Introducing the Decisions API
OpenAI recently released a guide on its Decisions API. These APIs evaluate text and images and return typed predicates, choices or rubric scores. The main advantage is speed: the Decisions API is roughly ten times faster than the older Responses API for classification, routing and prioritization tasks.
Why speed matters for small businesses
In our experience with small and medium enterprises, processes such as customer reply handling, ticket management or internal request routing depend on quick decisions. A delay of a few seconds can lead to an unhappy customer or a staff member waiting for information. Faster responses help reduce wait times and improve overall experience.
How a Decisions API works
A Decisions API receives either a text input (for example a support email) or an image (for example a photo of a damaged product). The model analyses the content and returns a structured result, for example:
- predicate: a boolean statement (e.g. "urgent request").
- choice: a selection among predefined options (e.g. "assign to sales team").
- rubricScore: a numeric score on a scale (e.g. 0-100 for priority).
Each result includes a short explanation, allowing you to verify it without deep model knowledge.
Practical examples for small businesses
1. Support request routing. An online shop receives many support emails. With the Decisions API we can automatically classify emails into categories ("order", "return", "technical issue") and assign them to the right department. The result is a lower average response time and fewer routing errors.
2. Internal ticket prioritization. A marketing agency manages several projects at once. By analysing Slack messages or project notes, the API can assign a priority score based on keywords such as "urgent", "deadline" or "blocked". Managers can instantly see the most critical tickets.
3. Product image evaluation. A retailer wants to verify that supplier photos meet quality guidelines. The API can return a compliance score and a choice ("accept", "request revision"). This automates a task that would otherwise require hours of manual review.
What to evaluate before adopting the Decisions API
Before integrating this technology, we recommend checking:
- Input data quality. Short, clear texts give more reliable results. If emails are long or contain informal language, you may need to pre-process the text.
- Choice definition. Output options (for example ticket categories) should be well defined and limited; otherwise the model may return ambiguous results.
- Operational costs. The Decisions API is faster, but pricing is based on tokens processed. Estimating monthly request volume helps understand budget impact.
- Human oversight. For critical decisions (for example payment refusals) keep a manual review step, at least during the early adoption phase.
Concrete steps for a pilot
1. Identify a low-risk use case, such as classifying newsletter sign-up emails. 2. Create a test endpoint using the development credentials provided by OpenAI. 3. Define the output categories and prepare a small set of example prompts. 4. Measure response times and compare them with your current method. 5. Assess accuracy by comparing automatic decisions with human ones.
Conclusion
The OpenAI Decisions API is a practical tool for small businesses that want to automate classification, routing and prioritization decisions. The improved speed allows you to cut waiting times and free staff for higher-value work. As always, we suggest starting with a pilot, monitoring results and fine-tuning settings before scaling to more critical processes. If you would like to explore how to integrate these APIs into your workflows, we are happy to discuss a tailored approach.