PeRCI, your friendly Public AI Assistant

General query tips:

• 1

  Clarity:

  When writing questions to me, clarity and specificity is key. With more detailed questions, I will better understand what you are looking for and can give more targeted answers.

• 2

  Iterate:

  If I am having trouble finding what you are looking for, try asking another way, giving more information, or providing examples.

• 3

  Converse:

  Unlike a traditional chatbot, I am great at conversing. You can talk to me like you would talk to a friend, colleague, or teacher. If you do not understand something or want more information, do not end the conversation - ask follow-up questions, for clarification, or for specific improvements in my answer.

• 4

  Always check facts:

  I am constantly learning and improving. However, sometimes I get things wrong so please review the resource links I provide with my answer to ensure the information is correct.

• 5

  Use my context memory:

  I will use the last 50 questions and responses in our conversation to help build context and understanding of your enquiry. If I do not get it right the first time ask again with additional context or detail, and it will improve my response. To clear my memory, refresh the webpage.

Basic Examples:

Here is an example of how to incorporate the basic tips and best practices when asking me a question.

• 1

  Poor query:

  What PeRCI reports completed before 2018 have studied cracking?

  REASON IT’S A POOR QUERY: Besides offering little in terms of context beyond the word cracking, looking for reports before a certain time will be difficult for me as my search functions are not based on key words/fields. For better results avoid searching based on date and instead focus on the topic and providing enough context for me to find it.

• 2

  Better query:

   

  Tell me about ILI data and pipeline failures due to cracks to better understand susceptibility of stress corrosion cracking.

• 3

  Even Better (Follow up query to above):

   

  Can you summarize the key results and conclusions of the above reports on stress corrosion cracking.

Specific Examples:

• 1

  Asking for summaries about a topic:

  You can ask me to give a high-level summary of any topic covered by projects or research in PeRCI’s database. Keep in mind that detailed or factual claims should be checked against the linked resources for accuracy.

  Example query:

  Can you explain how to model external corrosion for performing quantitative risk assessment?

• 2

  Asking for a list of reports:

  Asking me to list reports can encourage me to broaden my resource search so I can find more relevant data links.

  Example query:

  Can you list reports related to performing In line inspections on natural gas pipelines to detect corrosion?

• 3

  Summarize a report:

  To summarize a report effectively, you may need to ask me some leading questions to give additional context to improve my ability to create a proper summary. Additionally, it will help me find more information regarding the report if you give context about the report’s content. Please note that this is just a rough draft, so please review the summary I give and make any necessary corrections or clarifications before sharing it with anyone.

  Example query:

  User: Provide reports related to Large-Scale Cathodic Disbondment Testing for Coal Tar Enamel
  Prof PeRCI: <responds with list of related resources>
  User: Could you give a short summary of each of the above reports?
  Prof PeRCI: <Summarizes each of the reports>
  *Note: this is to enter the report summaries into my memory to give more context for the final query below*
  User: Using the above summaries, could you create a detailed summary of the report:
  Large-Scale Cathodic Disbondment Testing for Coal Tar Enamel (CTE)