MacroAgenticStrategy layer for agent workforces
Scale 1 min = 9 pxRev CSign in
Run of record
Project
GOAL-0417
Acceptance
≥ 88% gross retention, booked by 30 SEP
Executors
5 agents · 2 tools · 2 people
Portfolio
5 functions · 24 executors
Tasks
12, of which 8 critical
End to end
02:12
Planned
$359.50
Actual
$400.15
Ceiling
$450.00

Not a mock-up. Every drawing below is plotted from this table.

The Agentic CEO for
your agentic workforce.

MacroAgentic is your strategic co-CEO, develops and runs a plan across every agent available, tool and person you already have — and manages, to the cent, the unit economics.

00:0000:1500:3000:4501:0001:1501:3001:4502:00+4m float+18m float+6m float+6m floatSEGMENT ACCOUNTSDRAFT RETENTION PLAYSSEND WAVE 1RECONCILE CRMSCORE CHURN RISKEXEC REVIEWCALL TOP 40BOARD READOUTPULL USAGE TELEMETRYPRICE APPROVALSBUILD SEQUENCESTRIAGE REPLIEST1T2T3analyst-01 · 22mT4writer-03 · 18mT5rev-ops · 26mT6vp-success · 20mT7T8T9voice-01 · 34mT10support-04 · 28mT11T12T1T2T3analyst-01 · 22mT4writer-03 · 18mT5rev-ops · 26mT6vp-success · 20mT7T8T9voice-01 · 34mT10support-04 · 28mT11T12NOW

FIG. 1Retain the Q3 renewal book resolved to an execution network. 12 tasks across 9 executors; the red spine is the critical path and the hatched tails are total float — the slack the runtime can spend before the finish date moves.

Critical path — 02:12Off the critical pathTotal floatNOW — runtime position
DET. A

Goal reasoning

You state an outcome. It writes the work breakdown.

There is no prompt chain here. The layer decomposes the outcome into objectives, then into tasks, and attaches an acceptance test to every leaf before a single agent is woken up. It decides what needs an agent, what is a plain tool call, and what a person has to sign.

A leaf that fails its test goes back to the planner, which re-cuts the branch. That is the difference between a system with a strategy and a system with a retry loop.

1.1SEGMENT ACCOUNTSAGENTPASS IF every ARR $ assigned to one tier1.2PULL USAGE TELEMETRYTOOLPASS IF 90d coverage, < 2% null1.3SCORE CHURN RISKAGENTPASS IF scored book reconciles to CRM ARR1.0KNOW WHO IS AT RISK2.1DRAFT RETENTION PLAYSAGENTPASS IF one play per risk tier2.2PRICE APPROVALSHUMANPASS IF discount > 15% signed off2.3EXEC REVIEWHUMANPASS IF every save offer signed by the sponsor2.0DECIDE THE PLAY3.1SEND + CALL THE BOOKAGENTPASS IF 100% of tier 1 contacted3.2TRIAGE REPLIESAGENTPASS IF no reply older than 4h3.3RECONCILE + READ OUTAGENTPASS IF CRM matches the ledger3.0RUN IT AND SETTLE UP0RETAIN THE Q3 RENEWAL BOOKACCEPT IF ≥ 88% gross retention

FIG. 2Work breakdown for GOAL-0417. Assignee class is decided at plan time, not at run time; the acceptance test on each leaf is what the layer will actually check before marking it done.

DET. B

Orchestration runtime

The plan is the program.

The breakdown compiles to a dependency graph and the runtime executes it — wide wherever the graph allows, serial wherever it does not, with the critical path computed rather than guessed. State is checkpointed per task, so a dead agent, a rate limit or a redeploy resumes instead of restarting.

Executors are whatever you already run: Claude agents, in-house services, MCP servers, cron jobs, people. The layer does not care what is on the other end of the task, only whether it passed.

123SERIAL 26MSERIAL 12MSERIAL 26MPEAK 3 AT ONCE00:0000:3001:0001:3002:00

FIG. 3Concurrency profile for the same run. The dimensioned spans are serial pinches — stretches where the dependency graph forbids parallelism, and where adding agents buys you nothing.

DET. C

One CEO, every function

Sales, marketing, engineering, training, customer success — the same machinery.

A workforce is not one team. The layer runs 5 goals at once across 5 functions, over 24 executors and 6 people, on one clock and one ledger. Sales is working the pipeline while engineering ships the billing migration and training certifies the cohort that sales is about to hire into.

Nothing about the machinery changes between them. The same decomposition, the same critical path, the same ceiling — which is the point. A function does not need its own agent stack, it needs its outcome stated.

FUNCTION · GOALTASKSEND TO ENDSPENT00:0000:2000:4001:0001:2001:4002:00SALESGOAL-0431 · Convert the Q3 pipeline5 agents · 1 person601:38$205.90MARKETINGGOAL-0428 · Launch the category campaign5 agents · 1 person601:41$195.15ENGINEERINGGOAL-0435 · Ship the billing migration3 agents · 1 person601:48$127.90CUSTOMER SUCCESSGOAL-0417 · Retain the Q3 renewal book7 agents · 2 people1202:12$400.15TRAININGGOAL-0442 · Onboard the new AE cohort3 agents · 1 person501:36$169.75ALL FUNCTIONS · 5 GOALS AT ONCE · 24 EXECUTORS3502:12$1,098.85

FIG. 45 concurrent goals, one per function, on a shared clock. Red is each goal’s own critical path; the totals are the whole portfolio, not one team’s slice of it.

