AI for business: May 2024

Terasu Rising · October 18, 2026 · 1 min read

In May 2024, the leading AI providers made their models faster, cheaper to use and able to work with voice, images and very long documents.

What happened

OpenAI released GPT-4o (May 13). The model handled text, images and audio in one system and responded quickly enough for natural spoken conversation. OpenAI also said it would be faster and cheaper through its API than the previous GPT-4 Turbo model.

Google expanded Gemini at its I/O conference (May 14). Google introduced Gemini 1.5 Flash, a lighter model for fast, high-volume tasks, and announced that Gemini 1.5 Pro would handle up to two million tokens of input for developers on a waitlist. In practical terms, that is enough to analyze very large collections of documents in a single request.

What it meant for business owners

• Voice became a practical channel. Phone enquiries, which many smaller businesses handle manually, became a realistic candidate for AI assistance, with a person taking over for complex calls.
• Documents and images became usable data. Scanned invoices, photographs of equipment, contracts and lengthy reports could be read and summarized by the same system.
• Falling prices changed project economics. Tasks that were too expensive to automate at the start of the year became affordable.

A caution

Faster, more fluent systems are also more convincing when they are wrong. As AI took on customer-facing work, the need for review, clear escalation to a person and records of what the system said became more important, not less.

One action for owners

Identify the documents your team reads or retypes most often: purchase orders, contracts, delivery notes, applications. Measure the hours spent on them each month. These are strong candidates for an AI-assisted pilot with human review.

Previous: April 2024. Next: June 2024. See where else AI pays in Business and AI.