Boolean & Beyond  —  Field Notes
Nº 03 · Implementation Guide · 2026
Field Notes: Implementation Guides

A demo is a vibe. An eval is a contract.

Every LLM feature that failed in production passed a demo first. This is the method we use to build evaluation suites that gate releases: a golden set drawn from real questions, graders you can trust, and a regression loop that turns every bad answer into a permanent test.

AI Engineering · How-To/Bangalore · Coimbatore/read time: about 10 minutes
30–50
cases to start, not 5,000
4
numbers tracked per release
1 rule
no pass, no ship
~30 min
per week of upkeep
§01
Why this guide exists

The failure mode.

Large language model (LLM) features fail in a specific, predictable way. The demo is impressive, the first week is quiet, and then the long tail arrives: questions phrased differently, edge cases nobody rehearsed, a model update that shifts behaviour overnight. Without measurement, every one of those surprises reaches a user before it reaches you.

An evaluation suite inverts that. It is a set of real questions with known-correct answers, run against the system on every change, with hard numbers that decide whether a release ships. Teams that have one iterate quickly and sleep well. Teams that do not have one re-test by hand, ship on hope, and find out in support tickets. We covered why this matters in the pillar essay; this guide is the how. Read the pillar essay.

§02
What to measure

The four numbers.

Before building anything, agree on what a passing release means. We track four numbers, and only four, on every run.

01
Answer correctness
did we get it right

The share of golden cases where the answer matches the expected result, judged by the grader assigned to that case.

02
Grounding accuracy
did we prove it

The share of answers whose citations actually support the claim. An answer that is right for the wrong reason is a latent failure.

03
Refusal correctness
did we know our limits

The system should refuse when the knowledge is missing. Track both false confidence and false refusal; both erode trust.

04
Latency, p95
was it usable

Quality that arrives after the user gives up does not count. Watch the tail, not the average.

§03
Step 1

Build the golden set from real questions.

A golden set is a list of question-and-correct-answer pairs that represents what your users actually ask. Do not invent questions in a workshop. Harvest them: support tickets, sales call notes, search logs, the questions your own team asks the system in its first week. Thirty to fifty cases is enough to start, provided each one is real.

Every case needs the same anatomy: the question as a user would phrase it, the expected answer or the facts it must contain, the source documents that answer should be grounded in, the grader that will judge it, and a tolerance note for what counts as acceptable variation. Write the expected answer with the source open, not from memory.

case-041golden set · one case
questionWhat did we agree with Acme on payment terms in the March review?
expectedNet 45, revised from net 30, effective April; approved by finance
sourcesacme-march-review.transcript · commercial-terms-v4.pdf, §2.1
graderLLM-as-judge, rubric: facts present, no invented terms
tolerancewording free; the three facts and the citation are mandatory
Fig. 01  ·  The anatomy of one golden case: five fields, written with the source document open
§04
Step 2

Choose graders you can trust.

A grader is whatever decides that an answer passed. Use the cheapest grader that can actually judge the case, and escalate only when the case demands it.

Deterministic checks
exact and structural
  • ·Exact match for codes, numbers, and names
  • ·Schema validation for structured output
  • ·Citation-presence and link-resolves checks
  • ·Fast, free, and never wrong about what they check
Rubric scoring
facts as a checklist
  • ·Expected facts listed per case, checked one by one
  • ·Catches partial answers and invented extras
  • ·Human-writable, machine-checkable
  • ·The workhorse for knowledge questions
LLM-as-judge
judgement at scale
  • ·A second model scores against the rubric
  • ·Needed for tone, reasoning, and open-ended answers
  • ·Known biases: position, verbosity, self-preference
  • ·Calibrate against 20 human-graded cases first

The judge deserves suspicion. Before trusting an LLM-as-judge, run it over twenty cases a human has already graded and measure the agreement. Use a different model family for judging than for answering, keep the judge prompt versioned like code, and spot-check a sample of its verdicts every month. A judge you never audit is just a second opinion you never questioned.

§05
Step 3

Wire it into CI and let it say no.

Run the full suite on every change that can shift behaviour: prompts, models, retrieval settings, chunking, tool definitions. Publish the four numbers as a report artifact, compare against the last accepted run, and fail the pipeline on regression. The suite earns its keep the first time it blocks a bad release at ten in the morning instead of ten at night.

$ eval run --suite=release --cases=41 --baseline=v1.6.2
answer correctness39/41-2FAIL
grounding accuracy40/41±0PASS
refusal correctness41/41±0PASS
latency p951.4s+0.1sPASS
case-017, case-034 regressed vs v1.6.2 → release blocked
Fig. 02  ·  A release blocked by two regressed cases. This is the suite doing its job.
§06
Step 4

Close the loop: every bad answer becomes a test.

