Vxceed builds the proper gross sales pitch for gross sales groups at scale utilizing Amazon Bedrock

This submit was co-written with Cyril Ovely from Vxceed.

Shopper packaged items (CPG) firms face a crucial problem in rising economies: how one can successfully retain income and develop buyer loyalty at scale. Though these firms make investments 15–20% of their income in commerce promotions and retailer loyalty applications, the uptake of those applications has traditionally remained under 30% resulting from their complexity and the problem of addressing particular person retailer wants.

Vxceed’s Lighthouse platform tackles this problem with its progressive loyalty module. Trusted by main international CPG manufacturers throughout rising economies in Southeast Asia, Africa, and the Center East, Lighthouse supplies area gross sales groups with a cutting-edge, AI-driven toolkit. This resolution makes use of generative AI to create customized gross sales pitches primarily based on particular person retailer information and developments, serving to area representatives successfully interact retailers, handle frequent objections, and enhance program adoption.

On this submit, we present how Vxceed used Amazon Bedrock to develop this AI-powered multi-agent resolution that generates customized gross sales pitches for area gross sales groups at scale.

The problem: Fixing a income retention drawback for manufacturers

Vxceed operates principally within the rising economies. The CPG trade is going through challenges comparable to fixed change, excessive buyer expectations, and low boundaries to entry. These challenges are extra pronounced within the rising economies. To fight these challenges, CPG firms worldwide make investments 15–20% of their income yearly in commerce promotions, typically within the format of loyalty applications to retailers.

The uptake of those loyalty applications, nonetheless, has historically been decrease than 30% resulting from their complexity and the necessity to handle every particular person outlet’s wants. To make this problem extra complicated, in rising economies, these loyalty applications are primarily bought by way of the sphere gross sales group, who additionally act within the position of order seize and fulfilment, and the size of their operation typically spans throughout hundreds of thousands of retailers. To uplift the loyalty applications uptake, which in flip uplifts the manufacturers income retention, the loyalty applications wanted to be tailor-made at a personalised stage and pitched correctly to every outlet.

Vxceed wanted an answer to resolve this drawback at scale, creating distinctive, customized loyalty program promoting tales tailor-made for every particular person outlet that the sphere gross sales group can use to promote the applications.

This problem led Vxceed to make use of Amazon Bedrock, a totally managed service that provides a selection of high-performing basis fashions (FMs) from main AI firms by way of a single API.

Resolution overview

To deal with the challenges of personalization, scale, and placing the answer within the palms of tens of 1000’s of area gross sales groups, Vxceed developed Lighthouse Loyalty Promoting Story, an AI-powered resolution. The Lighthouse Loyalty Promoting Story structure makes use of Amazon Bedrock, Amazon API Gateway, Amazon DynamoDB, and AWS Lambda to create a safe, scalable, AI-powered promoting story technology system. The answer implements a multi-agent structure, proven within the following determine, the place every element operates throughout the buyer’s personal AWS setting, sustaining information safety, scalability, and intuitive person interactions.

The answer structure is constructed round a number of key parts that work collectively to supply a curated gross sales enablement expertise that’s distinctive for every retailer buyer:

• Salesperson app – A cellular utility is utilized by area gross sales groups to entry compelling program gross sales pitches and work together with the system by way of a chat interface. This serves as the first touchpoint for gross sales representatives.
• API Gateway and safety – The answer makes use of the next safety providers:
• Clever brokers – The answer makes use of the next Lambda primarily based brokers:
  • Orchestration Agent coordinates the general move and interplay between parts.
  • Story Framework Agent establishes the narrative construction.
  • Story Generator Agent creates customized content material.
  • Story Evaluate Agent maintains high quality and compliance with model pointers.
  • Model Pointers Agent maintains model consistency.
  • Enterprise Guidelines Agent enforces enterprise logic and constraints.
• Knowledge providers layer – The information providers layer consists of the next parts:
  • Knowledge API providers present entry to crucial enterprise data, together with:
    • Outlet profile information
    • Loyalty program particulars
    • Historic information
    • Buy profile data
  • Integration with Lighthouse synthetic intelligence and machine studying (AI/ML) fashions and information lake for superior analytics.
  • Amazon Bedrock Data Bases for enhanced context and knowledge.
• Superior capabilities – The answer affords the next extra capabilities:
  • Q&A Service permits pure language interactions for gross sales queries.
  • CTA (Name-to-Motion) Service streamlines the retail outlet signup course of.
  • An Amazon Bedrock giant language mannequin (LLM) powers clever responses.
  • Amazon Bedrock Guardrails facilitates acceptable and compliance-aligned interactions.

The structure implements a safe, scalable, and serverless design that makes use of AWS managed providers to ship a complicated gross sales enablement resolution.

Multi-agent AI structure for safe orchestration

Vxceed constructed a multi-agent AI system on Lambda to handle customized gross sales storytelling. The structure contains specialised brokers that work collectively to create, validate, and ship compelling gross sales pitches whereas sustaining alignment with enterprise guidelines and model pointers.

The next is an in depth breakdown of the multi-agent AI structure:

• Orchestration Agent – Coordinates the workflow between brokers and manages the general story creation course of, interfacing with the Amazon Bedrock LLM for clever processing.
• Story Framework Agent – Establishes the narrative construction and move of gross sales pitches primarily based on confirmed storytelling patterns and gross sales methodologies.
• Story Generator Agent – Creates customized content material by combining information from a number of sources, together with outlet profiles, loyalty program particulars, and historic information.
• Story Evaluate Agent – Validates generated content material for accuracy, completeness, and effectiveness earlier than supply to gross sales personnel.
• Model Pointers Agent – Makes positive generated content material adheres to model voice, tone, and visible requirements.
• Enterprise Guidelines Agent – Enforces enterprise logic, buyer model compliance necessities, and operational constraints throughout generated content material.

