FAQ Knowledge Builder

1. Overview

The process reads a policy or guideline document (provided as a PDF) and converts the key information into a list of question‑and‑answer (Q&A) pairs that can be added to the company’s help‑center. Each pair captures a typical user query and a concise, accurate answer drawn directly from the source document.

2. Business Value

• Consistent answers – support staff can respond quickly with a standard answer.
• Reduced tickets – customers find answers themselves, lowering support volume.
• Up‑to‑date knowledge – the FAQ list reflects the latest policy automatically.
• Faster onboarding – new support team members can reference the same Q&A set.

3. Operational Context

• When to run – each time a new policy is published or an existing policy is updated.
• Who uses it – Support Lead, content managers, and anyone maintaining the help‑center.
• Frequency – once per policy change (one run per document).

4. Inputs

Primary Input – Policy Document (PDF)

Note: The PDF must be machine‑readable; if the file is an image‑only scan, the process cannot continue without a text‑based version.

5. Outputs

Output – FAQ Q&A List

6. Detailed Plan & Execution Steps

1. Collect the PDF – verify the file is accessible, readable, and in the correct language.
2. Extract text – open the PDF and extract all text into a working document (e.g., Word or plain‑text file). Ensure the extraction contains no missing sections.
3. Identify sections – locate headings, sub‑headings, or numbered sections that separate topics (e.g., “Eligibility”, “Process”, “Exceptions”).
4. Generate potential questions – for each section, ask what a typical user might want to know (e.g., “How do I request a refund?”). Write each question as a clear, stand‑alone sentence that begins with a question word (What, How, When, etc.).
5. Craft answers – locate the paragraph(s) that answer each question. Summarise the answer in plain language, keeping it short (1‑3 sentences) and in present tense. Do not add information not present in the source text.
6. Pair and format – place each question and its answer into a row of the final table. Keep punctuation, capitalization, and tone consistent.
7. Quality review – confirm each answer directly reflects the source paragraph, that there are no duplicate questions, and that no personal or confidential data appear.
8. Output generation – compile the rows into the final “FAQ Q&A List”. The result is a plain‑text table that can be copied into the help‑center system.

7. Validation & Quality Checks

• Coverage – each major policy topic has at least one Q&A pair.
• Accuracy – each answer matches the source text; verify by cross‑checking the source paragraph.
• Clarity – questions are clear, answers are concise, and both are written in plain language.
• Data privacy – scan for any personal or confidential information; remove if found.
• Formatting – each row has a Question column and an Answer column; no missing columns.

If any check fails, flag the item for manual review; do not produce the final output until the issue is resolved.

8. Special Rules / Edge Cases

• Non‑text PDFs – if the PDF is an image‑only scan, mark the run as Error – unreadable PDF and request a text‑based version.
• Very long documents ( > 50 pages ) – split the document into sections and run the SOP for each section separately.
• Ambiguous content – if a paragraph cannot be turned into a clear question, note “Manual review required – no clear Q&A found”.
• Duplicate questions – merge identical questions, combining the answers if they differ.
• Missing sections – if a required policy section (e.g., “Refund process”) is missing, create a placeholder entry: “TBD – review required”.
• Error handling –
  • If the PDF cannot be opened, set status Error – File not Accessible and stop.
  • If any required input detail (e.g., language) is missing, set Error – Missing Information and stop.

9. Example

Input

• File name: “Return_Policy_2025.pdf”
• File format: Text‑based PDF (30 pages, English).
• Content summary:
  • Customers may request a full refund within 30 days if the product is unused and in its original packaging.
  • A request must be submitted via the online portal with a purchase receipt.
  • Refunds are processed within 5–7 business days after receipt of the returned item.

Output – FAQ Q&A List

Appendix A – FAQ (About the Process)

Q1: The policy is written in a language other than English. What should I do? Translate the document into English before starting the process. This SOP does not include translation.

Q2: The PDF is a scanned image. Can the process still run? No. Request a text‑based version of the document; the SOP requires selectable text.

Q3: The policy changes after I have created the FAQ. How do I update it? Run the SOP again with the updated PDF and replace the old Q&A entries with the new list.

Q4: Who should review the final Q&A list? The Support Lead or designated content manager should perform a final review before publishing.

Q5: Can I use this process for marketing brochures? Only if the brochure contains factual, policy‑type information suitable for a help‑center. Otherwise, use a different process.

Q6: What if a question needs a numeric answer (e.g., “How many days…”) Include the exact number from the source; keep the answer concise.

Q7: Should I keep internal references like “Section 2.3” in the answer? No. The answer should be self‑contained; remove internal references.

Q8: What if there is a conflict between two sections? Use the most recent version of the policy. If the conflict is unclear, flag it for legal review.

Appendix B – Glossary

Appendix C – Style Guide for FAQ Creation

1. Tone – Friendly and professional; aim for clarity and brevity.
2. Question format –
   • Start with a question word (What, How, When, Who, etc.) or a verb.
   • Keep under 15 words if possible.
   • Use a question mark at the end.
3. Answer format –
   • Write in plain, present‑tense language.
   • Keep to 1‑3 sentences.
   • Use complete sentences and end with a period.
   • Do not include internal codes, IDs, or personal data.
4. Consistency –
   • Capitalize the first word of each question.
   • Use the same punctuation style throughout.
   • Do not embed bullet lists inside an answer.
5. Prohibited content –
   • Personal data (names, emails, phone numbers).
   • Internal reference numbers (e.g., ticket IDs, policy numbers).
   • Confidential legal or financial advice not present in the source.
6. Formatting –
   • Question – “What is the refund period?” Answer – “Refunds are issued within 5‑7 business days after the returned item is received.”

Good example: Question: “What is the refund period?” Answer: “Customers can request a refund within 30 days of purchase if the product is unused and in its original packaging.”

Bad example: “Refunds? 5‑7 days.” (missing subject and not a full sentence).

Additional Notes

• Version control – Keep the original PDF and record the date the Q&A list was generated for future reference.
• Collaboration – Share the generated Q&A list with the content team for a final editorial review before publishing.
• Future updates – When a policy is updated, re‑run this SOP with the new PDF and replace the old Q&A list.
• Audit trail – Record the file name, date processed, and name of the person who performed the conversion to maintain accountability.