Modern support teams juggle multiple tools, constant context switching, and overwhelming ticket queues. Even with advanced systems, agents often face fatigue and delayed response times—both of which directly affect customer satisfaction.
In this webinar, Software Developer Anto Gibson showed attendees how a single BoldDesk® AI agent can streamline support operations, reduce the load on human agents, and deliver instant, accurate answers around the clock.
If you missed the webinar or would like to watch it again, the recording is posted to our YouTube channel and embedded below.
BoldDesk AI Agents
BoldDesk AI Agents are far more than automated responders. They understand customer queries, search knowledge bases, execute API-driven actions, and respond instantly.
They operate continuously without breaks, burnout, or delays—providing immediate support whenever customers need help.
Live demo: Building an AI Agent in BoldDesk
Anto demonstrated how to:
- Create and publish an AI Agent.
- Embed the agent directly into a live chat widget.
- Connect the agent to select knowledge base articles.
- Add API actions for order lookups, address updates, and bulk order pricing.
Real-world scenarios from the demo
Within minutes, the AI Agent successfully:
- Answered questions about the coffee shop.
- Retrieved and filtered an order history.
- Updated a shipping address.
- Calculated bulk order amounts.
Why AI Agents outperform traditional support
AI agents eliminate the biggest challenges support teams face: fatigue, context switching, and tool overload.
They also leverage MCP servers and guardrails to safely access APIs, validate user queries, and determine when to proceed or decline actions.
Agenda
- [00:00] Welcome and Introduction
- [00:27] Agenda Overview
- [01:30] Why Support Teams Still Struggle
- [03:10] What Makes AI Agents Different
- [04:22] Benefits of a Unified AI Support Agent
- [05:23] Demo Start: Building an AI Agent
- [06:22] Creating and Publishing the AI Agent
- [07:17] Embedding the Agent in Live Chat
- [08:46] Adding Knowledge Base Content
- [09:50] Selecting Specific Knowledge Base Articles
- [11:28] Integrating APIs for Advanced Actions
- [12:07] API Action 1: Fetch Existing Orders
- [14:05] API Action 2: Change Shipping Address
- [17:08] API Action 3: Bulk Order Calculation
- [19:24] Testing the Agent in Real Scenarios
- [23:00] Demo Summary and Key Takeaways
- [25:03] AI vs. Human: Speed and Efficiency Comparison
- [26:46] Extending with MCP Server and Guardrails
- [27:29] Final Thoughts and Trial Information
Q&A on one agent vs a whole team
Q: How do I add parameters to the API query based on the context? E.g., to avoid a user querying the orders of another user.
With the use of guardrails, the current user can be verified with an OTP sent by email, and then the current user’s active orders can be retrieved. This way, we can prevent users from querying another customer’s orders.
Q: Will it save all the information inside the database?
Each conversation history is saved as trackable records and reports but not specifically in a database. So, once a conversation is closed, the history is stored as reports, not as a data source for the AI.
Q: How does the pricing work? Are there any limits to the number of knowledge sources?
You can create as many knowledge sources as needed. Check out our pricing page for details on our pricing plans.
Q: Can you create AI agents using the Syncfusion Community License?
No, you cannot create a BoldDesk AI Agent using a Syncfusion Community License.
Q: How flexible can the knowledge source be? Does it answer from table-like data, too?
Yes, the agent can read data from tables and text like most AI tools. Images can get tricky, so it is better to maintain a text format.
Q: Can the knowledge sources be images, too? If I have images from PDFs, does it analyze those, too?
No, it is not proficient enough to analyze data or text inside an image, so keeping it text-only will improve accuracy.
Q: What if we find that the AI Agent is hallucinating? Can an administrator force some behavior?
Yes, administrators have full control over the AI Agent’s behavior, so hallucinating can be controlled, and specific brand tones can be set.
Q: We have a n8n workflow already in place. How easy is it to integrate such with BoldDesk?
n8n can call BoldDesk’s REST APIs (ticket creation, updates, customer info) using its HTTP Request node, so integration would be very easy.
Q: Are there any new AI features on your roadmap?
Yes, we have many AI features planned that are in progress. You can check out our roadmap page.
Q: What type of output can an AI Agent generate? Is it like ChatGPT?
It can generate most responses that are text-related. Unlike ChatGPT, BoldDesk AI Agents focus on analyzing data or performing actions rather than being a general AI tool.
Q: Are the interactions tracked for later audits?
Yes, all AI agent interactions are tracked and maintained in a dedicated report section.
Q: Can you track the source used for the answer?
Yes, the answer sources, like knowledge base articles or webpage references, are trackable.
Try BoldDesk AI Agents
Start a free 15-day trial to explore how our AI features can transform your support workflow as briefed in this webinar show notes.
Related resources



















Email Ticketing System
Shared Inbox Software
Multi Brand Help Desk
Internal Help Desk Software
Trouble Ticketing Software
Mobile Help Desk 










