theLLMs

Last checked: 2026-05-25

Scope: Global. Sources checked as of 2026-05-24.

AI draft model: gpt-5.4-mini

AI review model: llm-editor (deepseek-v4-pro)

Key terms:

>_ token >_ function calling / tool use >_ GPU (Graphics Processing Unit)

AI feature unit economics: cost per user, task and successful answer

Quick answer

AI unit economics means measuring what an AI feature costs per meaningful outcome: per active user, per task attempted, per successful answer, per ticket deflected or per decision reviewed. Token spend is only useful when tied to value and failure rate.

Why this matters

A team can celebrate low token cost while losing money on retries, human review, escalations or poor completion. Another team can panic about a high model bill even though the feature saves expensive staff time. The useful number is not raw spend; it is spend per useful result. The practical danger is not usually that a team misunderstands the academic definition. The danger is that the team makes a buying or architecture decision from a demo-sized understanding, then has to unwind it after users, documents, policies and invoices become real.

A useful operator view asks three questions. First, what decision does this capability support? Second, what evidence would make the answer trustworthy? Third, what will happen when the evidence is missing, stale, private, expensive or ambiguous? If the article does nothing else, it should push the reader away from magic-word thinking and toward those operating questions.

The practical model

Think of the feature as a small system rather than a model call. There is an input, some context, a decision rule, an output, a cost, a failure mode and usually a human who inherits the mess when the system is wrong. The model may be the most visible part of the workflow, but it is rarely the only part that determines whether the workflow works.

For an early build, the aim is not perfection. The aim is a bounded version that can be inspected. That means the team should know what data entered the system, why the answer was produced, how much the attempt cost, where the answer should be checked, and when the system should refuse, escalate or fall back.

Decision framework

Use this as the first-pass checklist before buying a tool, switching models or publishing a feature:

  • Define the unit: user, session, task, successful answer, resolved ticket, approved draft or completed workflow.
  • Track attempts and outcomes separately. Failed attempts still cost money.
  • Include hidden AI costs: retries, validation calls, fallback models, logs, evals and human review.
  • Estimate value per successful unit. Deflected support tickets, faster research or better conversion each have different economics.
  • Set guardrails before launch: max cost per task, max retry count and human-escalation threshold.

If the team cannot answer these checks in plain language, it is not ready for a bigger implementation. It may still be ready for a prototype, but the prototype should be labelled as a learning tool rather than a production assumption.

Worked example

An AI support assistant costs £0.04 in model calls per conversation. That sounds cheap. But only 45% of conversations are resolved without human help, and failed attempts add £0.02 in retries plus staff time. Cost per successful deflection is closer to £0.13 before staffing. If a human-handled ticket costs £3, the feature is still attractive. If the same assistant handles low-value website FAQs where self-serve search already works, it may be waste.

The important point is not the specific vendor or model. The useful pattern is to decompose the workflow. Ask what is retrieved, what is generated, what is validated, what is cached, what is logged, and what is handed to a human. That decomposition is where most cost, quality and safety decisions live.

Where teams get it wrong

  • Reporting blended token spend without success rate. It hides expensive failure modes.
  • Ignoring output length. Verbose answers can double cost without improving outcomes.
  • Counting automation as savings before measuring whether humans still rework the result.

A quieter failure mode is overfitting to launch week. The team tunes a prompt, route or model choice against a small set of internal examples, then assumes the result will hold when users ask shorter questions, upload worse files, use different language, or hit the feature from a mobile connection. The fix is not to make the first version huge. The fix is to keep a small evaluation set and review failed cases deliberately.

What to measure before scaling

At minimum, track four numbers: volume, success rate, unit cost and review burden. Volume tells you whether a small flaw will become a large one. Success rate tells you whether the feature is doing useful work rather than producing attractive output. Unit cost connects quality to budget. Review burden shows whether humans are truly being helped or simply moved downstream.

For higher-risk features, add sampled qualitative review. Read the bad answers. Read the boring answers too. Boring high-volume cases often contain the biggest savings, while rare edge cases often contain the biggest risk. The operating posture should be: measure enough to know whether to continue, not so much that evaluation becomes theatre.

Stable advice versus volatile claims

The stable advice is architectural: separate evidence from generation, exact lookup from fuzzy matching, and model capability from product reliability. The volatile claims are provider-specific: prices, model rankings, context limits, cache discounts, supported file types and benchmark standings. Those should be checked near publication and dated in the page.

Avoid phrases like “the best model” unless the article immediately says “for what workload, on what date, under what constraints”. A model can be best for a leaderboard and wrong for a workflow. A cheap model can be expensive if it causes retries. A strong model can be a poor fit if the data terms, latency or tooling do not match the product.

Reader checklist

Before committing, the reader should be able to write a one-paragraph operating note:

  • The task this feature is allowed to do.
  • The evidence or input it is allowed to use.
  • The condition where it should ask for help or refuse.
  • The cost metric that would make it unattractive.
  • The review process that catches bad outputs.
  • The date when assumptions should be rechecked.

That note is deliberately small. If it cannot be written, the problem is still fuzzy. If it can be written, the team has a starting point for a prototype, procurement conversation or editorial recommendation.

Sources and evidence notes

Sources used, checked 2026-05-27:

Stable concepts: retrieval quality, prompt length, output length, access control, evaluation design and review workflow do not disappear when a provider changes its models. The exact model names, prices, cache discounts, rate limits, benchmark rankings and feature availability are volatile. Editors should re-check live provider pages before publishing any hard number or ranking claim.

No hands-on claim: this draft uses accepted briefs and public documentation only. It does not claim that the site ran proprietary benchmarks, production traffic tests or vendor bake-offs.

What would change this advice

This advice should be revisited if a provider changes the API contract, pricing unit, cache semantics, supported media type, benchmark methodology or data-retention terms in a way that affects the decision. It should also change if the site later keeps a public evaluation artifact for this topic; at that point the article can cite the retained test directly rather than speaking only from public docs and operator logic.

Change Log

  • 2026-05-27: Added direct source URLs to all named providers and services; added Change Log section. Content unchanged.