Table of Contents
- Key takeaways
- What is a help desk knowledge base?
- Help desk knowledge base vs FAQ page vs knowledge base software
- Why does a help desk need a knowledge base?
- What are the types of help desk knowledge base?
- Customer help center
- IT help desk knowledge base
- Internal agent knowledge base
- How do you build a help desk knowledge base?
- The KCS article template
- How much time can a help desk knowledge base save?
- Which metrics show a help desk knowledge base is working?
- 5 tools for a help desk knowledge base
- 1. Zendesk: best for a customer help center inside a full help desk
- 2. Freshservice: best for an IT help desk knowledge base
- 3. SiteGPT: best AI answer layer on top of a help desk knowledge base
- 4. Jira Service Management: best free start for a small IT team
- 5. Help Scout: best for a small customer team
- How much does a help desk knowledge base cost?
- The stack: which tool does which job
- How to choose a help desk knowledge base
- Common help desk knowledge base mistakes
- Start with the articles you already have
- Frequently asked questions
- Basics
- What is a help desk knowledge base?
- What should a help desk knowledge base include?
- What is the difference between a help desk and a knowledge base?
- IT help desks
- What is an IT help desk knowledge base?
- What is an IT support knowledge base used for?
- What is the best knowledge base for an IT help desk?
- Building and maintaining
- How many articles does a help desk knowledge base need?
- How often should a help desk knowledge base be updated?
- What is Knowledge-Centered Service?
- Cost and AI
- How much does a help desk knowledge base cost?
- Can a chatbot use a help desk knowledge base?
- Does a chatbot replace a help desk knowledge base?
- Sources

- A help desk knowledge base holds how-to guides, troubleshooting steps, and known fixes inside or next to the ticketing tool.
- Self-service often fails on content. In a Gartner survey of 5,728 customers, only 14% of issues were fully resolved in self-service, and the most common failure was not finding relevant content.
- Build it from your ticket data. Write the top ticket reasons first, use one template, and have agents create and fix articles while they resolve tickets.
- Use the knowledge base your help desk already includes. Zendesk and Help Scout suit customer teams. Freshservice and Jira Service Management suit IT teams.
- SiteGPT is not a help desk. It is the AI answer layer that reads your articles and answers questions in chat, from $39 a month billed yearly.
Key takeaways
Type | Who reads it | Typical articles | Tool that fits |
Customer help center | Customers | How-to guides, billing questions, account steps | Zendesk, Help Scout |
IT help desk knowledge base | Employees and IT agents | Password resets, VPN fixes, software requests | Freshservice, Jira Service Management |
Internal agent knowledge base | Agents only | Escalation rules, known errors, internal procedures | Any of the above, with restricted articles |
AI answer layer | Anyone who prefers to ask | No new articles. It answers from the ones you have |

What is a help desk knowledge base?
Help desk knowledge base vs FAQ page vs knowledge base software
Option | What it is | Connected to tickets? | Use it when |
FAQ page | One page of short answers | No | You have under about 20 simple questions |
Help desk knowledge base | Searchable articles inside the help desk | Yes | You handle tickets and want fewer of them |
Standalone knowledge base software | A separate tool for documentation or company wikis | Only through an integration | Your docs serve more teams than the help desk |
Why does a help desk need a knowledge base?

