culture here that everyone works hard to protect and champion; from my first company-wide huddle, to the fortnightly team events, I’ve felt like part of the team from day one.

What excites me most about Sprout.ai is the growth trajectory. We’ve set ourselves ambitious goals, but so far, I have every reason to believe we will surpass them. With the growth plans we have this year, the company will soon look very different to the one I joined, but that brings new learning opportunities, new teammates, and even more scope to build something truly impactful. The future is very bright indeed.

Eduardo Rocha de Andrade – Senior Data Scientist

Eduardo spent three years working with R&D in the fields of optics and photonics for a telecommunication company before he decided to make the switch to data science. He first joined a startup and then one of Brazil’s top AI and software development companies, where he was able to work with applied AI to computer vision and video analytics.

It is very rewarding to be able to work with something that has the potential to change our society for the better whilst being intellectually motivating. In particular, I find it fascinating to work with data of multiple domains, for example as images and natural language. I’m very excited about being able to deploy our AI technology into production, which could positively impact the lives of thousands of people.

Dr Ivan Sorokin – Senior Data Scientist

Before joining Sprout.ai, Ivan had more than three years’ experience in building production-level deep learning models in the speech recognition area.

For me, working in Sprout.ai provides a huge opportunity to explore new AI technologies. In order to make a working solution for the insurance industry, you need to be able to work with different types of data: images, text, tabular data, and even audio or video. Having now been here for several months, I can say that all my expectations have been exceeded. In this short period of time our team has already successfully applied several OCR and NLP models of varying complexity.

This is just the beginning. There are still a lot of interesting problems to be solved. For example, how to effectively integrate different modalities, how to move from simple tasks to high-level claim scenario comprehension and action reasoning. I’m looking forward to meeting these challenges and more in the months ahead.

We look forward to catching up Tom, Eduardo and Ivan again in the months ahead, as well as introducing you to more new members of Sprout.ai’s growing team.

Posted on

What do self-driving car technology and insurance claims have in common?

As we advance into the 2020s, the use of AI is becoming increasingly commonplace and touches on almost every aspect of our lives. Of course, some applications of AI are more headline-grabbing than others. For example, self-driving car technology captures the imagination because it is the sort of thing we have all seen on TV and in movies, from Knight Rider to I, Robot – themovie that allegedly inspired Elon Musk’s ideas for the first fully autonomous vehicle.

Here at Sprout.ai, we also work with AI on a daily basis to revolutionise the customer experience in insurance claims through enabling insurers to settle claims in under 24 hours (as opposed to 25 days). When you examine the challenges that we face and the obstacles to be overcome, it soon becomes apparent that Sprout.ai has more in common with Tesla than you might think.

The power to imagine

Anyone who drives will appreciate that one of the biggest challenges to a self-driving car is the almost infinite possibilities of what can be encountered on the road. Unpredictable behaviour from other road-users, obstructions, sudden changes in weather, a mechanical failure, the list goes on and on. Yet we face exactly the same challenges when considering the variables attached to an insurance claim. For example, the medical documents used in claims have numerous variations, there is no single format, people use different abbreviations and colloquialisms and so on.

It is impossible to place every such eventuality in a “black box” so that the AI can look it up and find out how to react. But contextual AI approaches decision-making from a new direction, basically empowering the AI to use data and context to predict and react to situations, even if it has not experienced them before.

This approach and the use of synthetic data combine to negate the need for hundreds of hours of “training” or millions of historical data points both for Tesla and Sprout.ai. The algorithms we use can start to predict straight away.

Coping with infinite variations through synthetic reality creation

In both driving and insurance claims, you never know what’s around the next corner. In order to automate processes like these, we must have an engine that can deal with information inputs when it has little existing data or experience to draw from. This was always a fundamental stumbling block with existing AI technology like Computer Vision and OCR. However, at Sprout.ai, we have leveraged a novel way of leveraging synthetic data to overcome these challenges.

The net result is that our solutions can deal with any piece of information and process it efficiently, even if it is something it has never seen before. New formats, character based languages and handwritten information are mere bumps in the road that can be analysed, understood and processed in minutes or at most hours.

Conceptually, Elon Musk and his team face very similar challenges to the ones we have had to overcome. And just like Sprout.ai, Tesla is a trailblazer in its industry. There are certainly more similarities than differences between the two companies, which are both pushing the envelope in exploring new ways that humans and machines can work in harmony.

Posted on

Building transparency and trust in AI

In December 2020, the White House issued an Executive Order on the use of “trustworthy AI” across federal government agencies. It serves as the biggest indicator yet of the growing public awareness and concern regarding what we term ethical AI.

The role of AI in various industries, as well as government agencies, is becoming more significant with every passing month. Yet while public opinion is still so strongly influenced by five decades of science fiction, it is understandable that trust issues and a suspicion of the unknown can be causes for concern. Some of these were outlined by Tech Crunch in a piece they published at the end of last year.

Sprout.ai is one of the first companies in our industry to have a dedicated AI ethics lead. 

Just as we have a dedicated data protection officer to oversee compliance with GDPR regulations, we take our current and future AI accountabilities equally seriously. Here, we explore what that really means and also address some of the concerns raised in the Tech Crunch article.

Coming out of the black box

The problem is that previous generations of AI have been built around deep learning black box processes. The algorithm takes millions of data points, correlates specific features about them and draws conclusions. This leads to inevitable concerns that there is the potential for bias. Human nature is inherently suspicious of black box decisions and users are no longer prepared to blindly accept that the AI knows what it is doing without better transparency.

The technological solution lies in new data pipelines of the type currently being developed here at Sprout.ai. Conceptually, these work in much the same way as a relationship database. Different data nodes are all interlinked, and they can be clearly, intuitively and transparently communicated through knowledge graphs.

Bringing AI into the 2020s

For both corporations and government agencies, the real challenge lies in switching from these incumbent AI processes into the new world of causal data pipelines. Black box systems simply do not have this functionality, so the new virtual infrastructure that underlies next generation algorithms needs to be built from scratch.

That’s easily said, but it presents a genuine headache from a practical perspective when businesses are using existing systems on a daily basis. It is a scenario that we have seen time and again over the past 40 years with businesses that have invested heavily in legacy systems that rapidly become obsolete. The reality is that it would take hundreds of engineers months or even years to bring their AI processes into line with new generation AI. 

Contextual AI leading the way

In that sense, it becomes clear that while previous generation AI companies have played an important role in bringing AI technology into the mainstream public arena, they have also taken something of a misstep. The time has come to get AI back on the right path as we negotiate the AI forest, one that is user friendly and transparently ethical, and one that facilitates both supervised and unsupervised learning to deliver the most powerful results.

This philosophy lies at the heart of the Contextual AI solution we have developed at Sprout.ai. As the name intimates, it provides the context behind every recommendation produced by the AI engine. Having that context in place means that every decision can be reviewed and audited by human eyes and can be traced back to its source.

As well as delivering leading edge results, this capability has earned the trust of our users, trust that was in such short supply with previous generation solutions. Insurance claim handlers are confident that they can safely rely on the tool to do the groundwork, leaving them to step in with their intuition and expertise when it matters most.

Even more than that, though, the solution gives them a platform to push back or disagree with the recommendations. In this respect, human and machine are working like any good team. The process of questioning and providing feedback means the AI can take human opinions on board and get better at what it does.

Ethics and accountability

Ethical AI provides another example of how Sprout.ai is striving to be at the forefront of AI innovation, not just in technology development but also in the philosophical considerations that surround it. After all, AI is not intended to supersede humans, but it is does play a vital role in making many of our lives easier, more efficient and more fair.