Trunnion AI Engine
16+ AI features. All built into Zennvue.
Most CRMs bolt AI on as an afterthought. Zennvue is built around it. Every workflow, from lead scoring to proposal generation to revenue forecasting, runs through the Trunnion AI engine. We built it on Anthropic Claude, the model we picked for long-context reasoning on event vendor workflows. No multi-vendor shopping, no fallback chain. One model, one deep integration, one privacy posture.
16+ AI features
Built on Anthropic Claude
No-training configuration
16+
AI features built in
Claude
Anthropic foundation model
<2s
Avg AI response time
0%
Of your data trains models
Watch It Work
One inquiry in. Three AI outputs in seconds.
When a wedding inquiry hits your inbox, Trunnion AI does the work you would otherwise do manually. Score, draft, plan. All in under 2 seconds.
📥 Incoming Inquiry
Emma & Daniel R. · June 15 wedding · 180 guests · Austin, TX · $5K to $8K budget · Pinterest mood board attached
Routing through Trunnion AI...
Anthropic Claude
🎯 Lead Score
~340ms
94
/100
High Priority
Verified domain
Date specific
Budget aligned
180 guests = sweet spot
📝 Proposal Draft
~1.4s
Hi Emma & Daniel,
Thank you for reaching out about your June 15 wedding. A 180-guest celebration in Austin sounds incredible, and the inspiration in your mood board is right in my wheelhouse. I would love to capture your day. I have your date open and put together a custom package quote based on what you described.
Thank you for reaching out about your June 15 wedding. A 180-guest celebration in Austin sounds incredible, and the inspiration in your mood board is right in my wheelhouse. I would love to capture your day. I have your date open and put together a custom package quote based on what you described.
💌 Follow-Up Sequence
~520ms
Day 0
Initial response with portfolio link
Sent
Day 3
Custom package proposal + booking link
Scheduled
Day 7
Soft check-in if no response
Pending
The Engine
Built on Anthropic Claude.
Trunnion AI is Viceroy NM's in-house AI workflow framework. Instead of shopping across multiple AI vendors hoping one will work, we picked the best model for event vendor reasoning and built our entire stack around it.
One model. Picked carefully. Integrated deep.
Trunnion AI is the orchestration layer on top of Anthropic Claude. It owns prompt engineering, brand voice profile management, output validation, classification routing, and the tenant scoping that keeps your data inside your account. Anthropic provides the foundation model that powers the reasoning. Together they deliver 16 production AI features.
The other approach, shopping across Claude plus GPT plus Gemini, sounds smart but trades depth for breadth. We chose depth. Faster iteration, predictable output quality, one set of safety properties, one Data Processing Addendum, one privacy posture. The single best model for the job, fully owned.
-
Best Model for Event Vendor WorkflowsClaude excels at long-context reasoning, brand voice mimicry, and structured proposal generation. Exactly what wedding and event vendor AI requires.
-
Sub-2-Second ResponseStreaming responses, intelligent caching, parallel API calls for multi-output workflows. Snappy UX even under load.
-
Graceful DegradationIf the Anthropic API is rate-limited or degraded, Trunnion queues with exponential backoff and surfaces status to the user. No silent failure.
-
Privacy by DesignNo-training configuration on the Anthropic API. Your data never trains Anthropic's foundation models. All calls run server-side under a signed DPA.
1
Request arrives at Trunnion AI Router
↓
2
Router applies tenant scoping + brand voice profile
↓
3
Prompt sent to Anthropic Claude API
↓
4
Response validated, post-processed, and delivered to your CRM
The 16+ AI Features
Every workflow. AI-augmented.
Lead & Response AI
Score, route, and respond to every inquiry instantly
AI Lead Scoring
Every inquiry scored 1 to 100 with conversion probability and reasoning. Tag-based explanations.
AI Auto-Response
Drafts responses in your brand voice. Approve before sending or auto-send for high-confidence replies.
Sentiment Analysis
Detects tone in every inbound message. Flags frustrated, excited, or hesitant clients in real time.
Call Routing & Triage
Inbound calls routed by AI based on inquiry context, vendor availability, and lead priority.
Content Generation AI
Generate proposals, contracts, and emails in your voice
AI Proposal Generator
Custom proposals from a single inquiry. Pricing, packages, terms, all formatted and ready to send.
AI Email Drafter
Draft outreach, follow-ups, and confirmations in your voice. Multi-step sequences in one click.
AI Contract Builder
Industry-specific contract templates auto-customized to event details, deposits, and cancellation terms.
