The AI Project Pricing & Scoping Playbook Nobody Wrote (But Every Dev Consultant Needs)

Short summary

AI consulting projects require different pricing than traditional software: quality is subjective, token costs are hidden, and model updates break tuned prompts. The solution: split work into zones (fixed core, hourly prompt engineering, retainer unknowns), lock evaluation sets with pass-rate thresholds upfront. Use contract clauses protecting against model provider changes post-delivery.

• •Split projects into three zones: fixed-price core, hourly prompt engineering, retainer-based operational unknowns
• •Lock evaluation sets and pass-rate thresholds with clients before work begins to prevent scope creep
• •Add contract language protecting against model provider updates and include infrastructure cost breakdowns

Generated with AI, which can make mistakes.