1
Capture
same day

A user flags a wrong answer. Log the exact question, the answer given, and the sources retrieved, straight from the trace.

2
Triage
weekly, 30 minutes

Decide what failed: retrieval missed, grounding ignored, the model guessed, or the knowledge simply does not exist.

3
Encode
ten minutes per case

Write the golden case with the correct answer and source, and add it to the suite. The failure is now permanent memory.

4
Re-run
every release

The fix is proven by the suite going green, and that question can never silently regress again.

§07
Effort, honestly

Sizing and cost.

The first version is a two-to-three day build for one engineer: harvest thirty to fifty questions, write the expected answers with sources open, wire deterministic and rubric graders, and add the CI job. If you use an LLM judge, add a day to calibrate it against human-graded cases. From there, upkeep is the weekly triage half-hour.

Running cost is small and worth paying on every release: a 40-case suite with a judge model costs a few rupees per run at current API prices, and a private judge model makes it effectively free. Against one production incident traced in front of a client, the suite pays for itself the first month. When regulators or enterprise buyers ask how you verify your AI, and under the DPDP Act's due-diligence duties for significant fiduciaries they will, the suite's history is the evidence. See the DPDP Act field notes.

Case study · proof, not theory

An evaluation suite gating every release of an enterprise agent.

The enterprise AI agent we built and handed over ships behind exactly this loop: a golden set harvested from real workflows, rubric and judge graders, and a CI gate that has blocked regressions before users ever saw them. The client's own team runs it today.

→ Read the case study
§09
Field warnings

Five ways eval suites die.

×1The five-case suite
too small to mean anything, green by coincidence; harvest real questions until it can fail
×2Vanity metrics
a single blended score hides which capability broke; keep the four numbers separate
×3The judge grades itself
the answering model marking its own homework inherits its own blind spots; use a different family and calibrate
×4The stale set
a suite frozen at launch tests last year's product; the weekly loop is the suite's pulse
×5Advisory-only results
a report nobody must act on changes nothing; the suite must be able to block the release
§10
Asked first

Questions, answered.

01What is an LLM evaluation suite?+

A set of real questions with known-correct answers, run automatically against your AI system on every change. Each case is judged by a grader, the results roll up into a small set of numbers, and the release ships only if those numbers hold. It does for AI behaviour what a test suite does for code.

02How many test cases do we need to start?+

Thirty to fifty real cases beat thousands of synthetic ones. Harvest them from support tickets, sales calls, and logs so they reflect what users actually ask. The suite then grows organically: every bad answer reported in production is encoded as a new case, so coverage concentrates exactly where the product hurts.

03What is a golden set?+

The curated list of question-and-expected-answer pairs at the heart of the suite, each with its source documents, an assigned grader, and a tolerance note. It is called golden because the answers are verified against sources when written, not recalled from memory, and the set is version-controlled like code.

04What is LLM-as-judge and can we trust it?+

LLM-as-judge means using a second model to score answers against a rubric, which is how evaluation scales past what humans can grade by hand. It is trustworthy only after calibration: run it over cases humans have already graded, measure agreement, use a different model family than the one answering, and spot-check verdicts monthly. Uncalibrated judges have known biases toward verbose and familiar-sounding answers.

05Should we buy an eval tool or build our own?+

Start with the loop, not the tool. The golden set, the graders, and the ship-or-block rule are the substance, and they fit in a repo with a CI job in a few days. Dedicated platforms earn their place once volume grows, and because the golden set is plain data you can adopt one later without starting over. What does not work is buying a dashboard and skipping the discipline.

06How do we evaluate RAG systems specifically?+

Split the measurement in two. Retrieval is checked by whether the right documents were fetched for each golden case, and generation by whether the answer matches the expected facts and cites those documents. When a case fails, the split tells you immediately whether to fix search or prompting, which are different repairs with different owners.

07How do we evaluate AI agents that take actions?+

Grade trajectories, not just final answers. Each golden case defines the expected outcome plus constraints on the path: which tools may be called, in what order, and what must never happen, like a destructive action without confirmation. Deterministic checks on tool calls plus a judge on the final state cover most agent behaviour, and refusal correctness matters double because an agent that guesses acts on its guess.

08How often should the suite run, and what does it cost?+

On every change that can shift behaviour, which in practice means every pull request touching prompts, models, retrieval, or tools, plus a scheduled run to catch provider-side model updates. A 40-case suite with an LLM judge costs a few rupees per run on hosted APIs, and near nothing with a private judge model. The weekly triage half-hour is the real cost, and it is the highest-leverage half-hour in the AI roadmap.

Colophon

Ship when the suite says so.

Bring one workflow and twenty real questions. In one conversation we can usually sketch the golden set, pick the graders, and size the build for your stack.

Start a conversation →
How to Build an LLM Evaluation Suite: Step-by-Step Guide | Boolean & Beyond | Boolean & Beyond