How I Solved It

Provide Career Guidance With Agentforce and Data Cloud | How I Solved It

By

Welcome to another “How I Solved It.” In this series, we do a deep dive into a specific business problem and share how one Awesome Admin chose to solve it. Once you learn how they solved their specific problem, you’ll be inspired to try their solution yourself! Watch how Team Agent 12 implemented an Agentforce solution powered by RAG and Data Cloud for enhanced job matching and personalized support for job seekers in the non-profit sector.


Key business problem

12 Beacons, a fictional nonprofit employment services platform, needs to address a critical business challenge: How can they improve the accuracy of career path prediction and guidance to ensure job seekers are consistently placed in roles that align with their skills, values, and long-term career goals, resulting in demonstrable impact?

Sarah Miller is one such job seeker. Passionate about environmental work, she finds herself navigating job placement, salary preferences, and interview preparation. 12 Beacons wants to provide personalized and effective support for Sarah and other job seekers, and they need a tool that streamlines these processes.

How I solved it

Our team, Agent 12, comprised of Rosie Goldman, Lucile Vea, Douglas Furey, Charlie Haws, and myself, Mike Fazio, hails from New York and Georgia. Rosie and I connected through the New York Salesforce User Group meetings. We met Lucile at the New York AI NOW Tour, and we first encountered Charlie and Doug at the NYC Hackathon, where they joined our team. Together, we created a service agent called Employment Services. This agent lives on our 12 Beacons Experience site for users like Sarah Miller to help find and apply for nonprofit jobs that match their interests and requirements. Let’s dive in!

The agent greets Sarah, who expresses her interest in applying for a job in environmental education. The agent, following the instructions under the Job_Placement_Assistance topic, asks Sarah for additional details to narrow down her search. These details include her job preferences such as job category or industry, location, work schedule (full-time, part-time, etc.), salary range, and relevant skills or experience.

Agent Builder showing a section for selecting a topic, with ‘Job_Placement_Assistance’ highlighted. It includes instructions for the agent to search for job openings based on user preferences like job title, industry, location, and more.

Sarah provides her job preferences: “Full time in Seattle, around $120K, with skills related to policy and research.” The agent selects the Job_Placement_Assistance topic and invokes the prompt template action called Find Job Openings. Unfortunately, the search does not yield any results, and the agent responds, “It looks like there aren't any current job openings that match your criteria. Would you like to adjust any of your preferences, or perhaps set up job alerts for new openings that match your criteria?”

Sarah then asks, “What about a range from $50-90K?” With her revised salary requirements, the agent is able to provide some job openings that meet her criteria.

Agent Builder focused on the Job_Placement_Assistance topic and the Find Job Options prompt template agent action.

The flex prompt template called Find Job Openings provides specific instructions, search parameters, and output fields. It utilizes Einstein Search with Retriever-Augmented Generation (RAG) for job openings. RAG enhances generative artificial intelligence (AI) capabilities by leveraging the most current and relevant proprietary data. This involves retrieving information from a knowledge library, including both structured and unstructured content such as emails, PDFs, chat logs, and social media posts.

The retrieved data augments the prompt, ensuring it generates a response based on the most appropriate and up-to-date information. This method helps deliver higher-quality, more relevant, and accurate AI outputs. By setting up an Agentforce Data Library and using the Answer Questions With Knowledge standard agent action, we can dynamically retrieve and augment prompts with relevant business knowledge, effectively implementing RAG with our agent.

Prompt Builder displaying instructions for analyzing job openings, sections for context, resolution, and response, and a configuration panel with search parameters.

Sarah sees a job she’s interested in and asks the agent to apply for the position. The agent responds, “Could you please provide your first and last name so I can proceed with the application for the Environmental Policy Analyst position?” After Sarah inputs her name, the agent, following the action under the Job_Placement_Assistance topic, applies to the position on her behalf.

Agent Builder showing the selected topic ‘Job_Placement_Assistance’, and ‘Apply for Job’ prompt template agent action.

When Sarah decided to apply for the job, the agent invoked the custom flow agent called Hackathon_Apply for Job. A few things about this flow:

  1. It gets the job candidates from Data Cloud using the input variables first name and last name from the agent.
  2. Then, we get the job opening from Data Cloud where the job title equals the (as an input variable) job title from the agent.
  3. Next, we create a job application record in Salesforce using the job candidate found in the first Get Records element and the job opening found in the second Get Records element.
  4. Lastly, we set a confirmation message made available to the agent (as an output variable) using an Assignment element.

Flow titled ‘Hackathon_Apply for Job - V5’. The flow outlines a process with the following steps: Autolaunched Flow (Start), Get Job Candidate, Get Job Opening, Create Application, SetMessage (Assignment), and End.]

The agent successfully submits Sarah's application and responds, “Your application has been submitted, Sarah! You will receive an email confirmation shortly. Is there anything else I can help you with?”

Recognizing the importance of interview preparation, Sarah asks, “What guidance can you provide for preparing for an interview?”

Our Employment Services agent has another topic, Career_Counseling, designed to help job seekers identify career goals and interests, and provide guidance on career paths and industry trends. Under this topic, an instruction is set up: ‘If the user asks for help with skills, interviewing, fundraising, grant writing, resume preparation or tips, attempt to answer from the knowledge base.’

The agent invokes the standard action labeled ‘Answer Questions With Knowledge’, searches the knowledge base, and provides several tips for Sarah’s interview.

Agent Builder for the Employment Services agent with the Planner showing the Career_Counseling topic selected and invoked Answer Questions With Knowledge standard action.

To provide those tips, the agent searches our knowledge base, which includes uploaded and indexed PDFs, ensuring that the user receives the correct information.

Agent Builder for the Employment Services agent, showing the Knowledge section where the PDFs are housed that are referenced in the Employment Services Library.

Business results

Building this agent helps job seekers in the nonprofit sector easily search and apply for jobs, receive career counseling, and get tips for interview preparation. Our solution utilizes Data Cloud, Agentforce, prompt templates, Agentforce Data Libraries, RAG, Experience Builder, and flows—all with no code!

Do try this at home

Trailhead! Trailhead! Trailhead! This has been so helpful for me to learn and build new skills, including the basics. The Configure Agentforce for Service project is a great resource to learn more about building agents.

Resources

Want to see more good stuff? Subscribe to our channel!

SUBSCRIBE TODAY