AWS re:Post introduces re:Post Agent

AWS re:Post Agent: a Generative AI-Powered Virtual Assistant

AWS re:Post Agent is a generative AI-powered virtual assistant that provides the first response to questions posted on AWS re:Post. Powered by Amazon Bedrock, a fully managed service for foundation models, re:Post Agent leverages high-quality AWS content to deliver more complete, actionable, and referenced answers to help accelerate your AWS development. The answers provided by re:Post Agent undergo review and validation by human experts, fostering a collaborative process that continuously builds collective wisdom. Community experts can earn reputation points by reviewing re:Post Agent's responses, helping to expand the availability of public knowledge across AWS services. re:Post Agent integrates seamlessly into the existing re:Post experience. When a question is asked, re:Post Agent provides a response within seconds, sharing relevant code samples, documentation links, and step-by-step instructions along with source citations. Customers with a Premium Support plan continue to receive prioritized responses, with re:Post Agent answers to their questions receiving expert reviews from AWS Support. Community members can provide feedback by upvoting or downvoting the answers from re:Post Agent. Question authors can Accept the answer, members with an “Expert” reputation status can review the answer and edit it for accuracy. The feedback provided is visible to all re:Post users, giving cloud developers the confidence in the information, while continuously improving re:Post Agent’s responses.

Opportunity with Generative AI

With more than 250 fully featured AWS services today, the sheer volume of knowledge and the pace of change make it challenging for human experts alone to keep up and effectively manage knowledge. The emergence of generative AI presents a transformative opportunity to augment and amplify human intelligence in the realm of knowledge management. By seamlessly integrating human subject matter expertise with the capabilities of advanced generative AI models, re:Post aims to be a knowledge service that addresses the limitation of traditional knowledge sharing approaches. Ultimately, re:Post Agent aims to empower AWS customers by providing them quick access to the most relevant insights and information, enabling them to make more informed decisions in their cloud development efforts.

Why is this important?

The launch of AWS re:Post Agent represents a significant milestone in re:Post's journey to integrate human expertise with the capabilities of advanced generative AI models. Humans bring contextual understanding, creativity, problem-solving skills, ethical decision-making capabilities, emotional intelligence, and empathy. Generative AI, on the other hand, excels at rapidly processing large data sets, identifying patterns and trends, continuously curating and managing content, and providing personalized recommendations and insights.

Frequently Asked Questions (FAQs)

Q1. Who can use re:Post Agent?

Any re:Post user who signs in and posts a question will receive a response by re:Post Agent. All re:Post users (signed in or not) can view the response from re:Post Agent.

Q2. How can I provide feedback to re:Post Agent?

Question authors can provide feedback by accepting the answer if it has answered their question. Community members with an “Expert” reputation status can review the answer. Upon reviewing, they can publish it as is or edit it for accuracy. All other members can upvote, downvote, or flag the answer. All feedback provided will help improve future responses from re:Post Agent.

Q3. Can I ask re:Post Agent a follow up question?

No, re:Post Agent is designed to provide the initial response to questions posted on re:Post. It does not engage in a conversational dialogue or provide follow-up responses to the same question.

Q4. Can I revise my question to obtain an updated response from re:Post Agent?

No, revised questions are not resent to re:Post Agent. We recommend you post a new question if it has not been asked by other community members.

Q5. How many points will I earn for reviewing an answer from re:Post Agent?

The initial expert reviewer will earn +5 points.

Q6. What language(s) can re:Post Agent respond to questions in?

re:Post Agent can respond to questions in English, French, Japanese, Korean, Simplified Chinese, and Traditional Chinese.

Q7. What questions can re:Post Agent respond to?

re:Post Agent will attempt to answer all general AWS guidance technical questions except for security related questions.

Q8. When is re:Post Agent unable to respond?

re:Post Agent is unable to respond to security related questions. The agent is also unable to respond to questions that do not relate to AWS technology.

Q9. How do I verify whether the answers from re:Post Agent are accurate?

re:Post Agent provides the sources used for each response so that you can research and validate the response. Because re:Post Agent responses are published and available for the community to view and for experts to correct and validate, you can monitor community upvotes/downvotes and expert review status to evaluate the quality of re:Post Agent responses.

Q10. What should I do if the response from re:Post Agent is inaccurate or incomplete?

If you believe the response is inaccurate or incomplete, please downvote the response. If you have earned Expert status, you can perform an expert review. If a response needs improvement, you can either correct or delete the response. If you believe the response from re:Post Agent is harmful, please flag the response and it will be escalated for moderator review.

Q11. Will I lose points for downvoting inaccurate answers from re:Post Agent?

No, community members will not lose points for downvoting answers from re:Post Agent. Members are encouraged to downvote answers that are inaccurate or misleading.

Q12. Does re:Post Agent earn points for answering questions?

The re:Post Agent earns points in a similar manner to community members, receiving points for answers that are validated by expert reviewers. However, the agent itself will not have a visible reputation status or be displayed on the community leaderboard, as its role is to support and empower the broader community.

Q13. What happens if a re:Post Agent answer is flagged?

Flagged responses will be reviewed by internal moderators, who can remove the answer from public view if the flag is valid. Flagged responses are used to further train the re:Post Agent to improve future responses.

Q14. How will users' privacy and data security be protected when interacting with the re:Post Agent?

re:Post Agent will only have access to the data and content that is publicly available on re:Post and other AWS authoritative sources. It will not have access to any private or sensitive customer data.