Skip to main content

Sales Assistant with Agents

The Generative AI Sales Assistant project was created to bridge the gap between technology and business teams and promote the adoption of innovative solutions. The vehicle sales sector continually faces challenges in onboarding new employees, requiring significant time and resources. The proposed solution seeks to streamline the onboarding process by utilizing a generative AI agent to offer immediate and personalized support to new employees, providing them instant information and guidance.

This example illustrates how a team developed an agent in StackSpot AI to serve as a virtual assistant for new employees in the vehicle sales sector. The Agent aims to:

  • Save time and resources by reducing the burden on veteran employees;
  • Ensure consistency in the information and guidelines provided;
  • Offer personalized support, adapting to the specific needs of each employee.

Step-by-step implementation of the generative AI Sales Assistant

Step 1. Create a Knowledge Source (KS)

The first step in development was creating a knowledge base using Knowledge Sources, which includes:

  • Complete operational context: Information about the vehicle sales sector, including guidelines, policies, and responsibilities of new employees;
  • Daily procedures: Detailed instructions on the tasks and processes that new employees must follow;
  • Frequently Asked Questions (FAQs): Answers to the most common questions from new employees.

Tips for creating a Knowledge Source

  1. Gather information: All vehicle sales sector guidelines, policies, and procedures were compiled into a file.
  2. Organize do conteúdo: Structure the content clearly and accessibly, categorizing the information by themes (e.g., sales, customer service, internal processes).
  3. Add the KS to the Agent: After creating a Knowledge Source, add it to the Agent.

Step 2. Create and Configure the Agent

After establishing the knowledge base, the next step involved developing the StackSpot AI Agent. The Agent was designed to emulate an experienced, amiable, professional mentor and offer tailored support to new employees.

Agent Configuration

In the system prompt section, the following was added:

  1. Persona definition: The Agent was configured to adopt the persona of an experienced mentor with a welcoming, clear, and motivating tone.
  2. Context explanation: The Agent needs to understand the vehicle sales sector scenario and the expectations for new employees.
  3. Task description: The Agent should answer questions, provide guidance, and assist with initial work processes.
  4. The tone of voice: A tone of voice was established to be welcoming and motivating for the Agent's responses.

After that:

  1. Link the Knowledge Source: Integrate the knowledge base with the Agent, ensuring it has access to the necessary information to answer employee questions.
  2. Test the integration: Perform tests to ensure the Agent meets your expectations.

Step 3. Validation and Curation

After configuring and integrating the Agent, it was made available to sales sector specialists for testing and feedback on adjustments and improvements.

  1. Test with specialists: Make the Agent available to sales sector specialists and conduct simulated interactions.

  2. Collect feedback: Gather feedback from specialists on the accuracy of responses, appropriateness of the tone of voice, and usability of the Agent.

  3. Adjustments and improvements: Based on the feedback, adjustments were made to the Agent and the knowledge base to improve the user experience.

Results

  • Comparison between LLM Tools The Agent was tested on three different LLM tools, using the same prompts to ensure a fair comparison. StackSpot AI performed better, offering more profound, cohesive responses aligned with business needs.

See the benefits of using StackSpot AI:

  • Accuracy in responses: The Agent effectively used the knowledge base, providing precise information.
  • Adherence to the defined Tone and Persona: Interactions remained consistent with the established tone of voice and persona.
  • Improvement in Onboarding: New employees who used the Agent reported a smoother and more satisfying onboarding experience.

Agents Usage

After conducting the tests, the team has decided to follow these steps for the use of Agents:

  • Expansion of use: The developed Agent is currently being used by part of the sales team. The plan is to expand its use across the entire company, allowing more teams to benefit from the solution.

  • Product: The next step is to proceed to the a product phase by incorporating the Agent into the company's official onboarding processes. This includes integration of the Agent with other internal tools and the creation of metrics to monitor its performance.

The Generative AI Sales Agent project has demonstrated that implementing a StackSpot AI Agent can significantly improve Onboarding, optimize resources, and enhance new employee experiences.