CellaNova's self-learning AI engine delivers fully itemized estimates — materials, labor, travel, permits — straight from a plain-language job description.
Product Demo
From a fresh lead to a fully itemized quote — see how Pricing Assistant moves a job from inbox to estimate, then scroll on to walk through each stage.
Features
Built for signage and graphics teams who need accurate, consistent estimates every time — without spending hours on spreadsheets.
Extracts billable area, unit counts, waste factors, and per-step labor drivers directly from plain-language descriptions. No templates, no manual entry.
Every approved estimate makes the next one more accurate. REINFORCE-style gradient updates continuously adapt the model to your real pricing history.
Consistent retail/wholesale markup strategy backed by a live material database. Full line-item audit trail from raw cost to final quote — no black boxes.
Inbound requests are screened, scored, and queued for estimation. Focus on high-value jobs while the AI pre-qualifies every inquiry automatically.
Upload past proposals as PDFs. The RAG system extracts your institutional pricing knowledge and applies it to every future estimate.
Track MAE trends, reward signals, and model drift in real time. Human-in-the-loop controls let you course-correct the AI at any time.
The Journey
A complete pipeline for sign & graphics teams: capture every request, qualify it, price it with AI trained on your own history, and watch it sharpen with every job. Here's how work flows through Pricing Assistant.
Connect your inbox and Pricing Assistant watches it around the clock. The instant a request arrives, it's pulled into your Leads dashboard — screened, scored, and queued. Reply, ask for more detail, and prepare drafts without ever leaving the board.

When you're confident a request is worth pursuing, promote it with a single click. It moves onto your Production & Estimation board, where every active job is tracked from AI-ready to approved — so your pipeline is always current.

Pricing Assistant decomposes the scope, then calculates materials, production and install labor, fees and travel — applying your markup strategy and returning a complete line-item breakdown with the reasoning behind every number.

Upload your historical proposals and Pricing Assistant reads between the lines — learning how you price, what you mark up, and how you break down a total. Every future estimate draws on your own institutional knowledge.

In the Training Lab, review the retrieval evidence behind each estimate, adjust labor assumptions, and send reward signals. Shape the model to your way of working from day one — and keep refining it as your business evolves.

The Analytics board tracks accuracy, reward trends, and how the policy adapts as feedback comes in — so you always know exactly how Pricing Assistant is learning, and how much closer each estimate gets to your numbers.

Process
Type a plain-language description — dimensions, materials, install location, timeline. No special syntax required.
The pipeline decomposes scope, retrieves context from past proposals, and assembles a fully itemized breakdown with rate guardrails.
Approve or correct the estimate. Each correction feeds back into the learning system, improving accuracy on every future job.