In short
A business school maintained a HubSpot chatbot by hand, with conflicting sources and a marketing team that had no time to spare.
Skolbot compared the content, resolved the discrepancies and handled the entire configuration of the Web Agent.
All the school had to do was test the agent and give feedback in the platform.
The agent was live 2 weeks after signing.
A fast launch hides the work
Many tools promise a chatbot live in 24 hours. In practice, the school writes the scripts and corrects its own content, and the project often lasts 6 months. This business school had a chatbot built in HubSpot. Its marketing team no longer had time to maintain it.
The decision trees and scripts took in-house writing and tuning. On top of that, the website, the brochures and the chatbot’s answers contradicted each other.
Marketing carries the whole project
Behind the chatbot, the school had to write the flows, configure the answers and sort out conflicting information.
Hand-built scripts
Every decision tree took work in HubSpot. The team also had to keep the scripts up to date as things changed.
Sources that disagree
Tuition fees, the required English test and the start date varied from one document to another. The right version had to be found before answering.
No time to spare
Marketing had to absorb this preparation on top of its other work. The launch depended on how much time the team could give it.
Skolbot prepares the sources and configures the agent
The Skolbot team handled the content preparation and the entire set-up of the Web Agent.
- 01
Read and compare every source.
Skolbot read the website pages, the brochures, the FAQ and the internal documents. Every piece of information was extracted, then compared with the other sources.
- 02
Resolve discrepancies between sources.
One page showed tuition fees of €8,900, against €9,400 in the current brochure. Skolbot kept the current brochure’s figure and applied it everywhere. The comparison continues after launch.
- 03
Configure the agent and keep HubSpot.
Skolbot did all the configuration, with no scripts for marketing to write. HubSpot remains the school’s CRM and still receives the leads.
- 04
Have the school test the agent.
The team tested the agent and left its feedback directly in the platform. Skolbot applied every piece of feedback. That was the only thing asked of the school.
A go-live with no in-house configuration
weeks from signing to go-live
The Web Agent was live 2 weeks after signing. The marketing team had no scripts to write and no configuration to do. Skolbot carried out all the preparation and set-up work.
The only ask of the school: test the agent and give feedback in the platform. Every question now has a single answer, applied everywhere. Leads still arrive in HubSpot.
“We had already tried other solutions, but setting them up took months. With Skolbot, everything was taken care of: configuring the agent, the knowledge base and the integration with our CRM. Two weeks after signing, our agent was live. No development and no scripts to build on our side.”
Frequently asked questions
How long does it take to set up an AI agent on a school website?
At a business school, Skolbot’s Web Agent went live 2 weeks after signing. Skolbot read the sources, compared the information and did the configuration. The school only tested the agent and gave feedback in the platform.
Who configures the AI agent and writes the scripts for the school?
The Skolbot team handles the set-up of the Web Agent. At a business school using Skolbot, marketing had no scripts to write and no configuration to do. It only tested the agent and left feedback in the platform.
Can you replace a HubSpot chatbot and keep HubSpot as your CRM?
Yes: a business school replaced its HubSpot-built chatbot with Skolbot’s Web Agent. HubSpot remains its CRM and still receives the leads. Skolbot handles the agent’s configuration from the school’s content.
How do you fix contradictions between a school’s website and its brochures?
Skolbot extracts the information from web pages, brochures, FAQs and internal documents, then compares it. Skolbot flags each contradiction with its sources and resolves it once. The same answer then applies everywhere.
How do you keep an AI agent’s answers up to date after launch?
At a business school, Skolbot keeps comparing the sources after the Web Agent goes live. Any new contradiction is spotted the same day. Skolbot corrects it so that each question keeps a single answer.