Perplexity for Enterprise Insights
A unified Streamlit dashboard that brings Salesforce deal data, public financials, and AI-driven market research into one place—slashing sales-meeting prep time and lowering cost-of-sale.
Inspiration
With six years as a Data Engineer at Workday, I saw firsthand how account executives spend hours cobbling together CRM reports, financial charts, and industry news before each pitch. I asked myself: why not let an AI do the heavy lifting? By tapping into Salesforce’s API, Yahoo Finance, and Perplexity’s Sonar AI, you can serve all the insights sales teams need—instantly.
What it does
- 📊 Opportunity Viewer: Live deals from Salesforce, with risk scoring, next-step guidance, and AI-powered industry trends & pain points.
- 🏢 Company Overview: Public financial metrics plus executive summary, with deep-dive AI briefings.
- 🔬 AI Research Hub: On-demand news trends, industry risks, competitor comparisons—trusted citations included.
- 🧭 Strategic Analysis: Auto-generate SWOT, Porter’s, PESTEL, VRIO frameworks with Pro-level reasoning.
- 🤝 Meeting Prep Generator: One-click agendas, talk tracks, and tailored questions by account, opportunity, and audience.
How we built it
- Streamlit for rapid UI development and state management across tabs.
- Salesforce REST API (via Simple-Salesforce) to pull Accounts, Opportunities, Contacts, Tasks, and Events.
- YahooQuery for live ticker data, price history, financials, and key stats.
- Perplexity Sonar API to power all AI prompts—real-time Internet search, chain-of-thought reasoning, and deep research.
- Session State to “remember” selected Account, Opportunity, and Ticker across all components.
- Clearbit Logo API for on-the-fly customer logos.
- Custom CSS/HTML for branded green tables, centered layouts, and downloadable CSV/PDF assets.
Challenges we ran into
- State Persistence: Coordinating selections (Account, Opportunity, Ticker) so every tab “knows” the context.
- API Reliability: Handling intermittent Salesforce and Yahoo Finance errors—added caching & fallback logic.
- Layout & Styling: Streamlit’s limited nesting rules required creative use of
st.markdownwith inline CSS. - Citation Tracking: Ensuring Sonar AI returns clickable sources for research outputs.
Accomplishments that we’re proud of
- Unified Experience: All data and AI insights in one place—no more bouncing between CRM, spreadsheets, and news sites.
- Time Savings: Sales teams can generate a complete meeting deck in seconds, not hours.
- Developer Friendly: Modular components and clear prompts make it easy to extend to other CRMs or data sources.
- AI Integration: Real-time Internet search, trusted citations, and chain-of-thought reasoning via Perplexity Sonar.
- External Industry Specific Reports: That enhances the experience and knowledge
What we learned
- Effective Streamlit session management is key for multi-page apps.
- How to cache external API calls without stale data.
- Best practices for error handling and user feedback with Salesforce and Yahoo APIs.
- Crafting business-quality AI prompts that guide Perplexity Sonar to deliver executive-grade content.
- Git - my first time with using Git so often and so "Pro" haha
What’s next for Perplexity for Enterprise Insights
- Multi-CRM Support: Add connectors for HubSpot, Dynamics, and custom in-house CRMs.
- User Authentication & Roles: Secure per-user data access and team-based dashboards.
- Custom Data Sources: Plug in internal product catalogs, collateral libraries, and knowledge bases.
- Analytics & Tracking: Measure usage patterns, AI prompt performance, and deal-outcome correlations.
- Mobile Optimization: Responsive design for field sellers on phones and tablets.
- *Slide Generation Feature *: To allow users to generate pptx slides with app outcomes
Built with ❤️ at Perplexity Hackathon by Kornel Pudło—empowering sales teams with AI-driven insights.
Built With
- pandas
- python
- requests
- salesforce
- sonar
- streamlit
- yahooquery
Log in or sign up for Devpost to join the conversation.