DET. D

System-wide cost visibility

Cost per outcome, not cost per token.

Every token, tool call and human minute is booked against the goal that caused it. Plan and actual are integrated on the same schedule, so an overrun surfaces while there is still time to act on it — not in next month’s invoice.

This run finished $40.65 over its own plan, at 11.3% variance, and stayed inside the standing ceiling. Ceilings are enforced by the runtime: it stops work before it crosses one.

$0.00$100.00$200.00$300.00$400.0000:0000:3001:0001:3002:00STANDING CEILING $450.00 — RUNTIME STOPS HERE+40.65 (11.3%)ACTUAL $400.15PLAN $359.50

FIG. 5Cumulative cost, plan against actual, on the run’s own clock. The hatched wedge is variance; the dashed rule is the ceiling the runtime will not cross without a signature.

DET. E

Authority & policy

An agent gets a spend limit, like anyone else on the payroll.

Each executor carries a standing limit and a blast radius. Work that would breach either is escalated for a signature instead of attempted, and the refusal is recorded alongside the spend. On this run voice-01 came within a few dollars of its cap — visible here while the run was still going, rather than discovered afterwards.

SPENT ▸◂ STANDING LIMITvoice-01$149.8094% OF CAPanalyst-01$70.3070% OF CAPwriter-03$31.1062% OF CAPsupport-04$36.2060% OF CAPops-02$16.4541% OF CAPpostmark.mcp$3.1012% OF CAPwarehouse.mcp$1.208% OF CAP

FIG. 6Spend against standing authority, per executor, for GOAL-0417.

DET. F

Performance review

Every run closes with a reconciliation.

The same table that scheduled the work settles it. Which executor earned its place, which one ran hot, and what the outcome cost in total — reconciled per run, not estimated per month.

ExecutorClassTasksMinutesPlanActualVar.
voice-01AGENT134$124.00$149.80+25.80
analyst-01AGENT343$60.00$70.30+10.30
rev-opsHUMAN126$52.00$52.00
vp-successHUMAN120$40.00$40.00
support-04AGENT128$33.50$36.20+2.70
writer-03AGENT118$29.60$31.10+1.50
ops-02AGENT225$16.10$16.45+0.35
postmark.mcpTOOL16$3.10$3.10
warehouse.mcpTOOL18$1.20$1.20
Total — GOAL-0417$359.50$400.15+40.65
DET. G

Closing the loop

Every run teaches the next one.

Plan, dispatch, measure, re-rank, plan again. Outcomes and costs go back into the registry, so the next composition ranks candidates on what they actually delivered rather than what they advertised. An agent that runs hot gets picked less; one that quietly beats its price gets picked more.

Across six passes at the same goal, cost per outcome fell 18.1% and then flattened — the loop converges rather than chasing its tail.

PLANdecompose + matchDISPATCHrun the graphMEASUREcost + outcomeRE-RANKupdate the registryEVERY RUNTEACHES THE NEXT ONEPASS 1PASS 2PASS 3PASS 4PASS 5PASS 6$400.15$327.6018.1%COST PER OUTCOME, SAME GOAL, SUCCESSIVE PASSES

FIG. 7The control loop, and what it bought. Each pass re-ranks against measured cost and outcome; the curve flattens when further swaps cost more than they save.

  1. P1First run. Candidates ranked on declared cost only.
  2. P2voice-01 capped; overflow routed to voice-02.
  3. P3Telemetry pull cached across goals.
  4. P4Draft step swapped to a cheaper writer at equal eval score.
  5. P5Exec review batched, off the critical path.
  6. P6Converged. Further swaps cost more than they save.
Notes

Applies to the whole sheet unless a detail says otherwise.

  1. 01The layer does not replace your agents. It sits above them and decides what they are for — which is the one job none of them can do for themselves.
  2. 02Memory is held at the goal, not in a chat. A run that is paused for three weeks resumes with the same plan, the same acceptance tests and the same ledger.
  3. 03People are first-class executors. A task assigned to a human gets the same dependency, the same clock and the same line in the cost report as one assigned to an agent.
  4. 04Re-planning is the default failure mode, not retry. A leaf that fails its acceptance test returns to the planner, which decides whether to re-cut the branch or escalate.
  5. 05Nothing is marked done without an acceptance test written before the work started. The layer will not accept its own output as evidence.
  6. 06Executors attach over MCP, plain HTTP or a queue. If it can take a task and return a result, it can be on the org chart.
  7. 07Functions do not get their own stacks. Sales, marketing, engineering, training and customer success run on the same decomposition, the same scheduler and the same ledger — what differs is the outcome each one states.
  8. 08The loop is the product. A run that is never measured teaches nothing, so cost and outcome are written back to the registry the moment a goal closes.
  9. 09Every figure on this sheet is computed, not drawn. The schedules, the float, the curves, the function bands and the ledger all come out of the same critical-path solver, run over real task tables.

Take the next run off your own desk

Bring one outcome you currently chase across four tools and three people. We plan it, run it and hand you the sheet.

Get started
Project
MacroAgentic
Drawing
Strategy layer — GOAL-0417
Scale
1 min = 9 px
Units
Minutes · USD
Tasks
12
Duration
02:12
Cost of run
$400.15
Sheet
1 of 1 · Rev C