Every agent is applied as a serverless Lambda perform, enabling scalable and cost-effective processing whereas sustaining strict safety controls by way of integration with AWS KMS and Secrets and techniques Supervisor. The brokers work together with the Amazon Bedrock LLM and guardrails to supply acceptable and accountable AI-generated content material.

Guardrails

Lighthouse makes use of Amazon Bedrock Guardrails to take care of skilled, centered interactions. The system makes use of denied subjects and phrase filters to assist stop unrelated discussions and unprofessional language, ensuring conversations stay centered on buyer wants. These guardrails display out inappropriate content material, set up clear boundaries round delicate subjects, and diplomatically handle aggressive inquiries whereas staying aligned with organizational values.

Why Vxceed selected Amazon Bedrock

Vxceed chosen Amazon Bedrock over different AI options due to 4 key benefits:

• Enterprise-grade safety and privateness – With Amazon Bedrock, you possibly can configure your AI workloads and information so your data stays securely inside your individual digital personal cloud (VPC). This method maintains a non-public, encrypted setting for AI operations, serving to hold information protected and remoted throughout the your VPC. For extra particulars, consult with Safety in Amazon Bedrock.
• Managed providers on AWS – Lighthouse Loyalty Promoting Story runs on Vxceed’s current AWS infrastructure, minimizing integration effort and offering end-to-end management over information and operations utilizing managed providers comparable to Amazon Bedrock.
• Entry to a number of AI fashions – Amazon Bedrock helps varied FMs, so Vxceed can experiment and optimize efficiency throughout completely different use instances. Vxceed makes use of Anthropic’s Claude 3.5 Sonnet for its skill to deal with refined conversational interactions and complicated language processing duties.
• Strong AI improvement instruments – Vxceed accelerated improvement by utilizing Amazon Bedrock Data Bases, immediate engineering libraries, and agent frameworks for environment friendly AI orchestration.

Enterprise affect and future outlook

The implementation delivered important measurable enhancements throughout three key areas.

Enhanced customer support

The answer achieved a 95% response accuracy fee whereas automating 90% of loyalty program-related queries. This automation facilitates constant, correct responses to buyer objections and queries, serving to salespeople and considerably bettering the retailer expertise.

Accelerated income development

Early buyer suggestions and trade evaluation point out program enrollment elevated by 5–15%. This development demonstrates how eradicating friction from the enrollment course of immediately impacts enterprise outcomes.

Improved operational effectivity

The answer delivered substantial operational advantages:

• 20% discount in enrolment processing time
• 10% lower in help time necessities
• Annual financial savings of two person-months per geographical area in administrative overhead

These effectivity good points assist Vxceed prospects concentrate on higher-value actions whereas decreasing operational prices. The mixture of quicker processing and decreased help necessities creates a scalable basis for program development.

Conclusion

AWS partnered with Vxceed to help their AI technique, ensuing within the improvement of Lighthouse Loyalty Promoting Story, an progressive customized gross sales pitch resolution. Utilizing AWS providers together with Amazon Bedrock and Lambda, Vxceed efficiently constructed a safe, AI-powered resolution that creates customized promoting tales at scale for CPG trade area gross sales groups. Trying forward, Vxceed plans to additional refine Lighthouse Loyalty Promoting Story by:

• Optimizing AI inference prices to enhance scalability and cost-effectiveness
• Including a Language Agent to current the generated promoting story within the native language of selection
• Including RAG and GraphRAG to additional improve the story technology effectiveness

With this collaboration, Vxceed goals to considerably enhance CPG trade area gross sales administration, delivering safe, environment friendly, and AI-powered options for CPG firms and types.

If you’re considering implementing the same AI-powered resolution, begin by understanding how one can implement asynchronous AI brokers utilizing Amazon Bedrock. See Creating asynchronous AI brokers with Amazon Bedrock to study in regards to the implementation patterns for multi-agent methods and develop safe, AI-powered options to your group.

Concerning the Authors

	Roger Wang is a Senior Resolution Architect at AWS. He’s a seasoned architect with over 20 years of expertise within the software program trade. He helps New Zealand and international software program and SaaS firms use cutting-edge expertise at AWS to resolve complicated enterprise challenges. Roger is obsessed with bridging the hole between enterprise drivers and technological capabilities, and thrives on facilitating conversations that drive impactful outcomes.
	Deepika Kumar is a Options Architect at AWS. She has over 13 years of expertise within the expertise trade and has helped enterprises and SaaS organizations construct and securely deploy their workloads on the cloud. She is obsessed with utilizing generative AI in a accountable method, whether or not that’s driving product innovation, boosting productiveness, or enhancing buyer experiences.
	Jhalak Modi is a Resolution Architect at AWS, specializing in cloud structure, safety, and AI-driven options. She helps companies use AWS to construct safe, scalable, and progressive options. Obsessed with rising applied sciences, Jhalak actively shares her experience in cloud computing, automation, and accountable AI adoption, empowering organizations to speed up digital transformation and keep forward in a quickly evolving tech panorama.
	Cyril Ovely, CTO and co-founder of Vxceed Software program Options, leads the corporate’s SaaS-based logistics options for CPG manufacturers. With 33 years of expertise, together with 22 years at Vxceed, he beforehand labored in analytical and course of management instrumentation. An engineer by coaching, Cyril architects Vxceed’s SaaS choices and drives innovation from his base in Auckland, New Zealand.