Learning Library

← Back to Library

IBM Cloud Packs Transform Auto Claims

Key Points

  • Insurance auto‑claims processing is currently slow, costly, and error‑prone, leading to high payouts, poor customer experiences, and pressure from disruptive tech‑focused insurers.
  • IBM’s Cloud Packs provide a flexible, modern application platform that enables insurers to transform legacy claim‑management systems into data‑driven, automated workflows.
  • An intelligent claims workflow can automate roughly 75 % of claim steps, using chatbots for data collection, AI models for fraud evaluation and complexity scoring, and automated task routing.
  • In a demo, a policyholder initiates a claim via a mobile app, the system automatically populates the claim, presents repair options, and delivers a settlement offer, with only the minimal human interactions highlighted in red.
  • The solution integrates with the insurer’s existing data‑center infrastructure, preserving mission‑critical legacy investments while adding the flexibility needed for rapid modernization.

Full Transcript

# IBM Cloud Packs Transform Auto Claims **Source:** [https://www.youtube.com/watch?v=bGwy98ra_E4](https://www.youtube.com/watch?v=bGwy98ra_E4) **Duration:** 00:05:57 ## Summary - Insurance auto‑claims processing is currently slow, costly, and error‑prone, leading to high payouts, poor customer experiences, and pressure from disruptive tech‑focused insurers. - IBM’s Cloud Packs provide a flexible, modern application platform that enables insurers to transform legacy claim‑management systems into data‑driven, automated workflows. - An intelligent claims workflow can automate roughly 75 % of claim steps, using chatbots for data collection, AI models for fraud evaluation and complexity scoring, and automated task routing. - In a demo, a policyholder initiates a claim via a mobile app, the system automatically populates the claim, presents repair options, and delivers a settlement offer, with only the minimal human interactions highlighted in red. - The solution integrates with the insurer’s existing data‑center infrastructure, preserving mission‑critical legacy investments while adding the flexibility needed for rapid modernization. ## Sections - [00:00:00](https://www.youtube.com/watch?v=bGwy98ra_E4&t=0s) **IBM Cloud Packs Transform Auto Claims** - The demo shows how IBM's Cloud Packs enable insurers to automate 75% of auto claim processing, using a data‑driven workflow, chatbot, and mobile app to speed resolutions and improve customer experience. ## Full Transcript
0:00i'm dirk deroos lead cloud pack demo 0:02architect 0:03we're going to take a look at how ibm's 0:05cloud packs can help insurance companies 0:06modernize their approach to automotive 0:08claims processing 0:09today processing auto insurance claims 0:12is inefficient slow 0:13and costly the claims adjustment 0:14workflow is filled with manual tasks 0:16and this is a huge factor in keeping the 0:18average time to resolution for claims 0:20high 0:21also adjusters under time pressure need 0:23to make judgment calls based on limited 0:25information 0:26this results in many claims having 0:27higher than needed adjustment payouts 0:29these delays and error-prone processes 0:32lead to poor customer experiences 0:34at the same time traditional insurers 0:36are under pressure from insurer tech 0:37startups who are disrupting the auto 0:39insurance space 0:40so to help insurers ibm has developed an 0:42intelligent workflow for processing auto 0:44claims 0:45at the center of this are ibm's cloud 0:47packs which enable you to build flexible 0:49modern applications this represents a 0:51transformative change in the way 0:52insurers handle claims 0:54so basically we're setting a data driven 0:56approach to support decision making 0:58and we're automating 75 percent of the 1:00claims processing steps 1:01we have a simulated insurance company 1:03that has provided its customers with 1:04this mobile app 1:05caitlyn smithers has just been in an 1:07accident where someone rear-ended her 1:09car so she's starting this claim from 1:10her phone 1:11we can see a chat bot engage with her 1:13and gather information as the claim 1:14progresses 1:15on the agent dashboard we can now see a 1:17new claim being initiated 1:19the app actually pushes a request to 1:21open the claim in the existing claim 1:22management application 1:24sitting in the data center there's no 1:26agent driving this process 1:27all the blue dots to the right of the 1:29tasks you can see here represent 1:31fully automated work and the red dots 1:33show tasks that a human needs to do 1:35for example where we see caitlyn 1:36responding to requests for information 1:38as she does this we can see the 1:40dashboard light up with data points and 1:41get fleshed out 1:43we can also see green dots next to some 1:45tasks as well 1:46like invoke fraud investigation these 1:48are tasks performed using ai 1:50capabilities 1:51so here the system feeds information 1:53about the claim to a fraud evaluation 1:55model 1:55and as a more complete