OutSystems AI for Apps and Agents
Picture a future where generative AI is the superstar of app development, pulling off magical transformations that'll make your customers’ jaws drop! Imagine apps that can write their own content, chat with users like they've got a PhD in linguistics, and morph to fit your company's every whim. It's like a tech fairytale, isn't it?
Of course, to join this adventure, you've got to integrate generative AI into your apps. If you think that means tackling a puzzling challenge that involves Pytorch, TensorFlow, Jupyter Notebooks, Terraform, OpenCV and more, you can relax. With OutSystems, you can infuse generative artificial intelligence into your apps and connect to large language models (LLMs) with a minimum of effort. LLMs are trained on vast amounts of data underpinned by a set of neural networks that extract meanings from a text sequence and understand the relationships between words and phrases in it.
Last year, OutSystems introduced the ChatGPT and Azure OpenAI connectors developed with Microsoft. And now, we're swinging open the doors to bring you some generative AI magic from Amazon Web Services (AWS). Step inside to learn more.
Digging down to connect to (Amazon) Bedrock
In collaboration with AWS, OutSystems has developed a connector to Amazon Bedrock. Amazon Bedrock is a fully managed service that offers a choice of high-performing foundation models (FMs) and LLMs used for generative AI. Foundation models are trained on a broad spectrum of generalized and unlabeled data. They can understand language, generate text and images, and converse in natural language.
Bedrock addresses use cases such as customer support, content generation, content comparisons, chat and Q&A, and more. You can customize (fine-tune) any foundation model with your own data privately–without writing a single line of code. With the Bedrock connector, which is available in OutSystems Forge, you can easily integrate state-of-the-art AI models from Amazon Bedrock into your applications.
Whether you use Amazon Titan, Claude 2 from Anthropic, or Stable Diffusion, the possibilities are endless. For example, in a recent blog post about how you can use Bedrock with OutSystems, OutSystems MVP Stefan Weber, Senior Director at Telelink Business Services, showed how you can use the OutSystems connector. He accessed Claude 2 and prompted, “Human: Why is OutSystems the leading low-code application platform on the market?”
This is what the prompt and the answer looked like:
Claude 2 explains why OutSystems is the best low-code platform
So, that’s super cool and all, but what else can you do with the connector? Here’s an example from e-commerce.
Tap into the power of generative AI to deliver personalized experiences
In the ever-evolving world of e-commerce, delivering a tailored shopping experience to your customers is key to staying ahead of the game. One area where this can truly make a revenue generating difference is on the order detail page, where customers check their purchase history and status. You can increase their propensity to purchase by not only showing them the list of items they’ve bought and an invitation to keep shopping, but also delivering product recommendations based on their browsing history. But what happens when you have to make recommendations to hundreds of thousands of customers with similar histories?
The challenge: Personalized recommendations at scale
One of your customers, Alessandra, has just purchased a pair of rugged hiking boots and a sturdy backpack for an upcoming adventure in the mountains. Now, it would make perfect sense for her to consider adding hiking sticks to this collection. But how do you make sure that thousands of other customers who’ve made similar purchases are also offered the hiking sticks? Manually crafting them for each customer could be an incredibly daunting task, especially for large e-commerce platforms with numerous daily orders.
The solution: A foundation (model) or two for recommendations
Amazon Bedrock offers an easy-to-use developer experience for working with FMs from leading AI companies like AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and Amazon. For this example, you can use the new OutSystems connector to access Anthropic’s Claude 2 (just like our MVP Stefan Weber) to generate text and Stability AI’s Stable Diffusion XL to generate images based on text. These two models give you everything needed to create personalized experiences powered by generative AI.
Watch the 7-minute deep-dive demo to see how the integration works.
Generative AI, with its ability to process vast amounts of data and generate personalized content, is the perfect tool for creating marketing content that speaks directly to your customers' needs. This content is:
- Scalable: Handling thousands of orders manually can be overwhelming. Generative AI streamlines the process, allowing you to reach every customer with tailored recommendations effortlessly.
- Personalized: Generative AI doesn't just create generic recommendations; it understands individual purchase histories and preferences, ensuring that every message resonates with the recipient.
- Timely: AI can deliver recommendations in real-time, capitalizing on the moment when customers are most engaged with your platform.
- Efficient: Say goodbye to hours of manual work. Generative AI automates the process, freeing up your marketing team to focus on strategy and creativity.
Additional use cases
The e-commerce example is just one of many use cases for this Forge component. Let’s take a look at a few more.
Healthcare Medical report
Build an app or add a capability to a healthcare system that uses generative AI to streamline medical report summarization for accessibility without compromising accuracy, while managing the balance between brevity and critical details.
Legal: Contract review
Use the Amazon Bedrock connector to enable an application to review legal contracts to identify discrepancies and potential risks, all while navigating the complexities and nuances of legal language.
Banking: Personalized financial advice
Offer customized financial guidance to bank customers based on prompts with generative AI. Answer questions about their unique financial goals and risk profiles, and build a financial plan while adhering to stringent financial regulations.
Insurance: Automated claims processing
Use OutSystems and the Amazon Bedrock Connector to automate the claims processing workflow in the insurance industry. Build apps that can quickly review and assess claims based on prompts and inferences, extracting relevant information from documents, photos, and customer statements to accelerate the claims settlement process.
Why OutSystems for AI
The Amazon Bedrock connector is just the latest of many AI innovations from OutSystems that enables developers to add AI features to their apps. But it’s not just about adding AI; it’s also about making AI part of the development process.
For more than five years, OutSystems has been infusing artificial intelligence across the platform to help developers, tech leads, and architects create software much faster and without sacrificing enterprise-grade quality. For example, the OutSystems AI Mentor helps developers at every development step, providing intuitive, in-the-moment help when they need it. It also automates repetitive tasks, eliminates tedious work, and validates applications before they go into production to ensure they are built and maintained to the highest possible quality standards.
Developers can download the Amazon Bedrock connector from the Forge and explore OutSystems platform features with a free trial version.
