What we build in the OpenAI API
Strong on classification and extraction, with a good range of model sizes for cost control. One of two providers we build with, chosen per task on measured accuracy rather than preference.
The OpenAI API is strong on classification and extraction, with a range of model sizes wide enough that model selection is the main cost lever. It is one of two providers we build with, chosen per task on measured accuracy rather than preference, and we build so the provider can be swapped without rewriting the surrounding workflow.
We build on it for high-volume classification on a small, cheap model, for structured extraction using function calling, for semantic search over your own documents, and for transcription feeding a downstream workflow. We do not treat its output as trustworthy without validation, and we pin model versions, because deprecations happen in a fast-moving market. It is priced per token, paid directly to OpenAI, and we take no margin.
What it is, and where it sits
OpenAI's API is the other AI provider we use. In practice the choice between it and Claude is made per task by measurement, not by allegiance, and we build so the provider can be swapped without rewriting the surrounding workflow.
Where it tends to fit: classification and extraction at low cost on the smaller models, function-calling patterns, and workloads where the model tier can be tuned tightly against a measured accuracy floor.
What we build in it
- Inbound message classification at volume, on a small, cheap model
- Field extraction from semi-structured text into typed output
- Lead qualification from free-text inquiry into structured CRM fields
- Content summarization for internal digests
- Semantic search over your own documents using embeddings
- Transcription workflows feeding downstream automation
How it is priced
Per token, input and output, varying substantially by model. Paid directly to OpenAI; we take no margin.
| Model | What you pay for | Best when |
|---|---|---|
| Per token, by model | Tokens in and tokens out, at a rate that varies substantially by model. | Always. Model selection is the meaningful lever. |
| A small model | The lowest per-token rate on offer. | High-volume classification and extraction. We benchmark the small model first and only move up if measured accuracy requires it. |
| Embeddings | Per token for the embedding call. | Semantic search. Cheap enough to be a rounding error for most SMB corpora. |
The gap between the smallest and largest models is large enough that running a classification task on an oversized model is one of the more common sources of unnecessary spend.
When to choose it, and when to think twice
Under the commercial and enterprise terms of both providers we use, your data is not used to train their models. If your compliance position requires that nothing leaves your infrastructure at all, we can work with self-hosted open-weight models instead, at some cost in capability. That gets decided during scoping, not afterwards.
We hold no reseller agreement with this vendor, take no affiliate commission, and do not resell licenses. You buy your own subscription directly. It is the only arrangement under which our recommendation means anything.
Want the cost model run on your numbers?
Tell us your workflows and volumes and we will show you the three-year cost across every option.
You will talk to Fayez or Kareem. Nobody else.