Technology Strategy: Scaling Enterprise AI
Evaluate investment priorities, operating models, and technology architecture to transition an enterprise software company into an AI-driven business.
Context
You are a Strategy Analyst at a B2B software company with an annual recurring revenue (ARR) of $120M. Enterprise clients want AI features integrated into their workflow systems. Management is currently considering allocating a large budget to build a proprietary large language model (LLM) from scratch, believing that this move will provide a long-term competitive advantage. The Chief Executive Officer has asked you to evaluate more efficient alternative AI strategies and present recommendations before tomorrow morning's meeting.
Your Task
Present your AI strategy analysis and options to the Chief Executive Officer. Your response must contain exactly three sections:
- Recommendation — one decision.
- Impact — business numbers and their impact.
- Implementation Plan — concrete steps (owner, sequence, timeline).
Available Data
Management has gathered cost and schedule estimates for three AI integration options:
| Strategy Option | Approach Description | Upfront Cost (Setup) | Annual Cost (OpEx) | Time to Launch | Projected ARR Growth |
|---|---|---|---|---|---|
| 1. Proprietary Foundation Model | Training a proprietary LLM from scratch with internal data | $25.0M | $14.0M | 18–24 months | +15% |
| 2. Fine-Tuning Open Model | Customizing an open-source model using RAG techniques | $3.5M | $4.2M | 4–6 months | +25% |
| 3. Third-Party API Integration | Connecting applications directly to commercial model APIs | $1.2M | $6.8M | 2–3 months | +18% |
Most directors initially supported Option 1, believing that owning a proprietary model is the only way to protect the business from competitors.