Wah Janney - A Sales Agent

• 

  Can you analyze the lead conversion rates by source

• 

  Find Number of Products Sold in Each Opportunity

• 

  For each Opportunity, give me the number of products sold, the total revenue and the max quantity sold from its OpportunityLineltems.png

• 

  For each Opportunity, give me the number of products sold, the total revenue, and the max quantity sold from its OpportunityLineltems

• 

  Get all the products with high engagement but low conversion

• 

  get products which have high conversion rate

• 

  Get the Best Selling Products within last month

• 

  Get the highest revenue product of this Quarter

• 

  Get the top 5 Best Selling Products

• 

  Give most demanded products within last month

Inspiration

We observed that sales teams spend excessive time sifting through reports to identify which products attract the most interest and which lead to actual orders. Combining opportunity and order data to assess product performance and lead conversion is time-consuming. Standard agents provide only surface-level summaries and lack detailed product-level insights. This inspired us to build Janny, a solution that automates end-to-end analysis, highlights high-engagement but low-conversion opportunities, and delivers actionable insights instantly, empowering sales teams to focus on closing deals instead of crunching data.

What it does

Janny automatically analyzes opportunity and order data to provide real-time product-level insights. It identifies items that attract attention but don’t convert, enabling teams to adjust strategies quickly. It delivers clear, actionable insights, streamlines sales performance analysis, and helps reps prioritize high-impact opportunities.

How we built it

We developed Janny using Salesforce Agentforce. We leveraged out-of-the-box templates to train the agent and created new agents by directly describing functionality to the AI. We explored all standard topics and actions within Agentforce and defined custom topics with simple descriptions, enabling tailored agent behavior without complex configurations.

Challenges we ran into

Conflicting Topics: Initially, multiple topic definitions overlapped, leading to inconsistent agent responses. We resolved this by narrowing the scope and refining the topic definitions.

Unreliable ID Hiding: Suppressing record IDs was inconsistent. Although discretion-based prompts helped, some IDs still slipped through. We implemented explicit logic to ensure ID suppression was consistently applied.

Accomplishments that we're proud of

Successfully built a tailored AI agent that provides actionable product-level insights. Automated what used to be a manual and time-consuming reporting process. Gained mastery over Agentforce’s capabilities, both standard and custom.

What we learned

Training and deploying Agentforce agents using templates and direct AI instruction. Explored and understood all standard topics and actions available in Agentforce. Learned how to define and fine-tune custom topics by describing them to the AI. Gained insights into how to adapt AI behavior for specific business requirements without extensive technical configuration.

What's next for Wah Janney - A Sales Agent

Expand Janny to include competitive analysis and pricing trends. Integrate with Marketing Cloud to align sales insights with campaign performance. Add proactive alerting for products with high views but zero conversions. Enable voice or chat-based interaction for field sales teams on mobile.

Built With

• agent-topics
• agentforce
• salesforce
• salesforce-platform