Implementation guides

Build a customer experience people can trust.

Clear, practical guidance for training your AI, installing the widget, capturing useful context, and handing the right conversations to your team.

JENNOX IMPLEMENTATION

A reliable launch has four parts

Approved knowledge
Tested answers
Human fallback
Measured improvement
Answer-first documentation

From first source to a working customer journey

These guides describe the product workflow without pretending that every business should automate the same conversations. Start with a narrow, testable use case and expand after the answers and escalation path are reliable.

Launch a website-trained AI chatbot

Create a chatbot, add approved website or document content to its knowledge base, review the resulting sources, and test representative customer questions before publishing the widget.

  1. 1.Create the chatbot and define its customer-facing name and purpose.
  2. 2.Add only current, customer-safe sources to the knowledge base.
  3. 3.Test common, ambiguous, and out-of-scope questions in the preview.
  4. 4.Configure lead capture and human handover before going live.

Add the Jennox widget to a website

Jennox provides a chatbot-specific embed snippet after setup. Add that snippet once near the closing body tag of the website, publish the change, and confirm that the widget loads on the approved domain.

  1. 1.Finish chatbot setup so the correct chatbot identifier is available.
  2. 2.Copy the generated embed snippet from the Jennox workspace.
  3. 3.Add it once to the public website template, not to an admin-only page.
  4. 4.Verify desktop and mobile loading, then send a real test conversation.

Design a reliable human-handover workflow

Use the AI for questions supported by approved knowledge and route uncertain, sensitive, or high-intent conversations to a person with the conversation context attached.

  1. 1.Define which topics the AI may answer and which require a person.
  2. 2.Set clear escalation language so visitors understand what happens next.
  3. 3.Collect only the contact details the team needs to follow up.
  4. 4.Review unanswered questions and improve the source content over time.

Turn conversations into usable CRM records

Capture a visitor's contact details and intent during the conversation, keep the transcript as context, and assign a clear next action instead of leaving an anonymous chat disconnected from sales or support.

  1. 1.Choose the lead fields that are genuinely required for follow-up.
  2. 2.Use conversation context to distinguish support, sales, and booking intent.
  3. 3.Add ownership, notes, and status so the next action is visible.
  4. 4.Use consent-aware exports or integrations for downstream workflows.

Offer booking at the right moment

Present booking after the visitor shows relevant intent, show only valid availability, and keep the contact and conversation connected to the resulting appointment.

  1. 1.Create the event type, duration, availability, and meeting instructions.
  2. 2.Confirm timezone and calendar behaviour with a test appointment.
  3. 3.Offer booking from relevant chatbot and team workflows.
  4. 4.Check the confirmation and follow-up experience as both host and guest.

Keep answers accurate, private, and reviewable

Treat the knowledge base as controlled customer-facing content: remove obsolete sources, avoid sensitive uploads, limit team access, and review real conversations for gaps and unsafe assumptions.

  1. 1.Upload only content that is approved for customer-facing answers.
  2. 2.Remove outdated policies, pricing, and product instructions promptly.
  3. 3.Use team roles and access controls appropriate to each responsibility.
  4. 4.Escalate legal, medical, financial, or account-specific decisions to people.

Need a workflow reviewed before launch?

Tell us what customers ask, what the AI may answer, and when your team should step in. We can help map the safest first version.

Discuss your implementation