Study Guide1,580 words

Stakeholder communication and lifecycle — reference

Stakeholder communication and lifecycle — reference

A decision-table companion. Each table names a decision, the documented input that settles it, and the claim it does not license.

1. Requirement or decision (LO1)

SentenceWhat it really isWhat it lets you check
"The assistant should be accurate"A wishNothing
"Use the fastest available model"A decisionNothing
"Use this named model"A decisionNothing
"F1 of at least 0.85 on a held-out set"A criterionWhether the release passes

Building a successful application starts with clearly defining success criteria and then designing evaluations to measure performance against them. The documented improvement is exactly this move: "good performance" becomes "accurate sentiment classification", and "the model should classify sentiments well" becomes an F1 score of at least 0.85 on a held-out test set.

The properties a criterion carries

PropertyThe documented demand
SpecificSay what you want to achieve, not that you want it to be good
MeasurableQuantitative metrics, or well-defined qualitative scales applied consistently
AchievableBased on benchmarks, prior experiments, research or expert knowledge
RelevantAligned to the purpose of the application and the needs of its users

⚠ Achievable is the one that gets skipped. Success metrics should not be unrealistic to current frontier model capabilities. A threshold nobody can hit is not ambition; it is a criterion built to fail.

⚠ One criterion is a finding, not a result. Most use cases need multidimensional evaluation along several success criteria, and the published list — task fidelity, consistency, relevance and coherence, tone and style, privacy preservation, context utilization, latency, price — is explicitly non-exhaustive.

2. Decisions and trade-offs (LO2)

FactorThe documented question
CapabilitiesWhat specific features will you need the model to have
SpeedHow quickly does it need to respond in your application
CostWhat is the budget, for development and production usage
EffortTrades intelligence for latency and cost within a single model

⚠ Tuning effort is often a better lever than switching models. A decision record that only weighed model swaps has left out the cheaper alternative.

Levers, and what each one trades

LeverGainedGiven up
Lower effortLatency and costIntelligence within that model
Fast modeUp to 2.5x higher output speedPremium pricing
Multi-model pairingMost tokens billed at the lower rateAn orchestration boundary to maintain

A defensible record carries the decision, the alternatives, and the evidence. The documented evidence step is to create benchmark tests specific to your use case — having a good evaluation set is the most important step in the process — then test with your actual prompts and data, then weigh performance and cost trade-offs.

⚠ Anthropic token cost typically dominates container infrastructure cost by an order of magnitude or more. A team optimising the hosting bill is working on the smaller number.

3. Service levels and expectations (LO3)

MeasurementWhat it covers
Baseline latencyPrompt processed and response generated, ignoring tokens per second
Time to first tokenFrom prompt sent to the first token of the response

⚠ Name the metric before agreeing the number. These are different intervals over the same request, and on a long answer they differ by a lot.

A service-level line needs three parts: the named metric, a threshold with the share of requests it covers (the documented multidimensional example attaches a percentage to its response-time criterion rather than stating a bare average), and a note of when it was measured. It is always better to engineer a prompt that works well first and then try latency reduction, because reducing latency prematurely might prevent you from discovering what top performance looks like.

Whose lifecycle governs

Where you deployWhose dates apply
Claude APIAnthropic's first-party commitment
Claude Platform on AWSThe first-party model lifecycle
Microsoft FoundryThe Claude API lifecycle schedule
Amazon BedrockBedrock sets its own lifecycle dates
Google CloudGoogle Cloud sets its own lifecycle dates

⚠ Every model ID, whether dated or dateless, has its own distinct deprecation and retirement schedule. The published date is a floor ("not sooner than"), not an appointment.

What a pinned model ID freezes

Fixed for the lifetime of the IDFree to change
WeightsThe request router
ConfigurationSafety classifiers
(Updates ship under a new ID)Sampling logic

⚠ Infrastructure updates occasionally produce minor differences in observable behavior even when the model ID and weights have not changed, and where you see unexpected behavioural differences on a previously stable ID, an infrastructure update is the most likely cause. A dateless ID is a pinned snapshot, not an evergreen pointer — the documentation names the opposite belief as a common misconception.

4. What a handover document answers (LO4)

DecisionWhy it is pre-launch
Session and state persistenceDefault local disk is lost on restart, scale-down or a move to another node
ObservabilityWithout telemetry you cannot see which tools ran, how long they took, or where a session stalled
Auth and secretsInbound at a gateway; outbound credentials injected by a proxy after the request leaves
Scaling and concurrencyBounded by how many subprocesses host RAM can hold
CostToken cost dominates infrastructure cost
Multi-tenant isolationShared filesystem state can leak one tenant's context into another's session

Known limitations to plan around

LimitationThe documented answer
No top-level session timeoutBound the tool-use round trips with a maximum turn count
Memory growth over long sessionsCap session length or recycle subprocesses periodically
Wide parallel-subagent fan-outs can hit rate limitsBreak work into smaller batches
No per-subagent wall-clock deadlineThe stall watchdog fires on no output; it is not a total-runtime deadline

⚠ The mirror error is the failure that leaves the run looking fine. When the SDK cannot deliver a batch to the store it drops the batch and continues the query — success, exit zero, record gone. Alert on these if store durability matters. Note the conditional: it is a design decision, not a default.

⚠ A session store mirrors transcripts only, not memory files or other working-directory artifacts, and the subprocess writes to local disk first with the store receiving a forwarded copy.

⚠ The published per-session memory figure is a floor, not the ceiling. Measure yours with a representative session at your target length under expected tool load, recording peak resident memory.

5. Lifecycle phases (LO5)

BoundaryWhat must cross itWhat is lost if it does not
Design → handoffDecisions, constraints, known limitations, environment assumptionsThe team rediscovers each one during an incident
Handoff → monitoringOwners, alerts, and which cost figure is authoritativeNobody is paged, and an estimate is read as a bill
Monitoring → iterationThe failure, as an automatically gradable eval caseThe same defect ships again

⚠ It works on my machine is an ordinary failure, not an exotic one. The documented first suspect for Python is that a container or service manager runs the application with a different path than your shell, so a local install is invisible to the process.

Which cost number goes in front of whom

QuestionWhere the answer comes from
How is this run trendingThe agent result's dollar fields
What are we actually billedThe Usage and Cost API, or the Console usage page
What do we charge a customerNot these fields at all

⚠ The total_cost_usd and costUSD fields are client-side estimates, not authoritative billing data. They are computed locally from a bundled price table and can drift from what you are actually billed. Use them for development insight and approximate budgeting; the documentation says plainly: do not bill end users or trigger financial decisions from these fields.

Making the estimate honest anyway

Read fromWhat it costs you
Per-step output_tokensA placeholder reported before the response was generated
Raw usage on a budget errorLeaves out the response that crossed the budget
Successful results onlyFailed conversations still consumed tokens up to the failure
total_cost_usd or modelUsageThe documented basis, where the choice exists inside the result

⚠ A crash emits a final error result that may carry every cost field zeroed; recover the totals from what arrived before it.

⚠ Tracing is in beta, and span names and attributes may change between releases. A dashboard is a promise about a schema.

Ready to study Claude Certified Architect - Professional (CCAR-P)?

Practice tests, flashcards, and all study notes — free, no sign-up needed.

Start Studying — Free