picture of the 1:57claim emerges the models actually 1:58predict 1:59a low likelihood of fraud and that this 2:01claim has low complexity 2:02based on this evaluation the insurer has 2:04now authorized the repairs 2:06and presents caitlyn with repair choices 2:08towing and transportation from the 2:10accident site 2:11and back in the app caitlyn is presented 2:13with a settlement offer which she can 2:14then accept 2:15so let's look at what's behind these 2:16apps in the intelligent claims workflow 2:18we're going to start with the insurance 2:20companies data center now for any 2:22established insurance company 2:23their existing it infrastructure is 2:25mission critical it's got decades of use 2:27investment and keeps their business 2:28running every day there are always needs 2:30for increased flexibility 2:32but there's very little appetite to 2:33build this in because migrating from 2:35that established infrastructure 2:36is expensive and disruptive as a 2:39starting point for the solution we just 2:41saw 2:41we're going to leave the established 2:42claims management application and any 2:44supporting software like databases in 2:46place 2:48next up is the cloud pack for business 2:50automation one of its major components 2:53is the business automation workflow 2:54tooling what we're looking at here 2:56is a detailed graph of over 70 tasks 2:59that make up the claims adjustment 3:00process 3:01a business process analyst and this 3:03isn't a coder has built this out as an 3:05automated workflow 3:06and in the app that we just saw it's 3:08actually this workflow that drives the 3:09whole automated claims adjustment 3:11process 3:13so now we're looking at cloud pack for 3:14data which is an ecosystem for everyone 3:16whether you're a data scientist or a 3:18business user 3:19to work together on data projects this 3:21is one of the tools 3:22auto ai which is training our fraud 3:24evaluation model 3:25a really important design point that's 3:27baked into this tooling is trust 3:29your data stewards can make sensitive 3:31data available for data scientists to do 3:33their work and at the same time 3:34define policies that actually hides 3:36customers private information 3:38and trust doesn't just apply to data it 3:40also has to apply to machine learning 3:41models 3:41we saw in the agent dashboard for 3:43example that caitlin's claim had a low 3:45likelihood for fraud 3:46and with club pack for data's model 3:48explainability we can show the claims 3:49adjuster the reasons behind that 3:51and now we're looking at watson 3:52assistant which powers that chatbot that 3:54we saw engaging with caitlin 3:55it integrates with our automated 3:57workflow giving caitlin an easy customer 3:59experience 4:00which ultimately helps her deal quickly 4:01with the accident 4:03cloudpack for integration is the glue 4:04that ties all these pieces together 4:06integration developers use appconnect 4:08api connect and kafka flows to enable 4:10the mobile app and the claims dashboard 4:12to talk with the claims management 4:14application in the backend data center 4:16and finally application developers have 4:19built a set of cloud native 4:20microservices that use all the 4:21capabilities that we've looked at 4:22to power the mobile app and the claim 4:24agent dashboard 4:26everything in this architecture diagram 4:28except the original infrastructure in 4:30the data center 4:30is running in openshift here i'm looking 4:33at the openshift container service 4:34i can see all the pods that are powering 4:36the cloud packs and also the new claims 4:38processing microservices for the 4:40dashboard and the mobile app this 4:42container platform has a common 4:43interface for devops 4:44and administration of this whole 4:46solution since it's based on openshift i 4:48can deploy and run this anywhere i want 4:49it can be in the same data center as the 4:51existing infrastructure 4:53i can be on ibm's cloud where i'm 4:54running it now or some other vendors 4:56cloud 4:57and what's worth calling out here is 4:58that while we've greatly modernized the 5:00claims processing practice 5:01we've kept the original systems intact 5:03this saves us money 5:05time and reduces risk and disruption 5:08what we've built here is an almost 5:10entirely automated process 5:12and this frees customer service agents 5:13and claims adjusters to focus on higher 5:15value work 5:16the insurance company provides a faster 5:18and more flexible clean process for 5:20their customers 5:21and also embeds machine learning models 5:23in the workflow that helps the claims 5:24adjusters 5:2525 of adjustment tasks are aided with 5:28these models 5:29this means adjusters can make much 5:30better decisions for difficult areas 5:32like 5:33assessing total loss amounts and the 5:34likelihood of liability 5:36by taking advantage of their existing 5:38customer base and wealth of data 5:39we've used ibm's cloud packs to help 5:41traditional insurance companies set a 5:43solid competitive advantage 5:44over new companies encroaching in this 5:50market 5:55you