AnswersIQ™ - Overview and Implementation Guide

AnswersIQ™ is designed to empower your agent to deliver results-driven, context-aware interactions, keeping your customers informed while your team focuses on more strategic, higher-value work. By matching lead questions to the best responses in your knowledgebase, AnswersIQ enables:

• Automated fact matching and timed delivery of relevant answers
• More natural, specific responses that improve customer satisfaction
• Simple fact management with multiple ways to add and maintain knowledge

With these advanced capabilities, your business can deliver a more seamless and engaging customer experience.

Key Features of AnswersIQ

• Automated Fact Collection from Customer Websites
  AnswersIQ collects essential information directly from your company's website, automatically compiling facts for quick reference by the assistant. This saves time and ensures your virtual assistant has accurate, up-to-date answers without requiring manual updates.
• Improved Matching of Answers to Lead Questions
  Our advanced matching algorithms enable AnswersIQ to link lead questions with the most relevant answers, ensuring your customers receive the information they need with minimal follow-up questions. This precise matching significantly reduces customer confusion and enhances user satisfaction.
• More Organic and Specific Content in Interactions
  With AnswersIQ, your assistant can provide responses that feel more natural and relevant to specific inquiries. This feature results in richer, more engaging conversations and allows the assistant to deliver responses that are contextually tailored to customer needs.
• Enhanced User Interface for Fact Management
  Managing facts is simpler and more intuitive with the enhanced UI in AnswersIQ. This feature makes it easy to review, edit, and organize information so your assistant always has the most relevant data at its fingertips.

How does it work?

Our system optimizes responses by matching your lead's questions to predefined answers and appending them to scheduled messages in a two step approach.

Step 1: Answer Matching

• Identifying Relevant Answers: The system analyzes each lead's question and finds relevant answers based on customer-defined information.
• Selecting the Best Match: If multiple answers are found, the system automatically chooses the answer with the highest confidence score.

Step 2: Appending to Scheduled Messages

• Adding Matched Answers to Messages: Once an answer is matched, it is appended to the next scheduled email from the assistant.
• Timing Consideration: If the next scheduled message is set for more than two hours later, the system sends an informational email with just the matched answer, ensuring timely delivery of information.

This ensures that leads receive the most relevant answers at the right time.

Updates to Leads to Review

When you are migrated over to AnswersIQ, there are updates to how Leads to Review are worked and flagged in the system.

When a conversation is flagged as Lead to Review, the system now automatically checks for a relevant answer. If a matching answer is found, an informational email is sent to the lead with the answer.

After the informational email is sent, no additional automated messages will be delivered. Reps can intervene by using the Reply as Assistant feature as needed to continue the conversation.

If the lead replies to the informational email sent with the Answer, the system will:

• Reclassify the response.
• Move the lead to the next conversation stage or repeat the informational email experience if necessary.

Ways to Add Facts to AnswersIQ

You can populate your AnswersIQ knowledgebase in three ways:

1. Manually in the platform — Create or edit facts directly via the AnswersIQ Fact Management. To learn more about creating and managing facts, see full article here.
2. Request a website crawl — Ask Support to run a crawl of your public website; we’ll harvest facts and add them to your factbase. This is useful for large public knowledge sources.
3. Send a CSV to Support — Prepare a properly formatted knowledgebase CSV and send it to Support for bulk upload.

Knowledgebase/CSV Requirements

To ensure successful ingestion, prepare your CSV exactly as specified below.

Required columns (must be present)
These three columns are required for every fact:

• topic — The category or subject area of the fact.
  Example: AI Sales & Marketing
• fact — The main factual content to be stored.
  Example: Conversica is a company that provides AI-powered sales and marketing solutions.
• queries — The list of search queries that should match this fact. This must be a valid JSON array stored in a single CSV cell, for example:
  ["What does Conversica do?", "What are the benefits of using AI?"]
  (Important: queries is NOT a plain comma-separated string. See the queries section above for full details.)

Optional columns
The following columns are optional. Include them when you want to associate a source or a target URL with the fact, or to mark the fact as a lead-capture item.

• source_type — The type of source where the fact was found. Defaults to website if not provided. Possible values: website, pdf (use lowercase).
  Example: website
• source — The source URL where the fact was found. Only used (and most useful) when source_type is website.
  Example: https://www.example.com/about-us
• target_url_value — A destination URL to associate with this fact; when present it can be surfaced alongside the fact.
  Example: https://www.example.com/pricing
• target_url_name — A display label (URL mask) for the target_url_value. Only used if target_url_value is provided.
  Example: View Pricing
• lead_capture_flag — Marks whether the fact is a lead-capture fact. This field is optional and defaults to false when not present. Use true or false (lowercase).
  Example: true

Required formatting (Example)

• Headers must be lowercase and exact. 
• Field name: source_type (not Source Type or source type). Values in source_type should be lowercase (e.g., website).
• No blank/missing cells. Remove empty rows and ensure required fields are filled.
• Remove newline characters inside cells (replace \n, \r\n, \r with a single space). Newlines inside fields commonly break parsing.
• Avoid control characters and unusual punctuation that may cause parsing issues.

• The queries column must contain a valid JSON array inside a single CSV cell. Do not provide queries as a plain comma- or semicolon-separated string — that will cause ingestion failures or cause fact and queries to merge. Example: 

  ["What are your hours?","When is the box office open?"]

  
  Note: The queries JSON array is required only for bulk CSV uploads that you send to Support for ingestion.

  If you add or edit facts directly in the AnswersIQ knowledgebase UI, you do not need to provide a JSON array — use the UI’s query input fields to enter multiple queries (the platform will handle the internal format). The JSON format is only necessary when submitting facts in a CSV for Support to upload.

  Removing newlines & cleaning text

• Replace line breaks with a space before uploading: \n, \r\n, \r → .
• Collapse multiple adjacent spaces to one.

• If your facts contain multiple paragraphs, convert them to a single-line format suitable for the CSV cell.

  Encoding & CSV Type

• Use UTF-8 encoding.
• Files saved with a UTF-8 BOM (byte sequence 0xEF 0xBB 0xBF) are commonly labeled CSV UTF-8 (Comma delimited) in some editors — that is acceptable.
• If no BOM is present, save as CSV (Comma delimited) with UTF-8 encoding.

Preserve the encoding when saving the cleaned file.

Quick Validation Checklist

• Headers are lowercase and exact.
• queries contains a valid JSON array in a single cell.
• No newline characters in any field.
• No blank or missing required cells.
• source_type is present and lowercase where applicable.
• File saved in UTF-8 (BOM or no-BOM).

Troubleshooting - Common Issues and How to Fix Them

If facts or leads behave unexpectedly, check the following:

• Malformed queries — Ensure queries is a JSON array in a single cell. If fact and queries look merged, fix the CSV so queries is a JSON array cell.
• Newlines inside fields — Remove all newlines inside cells. Multi-line cells typically cause rows to split.
• Header problems — Confirm headers are exactly as required and lowercase (e.g., source_type).
• Encoding issues — Use UTF-8. If you see strange characters, resave as UTF-8.
• Fact indexing — Allow a short time for uploaded facts to become searchable after ingestion.

FAQs

Q — How can I get access to AnswersIQ?
A — AnswersIQ is available per client request. Please contact your Customer Success Manager (CSM) to enable the feature and start the onboarding process.