Achieving 5x Faster Speed-to-Market for AI Initiatives within Salesforce CRM at a Fortune 100 P&C Insurer

The Customer is a Fortune 100 Property & Casualty (P&C) insurer. The Analytics & AI organization is responsible for running all AI initiatives, including building models and delivering business improvement. These AI initiatives are both business-facing (models for use by their teams of underwriters, sales executives, and agencies) and consumer-facing.  

The Customer’s Analytics & AI team faced significant delays and inefficiencies in getting the models ‘off the shelf’ and putting them to work for the business. 

Salesforce CRM is extensively used by the Customer’s sales, account management, and underwriting teams. 

Key Challenges: 

1. Delayed Time-to-Value and CRM Integration for AI Models: The process of integrating AI models into daily workflows within Salesforce needed hundreds of hours of work for every AI initiative. It required building extensive data pipelines, integrations with Salesforce, and developing the right user experience for AI within the business user’s workflow.  
2. Limited End User Feedback: The team struggled to evaluate model effectiveness in real-world scenarios. There was no standardized way to evaluate what was working and what was not, making it difficult to iterate based on business user feedback and pilot data before wider rollouts. 
3. Broken Prioritization: Getting AI insights into the hands of end-users within their daily workflows proved challenging. The team spent considerable time deliberating on what to build, how to prioritize projects, and creating business cases for AI initiatives. 

Solution:  

The challenge we set out to address with a solution was quickly and iteratively bringing insights from models that were “on-the-shelf" to the end business users they were designed for.  

Integrating AI model data rapidly into Salesforce 

AI Squared's pre-built connectors allowed the team to connect to both, the model batch results on AWS S3. The model results were post-processed by mapping them to associate data available on AWS Redshift. These steps were orchestrated through the AI Squared platform. 

The platform enabled the team to leverage the out-of-the-box integration with Salesforce CRM, which provided the capability to embed Data Apps with insights from the model, into Salesforce. 

This significantly accelerated the speed at which the customer integrates AI projects into Salesforce. 

Making AI easily consumable in the context of the business user’s workflow 

When the business user is viewing an Account on Salesforce, their objective is either to edit information or to gain the information and insights that would allow them to make an effective decision. In the absence of AI Squared’s integration into the Salesforce workflow, the user was isolated from AI.  

Now with AI Squared’s Data Apps, the user is armed with contextual insights at the point of decision-making, within the Account object on Salesforce.  

For example, now an underwriter can access AI risk score charts for an account on Salesforce. An advisor can make better decisions on what products to sell to a specific customer through AI product recommendations provided via AI Squared’s Data Apps embedded in Salesforce. 

Capture & analyze end-user feedback from pilot users 

AI Squared’s Feedback capabilities within Data Apps also enable the team with a standardized way to measure the effectiveness of the AI initiatives and also allow the team to prompt and capture user feedback on the performance and usage of the AI model. With this, the AI/Analytics team was able to test the AI model results with a select set of business users. By measuring model effectiveness and model usage performance, the team was able to understand what’s working and what’s not. 

With access to insights on model effectiveness, the team is able to prioritize AI initiatives that demonstrate the most business value. 

Ensure compliance with data security standards and privacy regulations 

AI Squared implemented strict access controls and permissions to safeguard sensitive data while securely accessing and syncing data from AWS to Salesforce. The end-to-end solution was deployed within the customer’s VPC to comply with the customer’s need for enhanced data security.   

Impact Delivered:  

1. 90% reduction in integration time: Slashed post-processing and integration time by 90%, reducing hundreds of development hours to just 25-30 hours 
2. Accelerating the time-to-value of AI models by 85%: Reduced the overall AI initiative evaluation timeline from 6+ months to under 30 days 
3. Faster speed-to-market for the highest impact AI initiatives: Enabled business and data teams to self-serve their AI implementation needs. Streamlined the feedback loop and iteration process, allowing teams to rapidly adjust and optimize AI solutions without technical bottlenecks 

Conclusion 

AI Squared significantly accelerated the time-to-market for AI initiatives at this Fortune 100 Insurance company.  Through rapid prototyping and iterations using AI Squared, and a seamless integration with Salesforce, the AI & Analytics team accelerated both the adoption and effectiveness of their AI models.