Short answer
Focus on including important terms, relevant synonyms, and common variations used in your documents, organized into logical Categories. Always include one “catch-all” Category for general terms. Review and update regularly to keep it aligned with your Index content.
Full explanation
A strong Semantic Map ensures AskTuring understands the language used in your documents and connects related concepts accurately. Categories help organize these terms into meaningful groups, making the map easier to manage and improving the AI’s accuracy.
Behind the scenes, this process is powered by AskTuring’s patented technology, which links terms, categories, and alternatives together so the AI can interpret user queries more intelligently.
1. Create Categories
Semantic Maps typically have 4–10 Categories.
Each Category is a group of related terms that make sense together, such as Departments, Processes, Regulations, or Client Names.
Always include one “catch-all” Category, usually called General Terms, for items that don’t fit neatly anywhere else.
2. Identify key terms from your documents
Review your Index content to find specialized terms, acronyms, product names, or industry jargon important for search accuracy.
3. Add synonyms and alternative phrases
For each key term, list variations that users might type into the search box. This includes:
Synonyms (e.g., attorney and lawyer)
Acronyms and full spellings (e.g., IRS and Internal Revenue Service)
Common misspellings (if they occur frequently in your documents)
4. Group terms under the right Category
Place each term in the Category where it makes the most logical sense.
Use “General Terms” only for items that do not clearly belong elsewhere.
5. Keep it focused
Avoid unrelated or rarely used words.
Too many loosely related terms can reduce precision.
6. Update regularly
Whenever new documents are added to your Index, check for new terms or acronyms and add them to the correct Category in the Semantic Map.
Why it matters
A well-maintained Semantic Map, enhanced by AskTuring’s patented technology, improves both:
Recall → finding all relevant documents
Precision → reducing irrelevant results
By organizing terms into Categories (with one “catch-all” Category for General Terms), you ensure AskTuring can handle both highly specific and more generic queries accurately.
Tips
Involve subject matter experts to ensure you capture the right terminology.
Use search logs to identify terms people search for that aren’t returning enough results.
Review and clean up your Semantic Map quarterly to keep it relevant.