TL;DR:
-
A virtual agent is only as good as the knowledge base behind it — structure that first
-
Hostnet handles 5,000–7,000 chats a month using an AI Chat Agent built on structured knowledge
-
AI augments agents, not replaces them: 65% of CX leaders say AI reduces burnout, 77% say it enables personalisation
-
Self-managed control matters as much as AI capability — ask how much your team can adjust without vendor dependency
-
Shared knowledge across virtual agents and human agents keeps answers consistent and cuts repeated customer effort
Chat volumes don't level off. They grow, then grow again — and at some point, adding more agents stops being the answer.
That's the position Hostnet, one of the Netherlands' largest hosting providers, found itself in: handling 5,000 to 7,000 chats a month across a wide range of query types, with service levels becoming harder to hold using human agents alone. The answer wasn't fewer people. It was giving routine queries somewhere else to go.
Self-service works when it's built on real knowledge
A virtual agent is only as useful as the knowledge base behind it. Without structured, well-maintained knowledge management, self-service tools guess. With it, they resolve.
That's the shift worth making:
- Structured knowledge, not scattered documents — agents and virtual agents draw from the same source, so answers stay consistent across self-service and live conversations.
- Case management that connects the dots — when a query does need a human, the full context travels with it, so customers never repeat themselves.
- Continuous improvement — every resolved (and unresolved) query becomes data that sharpens the knowledge base further.
AI augments agents — it doesn't replace them
Puzzel's Virtual Agent Suite, including AI Chat Agents like Hostnet's "Kevin," handles routine queries around the clock — freeing human agents for the complex, judgement-heavy conversations that actually need them.
This matters as much for morale as for efficiency. 65% of CX leaders say AI tools are essential for reducing agent burnout, and 77% say AI is crucial for delivering personalised interactions. The two aren't in tension: when routine volume is handled elsewhere, agents have more capacity to make the conversations they do handle genuinely personal.
Self-managed matters more than most vendors admit
One detail from Hostnet's experience stands out: the value wasn't only in what the AI Chat Agent did, but in how much control the team retained. Building and refining a virtual agent themselves — without depending on external IT for every change — gave Hostnet's support team real ownership of the tool.
For UK and Dutch organisations evaluating virtual agent platforms, this is worth testing directly. Ask not just what the AI can do, but how much of it your own team can manage, adjust, and improve without submitting a ticket to your vendor.
What this looks like in practice
A well-built virtual agent and knowledge management setup should mean:
- Customers get accurate self-service answers around the clock, not just during agent hours
- Agents spend less time searching for information and more time resolving complex issues
- Every handover from virtual agent to human agent carries full context, so nothing gets repeated
Puzzel's Virtual Agent Suite and Knowledge Management tools are built to work together — giving your team structured, AI-native self-service that's genuinely self-managed, not dependent on outside support every time it needs to change.
See how virtual agents and knowledge management work together in Puzzel.
FAQ
A chatbot typically follows scripted decision trees and struggles outside them. A virtual agent draws on structured knowledge management and AI to handle a wider range of queries accurately, and hands off to a human agent with full context when it can't resolve something itself.
No. Virtual agents are designed to absorb routine, high-volume queries so human agents can focus on complex or sensitive conversations. Puzzel's positioning is human and AI by design — AI supports agents rather than replacing them.
This varies significantly by vendor. Hostnet's experience with Puzzel's AI Chat Agent showed that a self-managed setup — where the support team can build and adjust the virtual agent themselves — reduces dependency on external IT and speeds up how quickly the tool improves.
When virtual agents and human agents draw from the same structured knowledge base, customers get consistent answers regardless of channel, and agents spend less time hunting for information — because it's organised, current, and searchable in one place.