- Fewer repeat tickets. One article can answer the same question for every person who asks it.
- Faster resolution. Agents search for a known fix before they work the problem from scratch.
- Consistent answers. Every agent gives the same steps, and new agents can learn from the same articles.
What are the types of help desk knowledge base?
Customer help center
IT help desk knowledge base
- Password resets and account lockouts
- Multi-factor authentication setup and new-phone steps
- VPN connection errors
- Wi-Fi, printer, and meeting room setup
- New laptop and new starter setup
- Software access requests and approved software lists
- Known errors with a workaround and a status
Internal agent knowledge base
How do you build a help desk knowledge base?
- Export 90 days of tickets. Group them by reason and sort by volume. The top 20 reasons are your first 20 articles.
- Pick one reader for each article. A customer article and an agent article on the same issue need different words and different detail.
- Agree on one template. Use the same fields in every article. The next section gives a simple one.
- Write titles in the reader's words. Copy the phrasing from ticket subject lines. A title such as "VPN says authentication failed" beats "VPN credential troubleshooting".
- Put search where tickets start. Show article suggestions on the request form, in the portal, and in the agent's ticket view.
- Assign an owner and a review date. Every article gets one named person. Review high-traffic articles every 90 days and archive what no one uses.
- Make articles part of closing a ticket. Gartner advises enabling reps to create knowledge as part of resolving an issue, and not as a separate process. Agents search first, then link, fix, or create an article.
The KCS article template
Field | What goes in it | Example: IT help desk |
Issue | The problem in the requester's words | VPN says authentication failed after a password change |
Environment | What is true for this person's setup | Company Windows laptop, VPN client, password changed today |
Resolution | The answer or the steps | Sign out of the VPN client, remove the saved password, and sign in with the new one |
Cause | Why it happened, if known. This field is optional | The client still holds the old saved password |
How much time can a help desk knowledge base save?
Input | Value |
Tickets a month | 1,500 |
Tickets from the top 20 reasons | 600 (40%) |
Share of those tickets solved by an article | 20% |
Tickets avoided a month | 120 |
Agent time per ticket | 15 minutes |
Agent time saved | 30 hours a month, or 360 hours a year |
Which metrics show a help desk knowledge base is working?
Metric | What it tells you | What to do when it looks bad |
Searches with no result | Which content is missing | Write the article or add the missing term to a title |
Tickets on documented topics | Whether people find the article | Improve the title, or show the article on the request form |
Tickets closed with a linked article | Whether agents use the knowledge base | Add article search to the ticket workflow |
Article flags and low ratings | Which articles are wrong or unclear | Let customers and agents flag content, then fix it |
Articles past their review date | How much content may be stale | Review it, update it, or archive it |
5 tools for a help desk knowledge base
- Knowledge base on the plan. The article tools are included, with a clear entry plan.
- Link to tickets. Agents and requesters reach articles from the place tickets start.
- Fit for the main reader. Customer teams and IT teams need different tools.
- Published price. A team can estimate the cost without a sales call.
1. Zendesk: best for a customer help center inside a full help desk

- Knowledge base, AI agents, and action builder on Suite Team
- Help center as part of every Suite plan
- Omnichannel routing, messaging, live chat, and telephony
- AI writing tools on Suite Professional
- Tickets, help center, and channels live in one product
- Very large review base: 7,120 reviews on G2
- 14-day free trial with no credit card
- The knowledge base is not on Support Team, the $19 plan
- Costs rise quickly per agent
- AI agents are billed on automated resolutions, so usage adds to the seat price
Plan | Per agent a month, billed yearly | Knowledge base |
Support Team | $19 | Not listed |
Suite Team | $55 | Yes |
Suite Professional | $115 | Yes |
Suite Enterprise | Contact sales | Yes |
2. Freshservice: best for an IT help desk knowledge base

- Incident management, knowledge management, and a support portal on Starter
- Service catalog and SLA management on Growth
- Problem and change management on Pro
- Freddy AI Copilot as a paid add-on, and Freddy AI Agent on Enterprise
- Built for IT teams and ITIL-style processes
- The knowledge base is on the entry plan
- 14-day free trial with access to all features
- Change and problem management start at Pro
- Freddy AI Agent is listed only on Enterprise, which has a custom price
- The pricing page shows local currency by visitor location, so check it for your region
Plan | Billing | Knowledge base |
Starter | Per agent a month, monthly or yearly | Yes |
Growth | Per agent a month, monthly or yearly | Yes |
Pro | Per agent a month, monthly or yearly | Yes |
Enterprise | Custom | Yes |
3. SiteGPT: best AI answer layer on top of a help desk knowledge base

- Trains on website URLs and sitemaps, so it can read a public help center
- Trains on files and text, plus Notion, Google Drive, Dropbox, OneDrive, Box, and GitHub
- Fallback message when the content has no answer
- Escalation to a person
- Weekly digest with the questions the chatbot could not answer (Growth and above)
- Auto refresh monthly on Growth and weekly on Scale
- Slack, Google Chat, Crisp, and Freshdesk Messaging from Growth, and Zendesk AI Agent from Scale
- No migration: it works with the articles you already publish
- Unanswered questions show which article to write next
- 95+ languages
- No ticket queue, article editor, or published help center
- Search works by meaning, so exact error codes and part numbers can match less reliably
- No free plan, and the 7-day trial requires a credit card
- Starter has 1 member and manual refresh only
- No Intercom or HubSpot integration
Plan | Billed monthly | Billed yearly | Messages a month | Pages |
$59 a month | $39 a month ($468 a year) | 4,000 GPT-4.1-mini or 400 GPT-4.1 | 1,000 | |
$129 a month | $79 a month ($948 a year) | 10,000 GPT-4.1-mini or 1,000 GPT-4.1 | 10,000 | |
$429 a month | $259 a month ($3,108 a year) | 40,000 GPT-4.1-mini or 4,000 GPT-4.1 | 50,000 | |
Custom | Custom | Custom | Custom |
4. Jira Service Management: best free start for a small IT team

