AI is a term used inflationarily in software marketing. Almost every new tool carries the AI label. This article explains transparently how AI quoting in PlotonIQ actually works.
What AI quoting means
Component 1: Language models for quoting logic
When you describe van partial graphics, both sides, logo plus address, Oracal 751 — the system understands: vehicle type, scope, material, service components. The language model translates free text into a structured quote.
Component 2: Structured calculation logic
What surfaces does a VW Transporter T6 have? How much film is needed with 15% waste? How long does installation take? What does Oracal 751 cost today? Rule-based industry knowledge — no guessing.
The process
- Input: free text or voice note
- Language processing: LLM identifies type, material, scope
- Database query: current prices and benchmarks
- Calculation: area, material quantity, labour time, travel, margin
- Output: complete quote with all line items
- Your check: review and adjust if needed
Material recognition
The AI knows catalogues from Oracal, 3M, Avery, Mactac, Metamark, Neschen, Sihl. For Oracal 651 white: current buy price, roll widths, waste benchmarks, whether laminate is needed. Unknown material: system asks or suggests the nearest equivalent.
How the AI learns from jobs
Your operational data: post-calculations feed in. The system learns: your film waste runs at 18%, not 15%. Your team takes 10% longer on vehicle graphics than the industry average. Industry-wide patterns: aggregated and anonymised. Your data stays your data.
AI survey: the visual component
Four models in cascade: Claude Vision (classification), Gemini Flash (localisation), Gemini 2.5 Flash (context), Segment Anything Model (pixel-accurate polygons). You enter a reference measurement, AI calculates all areas. More on photo measurement
What AI quoting cannot do
- Very rare materials outside the known catalogue
- Very unusual job structures with 10+ service types
- Client-specific special terms — must be entered manually
- Local planning restrictions
Practical example: complete quote in 45 seconds
Input: facade signage for bakery Smith, company sign 200x60cm white dibond with cut vinyl lettering, logo in colour, ground-floor installation, design still to be done. Output after 8 seconds: dibond 48 EUR, lettering and logo 67 EUR, design 1.5h 112.50 EUR, installation including travel 178 EUR. Net total 405.50 EUR, gross 481.55 EUR. Duration: 53 seconds. Traditional method: 30-45 minutes.
Especially valuable for small shops
No dedicated quoting department. Every saved quoting hour is more production or sales time. The AI learning effect over jobs is especially valuable for small teams because there is less accumulated experience in the team.
Conclusion: AI quoting is a tool, not a miracle
For 80% of standard sign-making jobs, AI quoting delivers in 45 seconds a result that takes 45 minutes manually. For edge cases: still your judgement. Together: the most effective quoting tool available for sign makers today. More on AI quoting | Sign making quoting guide | Start free — 9 EUR/month