Inspiration
Throughout my career in SP&O, I noticed a common mindset: companies are often viewed in isolation, as if they exist outside the context of the world. Targets are set to grow quarter over quarter — simply because growth is expected. And when performance shifts, the search for answers stays internal: sales training, marketing spend, headcount changes. But the external world plays a massive role—both positively and negatively. While major events like COVID are obvious, smaller yet impactful shifts often go unnoticed, especially if we think by region, by country, by industry etc. These connections are hard to track and usually understood only in hindsight. That’s where the idea came from: what if AI could surface those external influences in real time, making them visible, understandable, and actionable? How can we help companies make sense of the world around them—not just react with a delay that could cost a fortune in missed revenue?
What it does
It collects relevant global and industry news, structures and analyzes it with Agentforce, and delivers the most pertinent insights to employees—at the right time and through the right channels. It works in three ways: Proactively, it sends Slack notifications when significant events occur. Reactively, it provides insights via an employee agent when users ask questions like "why" or "what could have influenced" specific outcomes. In context, it analyzes recent account news, such as new contracts, mergers, or acquisitions (for a B2B use case). If something noteworthy happens, it notifies the account owner and drafts emails, using these events as a reason to reach out. The system also frames the product value in context.
How we built it
To build this solution, I combined several Salesforce tools (flows, prompts, agents, custom objects, api) to automate the collection, analysis, and delivery of relevant news insights.
- News Custom Object: I created a custom object to capture relevant global and industry news. Raw content is sent to the object from Make.com based on its request to the news API.
- Flow for Knowledge Article Drafting: A flow drafts knowledge articles based on the collected news (stored in the custom object).
- Workflow to Publish Articles: A workflow automates the publishing of the drafted articles.
- Flow for Readable Summaries: I implemented a flow that creates readable summaries of the articles for easy consumption.
- AI Prompt for Content Analysis: Using Einstein AI, I developed prompts that analyze content and return summaries with actionable insights.
- Impact Evaluation: I created flows and prompts to assess how news might impact the company's performance—both negatively and positively. They analyse news for negative and positive impact, and then create a relative score for each.
- Slack Notifications: Flows were built to trigger Slack alerts if news items exceed a specific impact threshold, ensuring timely awareness.
- Employee Agent: An agent was designed to provide real-time insights, answering questions like "Why did this happen?" or "What influenced this result?". This agent works based on a custom action. The custom action is powered by a flow that loops through recent knowledge articles, and a prompt that analyzes them and returns insights to the user.
- Account-Level Insights: Make.com is triggered to request news on a specific account. The returned content is analyzed by a prompt, which creates a summary field. Based on this summary field, sales can draft personalized emails, such as 'Congratulations on winning the new contract! I thought it might be the right time to...'
Challenges we ran into
One challenge was ensuring the AI could accurately identify news that had a direct impact on business performance, especially when considering how rapidly external factors change.
Accomplishments that we're proud of
It works! I successfully built a system that automates the process of collecting, analyzing, and delivering news insights in real-time, that can be used by a real org.
What we learned
How to use Agentforce ( :
What's next for Briefing Bruce
Refine news collection to focus on more relevant, less redundant information. Optimize prompt use and storage by reducing repetition (+recycling and removing redundant and duplicated information, but keeping the track of records) and improving context clarity. Expand Briefing Bruce’s ability to process diverse news sources, enhance contextual analysis for greater accuracy, and integrate customization based on user feedback. The goal is to provide precise, real-time insights that enable faster, data-driven decisions. Develop new use cases, such as building TAL based on global context, dynamic relative scoring that reacts on what's going on in the world, or planning marketing communications.
Built With
- api
- flow
- make.com
- newsapi
Log in or sign up for Devpost to join the conversation.