How the Smart Recorder search works
When using the Smart Recording to create a record, it displays related Knowledge articles, tickets, and templates. This topic provides details about how the search decides which records to display based on their relevancy.
Smart Recorder search does not support Rich Text Format (RTF) text entry.
Tickets and knowledge search
Remedy with Smart IT uses Remedy AR Server's Apache Lucene based Full Text Search (FTS) engine. FTS has a Relevancy Ranking algorithm which considers the Search text typed in by the user and the FTS indexes fields for Ticket and Knowledge forms. The relevancy ranking algorithm assigns a score to every record searched and the search results are order by decreasing score value.
How the scoring works
The score, and therefore the relevance of a ticket or knowledge is calculated by considering the following:
- The frequency of search terms across tickets or knowledge records - For example, if the word email is included in 8000 of 10000 tickets or knowledge records, but the word configuration is in a much smaller number of tickets or knowledge records, then the term configuration is considered more relevant than the term email
- The number of times a search term appears in a ticket or knowledge - For example, if the word email appears five times in one of the fields say Summary, that ticket or knowledge is considered more relevant than another ticket or knowledge that might only contain the term email once.
- The length of the text - For example, if a search term is found in a ticket or knowledge that contains a short amount of text, that ticket/knowledge is considered more relevant than a ticket/knowledge with a larger amount of text.
All these factors are considered while deciding the score and position of the search result.
The scoring is done for every field of the Ticket or Knowledge using these factors and the relevancy score at field level is aggregated for a final score for every Ticket or Knowledge.
Influencing the relevancy ranking
You can use some additional configuration within Knowledge Management and Remedy AR System:
- Relevancy Weights for certain fields - For example, searches against Title, Keywords or Environment fields can be ranked higher than rest of the fields.
- Knowledge Use or View Boost factors - Knowledge articles which are viewed or related frequently can be ranked higher than rest of the articles.
For Template matching, following searches are performed
- Full Text Search on Incident Templates. By default, restricted to those templates where agent has access to the Authoring Company defined against them. In Smart IT 1.5.01, there is a configuration in Smart IT Database (Business_Schema) to consider only those templates where the authoring support groups defined against the templates matches the support groups Agent is part of
- Full Text Search on Work Order Template
- Full Text Search on SRDs – restricted to those SRDs where Customer has required Entitlement defined against those SRDs
The Full Text matching is done using the Relevancy Ranking algorithm explained in earlier section. The ranking rules remain the same.
The final result is ordered as follows
- Incident Templates matched in descending order of score (maximum of 20)
- SRDs matched in descending order of score (maximum of 100)
- Work Order Templates matched in descending order of score (maximum of 20)
FTS search is performed on Outages based on CI Name or Outage (Unavailability) Description and restricted to the Company of the Customer (maximum of top 20)
Refer to attached spreadsheet to find out all the fields FTS indexed for Tickets, Knowledge and Templates.