Your Data and Our AI: A Transparent Look at How We Innovate Responsibly

Artificial intelligence (AI) is rapidly transforming how businesses use social to connect with their customers and unlock new levels of efficiency and insight. At Sprout Social, we focus on empowering you through AI, enhancing your experience, and enabling you to achieve your social media goals faster and smarter.

We also understand that transparency is paramount when it comes to your data and AI. You might be wondering, "How exactly is my data used with Sprout's AI features?" It’s a critical question, and we’re here to provide clear answers and build on the trust you place in us.

Understanding AI at Sprout Social

To help you confidently leverage the AI-powered tools within our platform, we want to provide a clear understanding of the technology at work. Let's explore the distinctions in our AI and how we ensure responsible innovation.

The Distinction: Classic ML vs. Generative AI

To understand the landscape of AI we're navigating, it's useful to differentiate between two primary forms. Firstly, there's Classic Machine Learning (ML). This is the engine behind many long-standing intelligent features, including spam detection and sentiment analysis. Such models learn from large, general, and deidentified datasets—not your specific customer information—to identify existing patterns. 

However, the focus of our discussion in this article will be Generative AI. This newer wave of AI is what powers features like AI-driven content generation or summarization tools. It operates by sending prompts, which can include contextual information you supply (like a webpage link or a subject), to foundational models from providers such as AWS Bedrock and OpenAI. It's key to note that this data is used to create your output but does not contribute to the training of these external models. Let's delve deeper into this exciting field of Generative AI. 

Are We Training LLMs With Your Data?

The number one question we hear is:

"Are you using my data to train Large Language Models (LLMs)?"

The short answer is a resounding no.

Your specific customer data is not used to train the 3rd party Large Language Models that power AI Assist features across Sprout Social.

Let's dive deeper into how we approach AI, safeguard your data, and deliver features that empower you.

Our Principled Approach to AI

Our development and implementation of artificial intelligence are guided by core principles that ensure it serves your best interests. We are dedicated to building AI-powered features that are not only innovative but also secure, transparent, and genuinely valuable to your work.

Our Approach: Principled AI for Your Benefit

Our philosophy on AI is built on a foundation of trust, security, and delivering tangible value to you. Our AI-powered platform, informed by extensive historical social data, helps you optimize content with AI-generated captions and trend identification via AI Assist, gain key learnings from performance data with analyst insights, and resolve questions faster with AI-powered replies and summaries, enabling you to manage your social presence with greater impact and efficiency.

Partnering with Industry Leaders for Foundational Models

We utilize leading AI providers like AWS, Anthropic, and OpenAI to access powerful, pre-trained Large Language Models. Critically, our contracts with these partners explicitly state that none of your data or prompts sent to these models are used for their training.  

Prioritizing Data Security with AWS Bedrock

We evaluate all of our AI-powered features for the best combination of performance and security. When possible, we utilize AI providers through AWS Bedrock, a fully managed service offering a choice of high-performing foundation models (FMs) from leading AI companies. This approach offers an added layer of security, as your customer data remains within our secure AWS Virtual Private Cloud (VPC). This means your data doesn’t leave Sprout’s network when interacting with these AI features.  

No Training on Your Specific Data

To be crystal clear, the content you create, the analytics from your social profiles, and any information you input into Sprout Social are not fed into global training sets for the 3rd party foundational AI models we use. Your data is your data. 

Addressing Bias and Fairness in AI Responses

We are deeply committed to ensuring that our AI features generate accurate, helpful, and fair responses. To achieve this, we have a robust evaluation process in place specifically designed to mitigate potential bias and promote fairness when using Large Language Models (LLMs). Every AI-powered feature that leverages LLMs undergoes a thorough evaluation by one of our Applied AI/ML Scientists. This evaluation step meticulously examines the LLM's responses to a diverse range of prompts, ensuring that they are unbiased, equitable, and align with our ethical standards. This dedication to rigorous testing and oversight is integral to our approach to building responsible AI solutions you can trust.

AI Features Designed to Enhance Your Sprout Experience

The AI capabilities within Sprout Social are built to provide you with actionable insights and intelligent automation. We offer a wide array of features designed to enhance your efforts across marketing, customer care, and social insights. For a comprehensive overview of all the AI and automation tools available please visit our support page. These features, and many others, are designed with one core principle: to use AI in a way that makes your job easier and your social media strategy more powerful, all while respecting your data and upholding your trust.

Looking Ahead

The world of AI is constantly evolving, and so are we. We are committed to ongoing investment in Responsible AI, aligning with best practices and standards to ensure our AI development is safe, trustworthy, and positions us as an industry leader you can count on. We will continue to be transparent, listen to your feedback, and ensure our AI innovations always serve your needs responsibly.