As AI adoption accelerates, the cost and volatility of AI investments are becoming significant challenges for organizations. Gartner estimates that GenAI costs could vary by as much as 500% to 1000%, with vendors raising prices by up to 30% as they integrate GenAI capabilities. This unpredictability is driven by factors such as data preparation, infrastructure needs, computational power, talent scarcity, token costs (price per NLP interaction), and regulatory requirements.

For those of us in mid-sized enterprises (MSE) who began exploring AI use cases a few years back, we may not yet be facing this level of cost volatility and cost spikes. While our models and applications haven’t fully scaled, our first-mover advantage lies in AI literacy and capability building. By diving in early, we engaged in hands-on, often scrappy AI projects, frequently co-innovating and co-investing with vendor partners. We built foundational machine learning models, applied large language models (LLMs) to generate human-readable results, and enabled interaction with existing models, all using platforms already familiar to our users. Successful use cases have generated financial benefits, with vendors offering additional resources to showcase our shared achievements.

My advice to MSEs:

• Dive in NOW – Leverage existing platforms, data, and cloud capabilities.
• Be Scrappy – Test and learn with vendor ecosystems; seek co-investment and co-innovation.
• Invest in AI Literacy and Capability Building – Consider AI boot camps, AI leadership day, or AI executive retreat
• Prepare to scale Enterprise-Wide – Establish governance, prioritize investment, and expand on successful use cases