- Embedded knowledge base on every plan, including Free
- Requests by portal, email, chat, Microsoft Teams, Slack, or an embedded widget
- Custom-branded help center from Standard
- Asset and configuration management from Standard
- Virtual service agent on Premium
- Free for 3 agents
- Published prices, with lower per-agent rates at higher team sizes
- Fits teams that already use Jira
- The Free plan has community help only
- The virtual service agent needs Premium
- The per-agent price depends on team size, so use the pricing calculator
Plan | 10 agents, billed monthly | 10 agents, billed yearly |
Free | $0 for 3 agents | $0 for 3 agents |
Standard | $26.90 per agent a month | $2,700 a year |
Premium | $61.60 per agent a month | $6,200 a year |
Enterprise | Contact sales | Contact sales |
5. Help Scout: best for a small customer team
- Docs knowledge base, with multiple knowledge bases on Standard
- Shared inboxes, live chat, and SMS on Standard
- Secure customer portal on Plus
- AI Answers through Envoy at $0.75 per resolution
- A free plan with one Docs site
- Simple to learn
- 15-day free trial of Standard or Plus with no credit card
- Built for customer teams, with no IT service management features
- Extra Docs sites cost $20 a month billed yearly, or $24 billed monthly
- SSO is on Pro
Plan | Per user a month, billed monthly | Per user a month, billed yearly |
Free | $0 | $0 |
Standard | $30 | $25 |
Plus | $54 | $45 |
Pro | $90 | $75 |
How much does a help desk knowledge base cost?

Plan | Yearly cost for 10 agents, billed yearly | Note |
Jira Service Management Free | $0 | Covers 3 agents only |
Jira Service Management Standard | $2,700 | Saves $528 a year against monthly billing, or 16% |
Help Scout Standard | $3,000 | Saves $600 a year against monthly billing, or 17% |
Jira Service Management Premium | $6,200 | Adds the virtual service agent |
Zendesk Suite Team | $6,600 | $55 × 10 agents × 12 months |
Zendesk Suite Professional | $13,800 | $115 × 10 agents × 12 months |
The stack: which tool does which job
Layer | Job | Tool |
Ticketing | Track, route, and resolve requests | Zendesk, Freshservice, Jira Service Management, Help Scout |
Knowledge base | Hold the correct answer in articles | The one built into your help desk |
AI answer layer | Answer in chat from those articles | SiteGPT, or the help desk's own AI |
People | Handle exceptions and write new articles | Your agents |
How to choose a help desk knowledge base
If you need... | Pick... | Why |
A customer help center inside a full help desk | Zendesk Suite Team | Tickets, channels, and the knowledge base are in one plan |
A knowledge base for an IT service desk | Freshservice | Knowledge management is on the entry plan |
A free start for up to 3 IT agents | Jira Service Management Free | The embedded knowledge base is on the Free plan |
A free help center for a small customer team | Help Scout Free | It includes 1 inbox and 1 Docs site |
An IT team already using Jira | Jira Service Management | It fits the tools the team already uses |
Answers in chat from existing articles | SiteGPT | It trains on your content with no migration |
IT answers in a Slack channel | SiteGPT Growth | Slack is included from Growth |
An AI agent that takes actions on tickets | Your help desk's own AI | SiteGPT answers questions and has no ticket queue |

Common help desk knowledge base mistakes
- Writing from memory. Teams guess at topics. Ticket data shows the real demand.
- Using internal language. People search for "VPN won't connect", and they will not find "remote access remediation".
- Having no owner. An article without an owner goes out of date without anyone noticing.
- Hiding search. If search is not on the request form, people open a ticket first.
- Adding a chatbot to bad content. A chatbot repeats outdated articles with confidence.
- Measuring only article views. Views do not show whether the problem was solved.




