Knowledge management · Autessa IT Desk
Knowledge search
Autessa IT Desk knowledge search blends keyword ranking with semantic search, so a question like "my VPN keeps dropping" finds the right article even when the words differ. Agents search every knowledge base in the workspace, and employees search published Employee articles in the Help Center.
The problem
People ask questions, not keywords
An employee types a sentence. A plain keyword search matches the wrong words, misses synonyms and ranks a long article that mentions "VPN" once above the short how-to that solves it. When search fails, people open a ticket.
What you get
- Hybrid ranking: rarity-weighted keyword coverage blended 60/40 with semantic similarity.
- Up to six search keywords per question: the article's author-supplied keywords first, then the question's own content words, with stop words dropped and plurals folded.
- Title matches count most, then summary and keywords, then the body; a question that names an article's title outright jumps it to the top.
- Semantic-only matches must show real similarity, so vague results do not crowd out the right answer.
- When semantic search is unavailable, results fall back to keyword ranking instead of failing.
- Expired and retired articles are never served, and the Help Center only returns published Employee articles.
How it works
How does knowledge search work in Autessa IT Desk?
Someone asks
An agent searches the knowledge home in the workspace, or an employee types into the Help Center search box.
Keywords are chosen
IT Desk picks up to six keywords: the article keywords authors wrote, then the meaningful words of the question.
Two kinds of matching
Keyword matching scans titles, summaries, keywords and bodies, and one semantic pass looks for articles with similar meaning.
One ranked list
Scores are blended, title hits get a boost, and the reader sees the best answers first with a short summary of each.
Every option
What affects ranking
| Option | What it does |
|---|---|
| Title | Highest weight for a matching keyword; +0.1 when the rarest matched keyword is in the title. |
| Question names the title | +0.3 when the question contains the article title outright. |
| Summary and keywords | Weighted 0.75: write good keywords and synonyms here. |
| Body | Weighted 0.55, saturated so long articles do not win just by length. |
| Semantic similarity | Blended at 40%; semantic-only results need a normalized score of at least 0.2. |
| Scope | Workspace: IT and Employee knowledge bases. Help Center: published Employee articles only. |
| Expiry | Articles past their valid-to date, and retired articles, are left out. |
Permissions
Who can do what?
| Role | Can |
|---|---|
| IT agent | Search both knowledge bases from the workspace knowledge home and global search. |
| Knowledge author | Search everything they can write, and improve keywords so articles rank well. |
| Employee | Search published Employee articles and catalog items together in the Help Center. |
AI assist
How the IT Desk Assistant helps
The IT Desk Assistant uses the same knowledge search to answer questions in the workspace and the Help Center, and lists the articles it based its answer on.
The assistant acts as you, with your role. It fills forms but never saves them, and anything risky waits for you to press Confirm.
About the IT Desk Assistant →Connected
How it connects to the rest of IT Desk
IT DESK FEATURE
Knowledge base
Keywords, summaries and expiry dates written in the knowledge base drive what search returns.
IT DESK FEATURE
Help Center
Portal search returns catalog items and Employee knowledge side by side, so people find a fix or the right request.
IT DESK FEATURE
IT Desk Assistant
The assistant answers from the same ranked articles and cites them.
IT DESK FEATURE
Global search
Workspace global search includes knowledge alongside tickets, people, assets, CIs and vendors.
FAQ
Questions about knowledge search
Is IT Desk knowledge search semantic?
Yes, it is hybrid. It blends keyword ranking with semantic similarity, so it finds articles that mean the same thing even when the words differ.
What happens if semantic search is unavailable?
Results fall back to keyword-only ranking. Search keeps working; it just matches on words for that query.
Can employees see IT runbooks in search?
No. The Help Center only searches published Employee articles. IT articles stay in the Agent Workspace.
How do I make an article easier to find?
Add keywords and synonyms people actually type, and write a clear title and summary. Those fields weigh more than the body.
100% customizable, code and all
Make Autessa IT Desk fully your own.
Use it out of the box, then change anything. Ask Autobot, Autessa’s AI builder, for a new field, screen, report, workflow or agent and it builds it into your copy of the app. Or open the code and edit it yourself. Nothing is locked: the app, its data and its code are yours.
Start free and make it yours