Credit Usage
TL;DR: The AI Assistant operates on a credit-based system where different operations consume different amounts based on computational complexity. Document preprocessing is the most expensive, but benefits all team members.
Overview
The AI Assistant operates on a credit-based system. Credits are consumed when you use AI-powered features, with different operations requiring different credit amounts based on their computational complexity and resource requirements.
Important: AI Search and Large Projects
When AI Search is enabled for a project, it automatically scans and preprocesses every PDF file. This is a very expensive operation. If you have a large project with many drawings and documents, it could consume your credit budget rapidly.
To control costs: Turn off AI Search at the beginning and manually queue only the files you need for processing.
Credit Costs by Operation
Understanding how credits are charged helps you optimize your AI usage and manage your credit budget effectively. Here's a detailed breakdown of credit costs for different AI operations:
Page Preprocessing
Cost: Approximately 1 credit per page
Each PDF page must be preprocessed before the AI can analyze it. This involves extracting text, understanding layout, identifying elements, and creating a searchable index. For a 100-page drawing set, this means approximately 100 credits just for preprocessing.
When it happens: Automatically when AI Search is enabled, or manually when you queue specific files for processing.
Document Search and Analysis
Cost: 5 to 20 credits per operation
Searching across project documents and extracting relevant information is the most expensive operation because it requires:
- • Semantic understanding of your query
- • Cross-referencing multiple documents
- • Context analysis and relevance scoring
- • Data extraction and synthesis
However, once documents are analyzed and preprocessed, all team members working on the project can utilize the preprocessed data to answer queries without additional preprocessing costs. This makes document analysis a worthwhile investment for collaborative projects.
BOQ Generation and Editing
Cost: Less than 5 credits per operation
Creating BOQ items, editing descriptions, restructuring your bill of quantities, or organizing items uses minimal credits. This includes:
- • Generating new BOQ items from measurements
- • Editing item descriptions
- • Restructuring BOQ hierarchies
- • Formatting and organizing items
General Questions
Cost: Less than 5 credits per operation
Asking about construction methods, terminology, QS processes, or general knowledge consumes minimal credits. Examples include:
- • "Explain ceiling coffers in simple terms"
- • "What is the standard method for measuring formwork?"
- • "How do I calculate concrete volume for a slab?"
ML Takeoff (Machine Learning Takeoffs)
Cost: Less than 5 credits per operation
Automated measurements and quantity extraction from drawings using machine learning. This includes:
- • Automated element detection
- • Quantity calculations
- • Measurement extraction from drawings
Optimizing Credit Usage
For Large Projects
If you're working with a large project containing hundreds of pages:
- Disable AI Search Initially – Turn off automatic AI search when creating the project
- Identify Critical Documents – Determine which documents will be frequently referenced
- Manually Queue Files – Process only the essential documents for AI search
- Process in Phases – Add more documents as needed throughout the project lifecycle
For Team Collaboration
Document preprocessing is a one-time cost that benefits all team members. Once a document is preprocessed, anyone on the team can query it without additional preprocessing charges. This makes the upfront investment worthwhile for collaborative projects.
Cost-Benefit Analysis: While document search and analysis is expensive upfront (5-20 credits per operation), the investment pays off when multiple team members can query the same preprocessed documents without additional preprocessing costs. For a 10-person team querying the same specs document 20 times, you pay the preprocessing cost once, not 200 times.
Best Practices
- ✓ Start with AI Search disabled for large projects
- ✓ Preprocess only the documents you'll frequently reference
- ✓ Use general questions (low cost) before document searches (higher cost) when possible
- ✓ Leverage preprocessed documents for team collaboration
- ✓ Monitor your credit usage regularly
- ✓ Process drawing sets in phases as you work through them
- ✓ Use BOQ generation and ML takeoffs liberally—they're low cost
Monitoring Credit Usage
Keep track of your credit consumption to ensure you stay within budget and optimize your usage patterns. AedisPro provides tools to monitor your credit balance and usage history.
Planning Tip: For a typical medium-sized project (50-100 pages of drawings plus specs), budget approximately 150-200 credits for initial preprocessing if you enable AI Search for all documents. Additional query costs will depend on usage patterns.