Parker

Engineer · AI usage, harness practices and where spend could have gone further.

Demonstration report. This engineer and their usage are synthetic.

This page is for the engineer and their manager to find cheaper defaults and harness fixes. It is never used to rank people.

Period

Summary

FY2027 to date

AI spend

$2,157

185 sessions · $12 each

Savings opportunity

$656

30% of spend, at list price

Practices on target

3 of 7

Focus: rtk coverage, Knowledge MCP use

Cache hit rate

80%

Team 78% · org 82%

FY2027 to date · estimated at list price

$656 of $2,157 spend

Largest saving first

Improve prompt caching$343
Cache hit rate 80% against an 85% target.
Keep stable context at the front of prompts and avoid switching models mid-session, so 85% of input reads from cache.
Right-size models$227
42% of routine spend ran on Opus: code research $304 of $708, documentation $140 of $336, ops $11 of $28, presentation $95 of $241.
Run docs, decks, ops and read-only research on Sonnet instead of Opus.
Keep context lean$68
9% of input tokens were past 100K of context.
Compact or /clear between tasks instead of carrying more than 100K tokens of context.
Route commands through rtk$17
61% of 1,570 Bash calls ran through rtk; 2.6M tokens of command output went into context.
rtk compacts test, build, git and search output before it reaches the model (about 70% smaller).

Compared with Data & Analytics and the organization

Practice	Parker	Team	Org	Target	Recommendation
rtk coverage	61%Needs attention	60%	56%	≥ 80% of Bash calls	Install the rtk hook so test, build, git and search commands return compact output.
Knowledge MCP use 8 findings saved back	30%Needs attention	30%	30%	≥ 40% of sessions	Search the knowledge MCP before re-researching a codebase, and save findings back with remember.
Cache hit rate	80%Needs attention	78%	82%	≥ 85% of input	Keep CLAUDE.md and tool lists stable, avoid mid-session model switches, and resume sessions instead of restarting.
Context past 100K	9%On target	8%	10%	≤ 10% of input	On target
Opus on routine work	42%Needs attention	55%	44%	≤ 25% of routine spend	Default routine sessions to Sonnet and keep Opus for harness, design and hard debugging work.
Sessions using subagents	45%On target	45%	50%	≥ 30% of sessions	On target
Tool error rate	3.3%On target	3.4%	3.0%	≤ 4% of tool calls	On target

Is the engineer moving toward the targets?

Month	rtk	Knowledge	Cache	Context	Model fit	Subagents	Savings opportunity
Feb 2026	–	–	60%Off target	8%On target	62%Off target	35%On target	$24
Mar 2026	–	–	66%Off target	7%On target	62%Off target	43%On target	$93
Apr 2026	–	–	63%Off target	8%On target	62%Off target	33%On target	$80
May 2026	28%Off target	–	73%Needs attention	11%Needs attention	55%Off target	44%On target	$113
Jun 2026	41%Needs attention	8%Off target	73%Needs attention	8%On target	53%Off target	38%On target	$98
Jul 2026	74%Needs attention	19%Needs attention	77%Needs attention	7%On target	40%Needs attention	43%On target	$51
Aug 2026	67%Needs attention	40%On target	78%Needs attention	11%Needs attention	38%Needs attention	41%On target	$107
Sep 2026	65%Needs attention	38%Needs attention	87%On target	6%On target	30%Needs attention	50%On target	$43
Oct 2026	69%Needs attention	33%Needs attention	84%Needs attention	11%Needs attention	27%Needs attention	58%On target	$47

FY2027 to date

By model tier, by month

Routine types (documentation, presentation, ops, code research) rarely need Opus

Cache reads are billed at a fraction of uncached input; a high share is what good caching looks like

Model	Requests	Uncached input	Cache reads	Cache writes	Output	Spend
Opus 5.5	2,144	11.7M	107.7M	7.6M	3.4M	$175
Opus 5	2,385	18.1M	115M	8.5M	3.3M	$284
Opus 4.8	1,648	19M	72.3M	5.8M	2.4M	$228
Opus 4.6	3,136	48.4M	127.9M	11.3M	4.3M	$485
Opus 4.5	58	1.4M	2.4M	240K	90K	$12
Sonnet 5.5	4,118	23.7M	214.3M	15.2M	6.3M	$170
Sonnet 5	5,902	39M	267.7M	19.6M	8.3M	$396
Sonnet 4.6	2,791	37.6M	112.8M	9.6M	4.1M	$245
Sonnet 4.5	49	1.1M	2M	195K	66K	$6
Haiku 5.5	6,897	41.2M	363M	25.8M	12.7M	$17
Haiku 4.5	5,081	46M	239.5M	18.2M	7.3M	$129
GPT Realtime (meetings)	105	1.3M	5M	402K	162K	$11
Total	34,314	288.6M	1629.5M	122.4M	52.6M	$2,157

Spend and tokens come from LiteLLM gateway logs and imported agent session history, priced at API list rates.
Savings are estimates: routine Opus work repriced at the Sonnet available at the time; uncached input moved to cache reads up to an 85% hit rate; rtk removing 70% of unfiltered command output, which stays in context for about half the session's remaining turns; and 50% of input past 100K context avoided by compacting. Levers overlap, so the total is an upper bound.
rtk and the knowledge MCP are scored from their rollout (rtk May 2026, knowledge MCP Jun 2026).
Generated Oct 20, 2026, 7:46 AM.