Brand Voice Profiles
Train AI on samples of your writing. Every output sounds like you, not a chatbot.
Forecasting & Insights AI
See revenue, churn, and demand before they happen
Revenue Forecasting
30, 60, and 90-day projected revenue with confidence intervals. Updated as deals progress.
Churn Prediction
Flags clients at risk of canceling based on engagement patterns and inbound sentiment.
Demand Forecasting
Predicts capacity strain weeks in advance. Plan staff, gear, and partnerships before you scramble.
Pricing Optimization
AI compares your pricing against market data and conversion rates. Suggests price tier adjustments.
Operations AI
Smart matching, scheduling, and execution
AI Seating Chart
Optimal seating from relationships, dietary, and accessibility. Drag-and-drop refinement with explanations.
AI Vendor Matching
Couples get top 5 vendor matches per category. Vendors get qualified inbound leads with context.
Timeline Generator
Hour-by-hour wedding day timelines from event details. Vendor cues, prep windows, sync points.
Daily Briefing
Personalized morning briefing: priority leads, contracts to review, payments due, action items.
Your Data, Your Control
AI without the privacy tradeoff.
How we handle your data with AI
Most platforms use your data to train their models or share it across vendors. Zennvue does neither. Here are the rules we hold ourselves to.
No model training
Your inquiries, proposals, and client data are never used to train Anthropic's foundation models. Confirmed by the no-training configuration on the Anthropic API as deployed in our integration.
Single Data Processing Addendum
One AI vendor, one signed DPA. Easier to audit, easier to disclose to your own enterprise customers, easier to keep current.
Server-side only
All AI calls run server-side. Your client data never goes browser-to-Anthropic. Tenant scoping is enforced at the router layer.
Granular AI controls
Disable specific AI features per workspace. Studio and Enterprise plans get per-user controls. See Privacy Policy and Security for full disclosures.
AI FAQ
Common questions.
Trunnion AI is Viceroy NM's in-house AI workflow framework that powers every AI feature in Zennvue. It owns the orchestration layer: prompt engineering, brand voice profile management, tenant scoping, output validation, and request routing. The underlying foundation model is Anthropic Claude, accessed through the Anthropic API with a no-training configuration. Trunnion is the layer Viceroy NM built; Claude is the model that powers it.
Anthropic Claude, via the Anthropic API. We picked Claude because it leads on long-context reasoning, brand voice mimicry, and structured proposal generation, which are the workloads that matter most for event vendors. We do not currently route to OpenAI or Google models. If we add additional providers in the future, this Section and our Terms of Service Section 7 will be updated to reflect the change.
No. Your client data, inquiries, proposals, and contracts are never used to train Anthropic's foundation models. The Anthropic API is configured with no training on customer data, and that configuration is part of our integration. Brand Voice Profiles you train from your own samples are scoped to your account only and are not shared across organizations. All AI calls run server-side under a signed Data Processing Addendum.
Yes. Each AI feature can be toggled on or off per workspace in Settings → AI. Studio and Enterprise plans get per-user AI controls so you can give some team members AI access while others work without it.
Upload 5 to 10 samples of your writing (past emails, proposals, blog posts, social copy). Trunnion AI analyzes tone, vocabulary, sentence rhythm, and signature phrases, then encodes those patterns into a Brand Voice Profile that gets injected into every AI request from your workspace. Every AI-generated piece of content is drafted in your voice. You can train multiple profiles for different contexts (formal proposals vs. casual follow-ups).
Trunnion AI queues requests with exponential backoff and retries when capacity returns. Status is surfaced in the UI so users know an AI step is delayed rather than silently failing. For Enterprise customers, status incidents and resolution timelines are communicated per the SLA in Security Section 04. We do not currently route to alternate providers because doing so would mean different output quality, different privacy posture, and a different DPA, all of which we have deliberately consolidated on Anthropic.
Lead scoring accuracy depends on the quality of inquiry data. Inquiries with full details (date, guest count, budget, location, references) score with high confidence. Vague inquiries score lower with explanations. Each score includes the reasoning so you can override the AI when context warrants it.
Most AI features are included on the Professional plan ($99/month) and above. Some advanced features (brand voice profiles, churn prediction, pricing optimization) are limited to Business plan and above. Studio ($499/mo) adds per-user AI controls and team brand voice profiles. Enterprise ($1,999/mo) adds custom training on your historical data and dedicated success management. See full pricing breakdown.
16+ AI features. 14-day free trial.
Try every AI feature on the full Professional plan free for 14 days. No credit card required. See what Trunnion AI on Anthropic Claude can do for your event business.