Introducing Wilbur - LeadTruffle's Internal AI Assistant (BETA)
📝 Wilbur is your AI assistant for inspecting your LeadTruffle setup, diagnosing issues, and making updates you approve before they apply.
Table of contents
- Overview
- Where to find Wilbur
- What Wilbur can see
- Diagnosing issues with Wilbur
- Making updates with Wilbur
- Reviewing and approving changes
- Previews that hand off to your team
- When Wilbur hands off to support
- Tips for the best results
Overview
Wilbur is the AI assistant built into LeadTruffle. Think of it like a ChatGPT-style chatbot that knows your account — you can ask it questions, track down why something isn't working, and make updates to your setup, all from one place.
Two things make Wilbur safe to lean on:
- It's scoped to your account only. Wilbur never looks at another company's data and never exposes API keys, passwords, or credentials.
- It never changes anything silently. Every update Wilbur recommends is shown to you as a proposed change with a clear before-and-after view. Nothing happens until you approve it.
A common example: your AI is telling leads it costs $50 for a service call, but it's actually $100. In the past, you'd have to dig through your knowledge base and every channel prompt to find where that wrong number is buried. With Wilbur, you can just say — "The AI is saying our service call is $50, but it's actually $100. Help me update this." — and Wilbur will find it and propose the fix.
Where to find Wilbur
Once Wilbur is enabled for your account, you'll see Ask Wilbur in the left sidebar, just below the home button.
You'll also see contextual "Ask Wilbur" prompts on certain setup pages — for example, while you're editing an AI prompt or related configuration.
To use Wilbur, you need to be an Admin or Manager on the account. Agency users can run Wilbur against any company in their agency.

What Wilbur can see
Wilbur reads from your live, company-scoped setup so it can answer questions with real evidence. That includes:
- Your account setup, enabled products, billing status, and usage
- Active AI agents and prompts across every channel — website texting, web chat, missed call text-back, AI voice, Yelp, Thumbtack, inbound email, Google LSA, and Email Gateway
- The company information your AI uses — description, products and services, pricing, contact details, and service areas
- Contacts, recent messages, lead inquiries, and conversation history
- Booking setup, booking attempts, calendars, and provider availability
- CRM integration settings and recent sync logs
- Custom lead source and webhook configuration, plus recent test and live events
- SMS and texting setup, 10DLC status, failed messages, opt-outs, and blocked numbers
- Notification subscriptions and follow-up schedules
- LeadTruffle help docs for setup and how-to questions
Diagnosing issues with Wilbur
Wilbur is a strong read-only diagnostic tool. Reach for it first when you're wondering things like:
- Which AI agents are active for this company?
- Which prompt is causing this bad response?
- Why did the AI ask for information the lead already gave us?
- Why didn't this lead inquiry book an appointment?
- Why does availability look wrong?
- Is texting set up correctly, and what does an SMS failure mean for this lead?
- Why did this CRM sync fail?
- Are notifications and follow-ups turned on?
- Can this user access the account?
- Where's the right help doc for this setup?
Wilbur works best when you give it a specific anchor — a contact's phone or email, a lead inquiry ID, a source or channel name, a prompt name, or a provider name.
Making updates with Wilbur
Beyond answering questions, Wilbur can make updates to your account. You describe the change in plain language, Wilbur figures out where it needs to happen, and then shows you exactly what it plans to do before touching anything.
Updates Wilbur can prepare and apply after your approval include:
- Prompt updates — targeted edits to a supported prompt, shown as a diff
- Company information — pricing, products and services, contact info, description, industry, website, and company-specific instructions
- Service areas — add, remove, replace, or adjust which areas the AI references
- Booking repair plans — coordinated changes across prompts, channel booking modes, and CRM booking fields
- Prompt cleanup plans — grouped edits when several prompts need to change together
- Notification settings — active state, medium, contact email or phone, subscribed events, or name
- Follow-up settings — enabled state, delay, quiet hours, or follow-up context
- CRM booking config — selected safe fields for Jobber, Housecall Pro, ServiceTitan, and Workiz
- User access — deactivate or reactivate a user, or change a role, with built-in safety checks
- Manual SMS lead creation — only with a first name, last name, phone, and certified written SMS consent
- Blocked phone numbers — add or remove account-level blocked numbers
- Custom lead webhook mapping — after a dry-run preview against a recent payload
Wilbur tends to prefer company information updates when a fact should apply everywhere (like stale pricing), and prompt edits for channel-specific wording or behavior. For prompts, expect small, surgical changes rather than full rewrites.
Reviewing and approving changes
When Wilbur recommends a fix, it creates a proposed change that includes review copy, risk notes, and a before-and-after view.
- Red text = something Wilbur is removing from your current setup
- Green text = something Wilbur is adding

Review the diff carefully, then hit Approve to apply — or decline it and ask Wilbur to adjust.
Previews that hand off to your team
A few actions generate a preview and supporting context, but Wilbur doesn't execute them directly today. Treat these as a head start for you or our support team:
- CRM test sync previews
- CRM sync retry previews for Housecall Pro, Jobber, and ServiceTitan
- ServiceTitan test booking previews
When Wilbur hands off to support
Wilbur will flag that a human should take over in situations like:
- Billing changes, refunds, cancellations, plan changes, or upgrade pricing
- Anything involving another company's account or cross-company access
- Requests for secrets, credentials, or API keys
- Urgent, service-impacting failures where the evidence isn't complete
- CRM or provider setup that needs external admin access
- Actions Wilbur doesn't support, such as sending login links
- Any case where a required detail is missing and can't be provided
When this happens, the chat shows you how to reach support, and the work Wilbur already gathered carries over as a starting point.
Tips for the best results
- Give Wilbur exact identifiers. A phone number, email, lead inquiry ID, prompt name, or provider name leads to sharper answers.
- Always review the diff and risk notes before approving. Approval applies the change immediately.
- If approval fails on a stale state, ask Wilbur to re-check and propose again — your setup may have changed since the original proposal.
- For booking problems, expect a multi-part fix. A prompt tweak alone often isn't enough when CRM or calendar readiness is the real issue.
- For webhook mapping, look for the dry-run preview before approving anything.
A few prompts worth trying:
- "Check why the AI asked this lead for their address again. Phone is
…." - "Find the prompt behind this Yelp reply and propose the smallest safe fix."
- "Check whether AI booking is ready and what's blocking it."
- "Trace lead inquiry
<id>and tell me why no appointment was created." - "The AI is saying our service call fee is $50 — it should be $100. Help me fix this."
Need help? Contact us at support@leadtruffle.com or message the team via the chat on this article or in-app.
Updated on: 14/07/2026
Thank you!