Optimize fundraising attention policy
Allocate the current fundraising attention epoch with age-aware controlled Markov arm values and an exact multiple-choice capacity knapsack: compare action versus passive continuation through later stages, price founder distraction and action cost, expose a dynamic attention index, and disclose that independently relaxed future capacity is not a globally certified multi-period schedule.
What it's for
Turns founder fundraising time into a recalculating decision policy: which live opportunity and intervention deserve attention now, what value that attention adds through later stages, and where evidence is too weak to automate.
What you give it
Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.
| Field | Type | Role | Required |
|---|---|---|---|
| age_bucket_boundaries | array of integer | Evidence | Optional |
| attention_actions | array of objects (8 fields) ≥ 1 item | Evidence | Yes |
| attention_capacity_units | integer ≥ 0, ≤ 1000 | Your calibration | Yes |
| current_opportunities | array of objects (6 fields) ≥ 1 item | Evidence | Yes |
| discount_factor | number ≥ 0, ≤ 1 | Your calibration | Optional |
| distraction_cost_per_attention_unit | number ≥ 0 | Your calibration | Optional |
| historical_stage_episodes | array of objects (5 fields) ≥ 4 items | Evidence | Yes |
| horizon_periods | integer ≥ 1, ≤ 52 | Your calibration | Yes |
| liquidity_value_per_primary_dollar | number ≥ 0 | Your calibration | Optional |
| max_detail_rows | integer ≥ 1, ≤ 500 | Numerical control | Optional |
| minimum_stage_period_exposure | number ≥ 0 | Your calibration | Optional |
| prior_strength | number ≥ 0 | Your calibration | Optional |
| stage_order | array of string ≥ 2 items | Evidence | Yes |
Each attention_actions
record
| Field | Type | Required |
|---|---|---|
| advance_hazard_multiplier | number (≥ 0, ≤ 20) | Yes |
| attention_units | integer (≥ 1, ≤ 1000) | Yes |
| cash_cost | number (≥ 0) | Yes |
| eligible_stages | array of string (≥ 1 item) | Yes |
| evidence_verified | boolean | Yes |
| id | string (non-empty) | Yes |
| loss_hazard_multiplier | number (≥ 0, ≤ 20) | Yes |
| proceeds_realization_multiplier | number (≥ 0, ≤ 2) | Yes |
{
"attention_actions": [
{
"advance_hazard_multiplier": 1.4,
"attention_units": 1,
"cash_cost": 1,
"eligible_stages": [
"contacted",
"diligence",
"term_sheet"
],
"evidence_verified": true,
"id": "founder-update",
"loss_hazard_multiplier": 0.8,
"proceeds_realization_multiplier": 1
},
{
"advance_hazard_multiplier": 2,
"attention_units": 2,
"cash_cost": 1,
"eligible_stages": [
"diligence",
"term_sheet"
],
"evidence_verified": true,
"id": "partner-meeting",
"loss_hazard_multiplier": 0.5,
"proceeds_realization_multiplier": 1
}
],
"attention_capacity_units": 2,
"current_opportunities": [
{
"age_periods": 1,
"current_stage": "diligence",
"evidence_verified": true,
"id": "fund-alpha",
"loss_cost": 20,
"potential_primary_proceeds": 600
},
{
"age_periods": 0,
"current_stage": "term_sheet",
"evidence_verified": true,Truncated for display — the full payload is 479 lines.
What you get back
This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.
{
"decision": "fundraising_attention_policy_supported",
"failed_gates": [],
"guardrails": [
"Censored histories supply exposure without fake outcomes. The controlled Markov arm value propagates attention effects through later stages and compares each current action with waiting while the opportunity continues to evolve.",
"The current-epoch multiple-choice knapsack is exact over returned arm values. Future capacity is relaxed independently by arm, so this is an index policy to recompute after each stage event, not a globally certified multi-period schedule.",
"Action effects, cash costs, opportunity amounts, loss costs and liquidity value are organization-owned and require prospective calibration. Do not infer investor receptiveness from protected traits, personal data, names, nationality, or covert surveillance; this is not solicitation, securities, or financing advice."
],
"method": "finite_horizon_controlled_markov_arm_values_and_exact_epoch_knapsack",
"opportunity_diagnostics": [
{
"best_action_id": "partner-meeting",
"best_action_marginal_dynamic_value": 41.4059,
"current_stage": "diligence",
"dynamic_attention_index_per_unit": 20.7029,
"opportunity_id": "fund-alpha",
"passive_dynamic_value": 422.2568
},
{
"best_action_id": "partner-meeting",
"best_action_marginal_dynamic_value": 17.8075,
"current_stage": "term_sheet",
"dynamic_attention_index_per_unit": 8.9038,
"opportunity_id": "fund-beta",
"passive_dynamic_value": 343.3485
}
],
"selected_actions": [
{
"action_id": "partner-meeting",
"attention_units": 2,
"marginal_dynamic_value": 41.4059,
"opportunity_id": "fund-alpha"
}
],
"solver": {
"current_epoch_selection": "globally_optimal_multiple_choice_knapsack",
"multi_period_policy": "single_arm_index_relaxation_not_globally_certified"
},
"summary": {
"attention_capacity_units": 2,
"attention_units_used": 2,
"expected_incremental_dynamic_value": 41.4059,
"opportunity_count": 2,Truncated for display — the full payload is 52 lines.
How it works
Markov & state-space control — Choose a policy over evolving states when today's action changes tomorrow's options.
- 1 Fit the same censored age-state competing-risk model, then freeze prospectively calibrated, stage-eligible attention actions with advance/loss hazard multipliers, proceeds effects, cash and capacity cost.
- 2 For each live opportunity solve a finite-horizon controlled Markov arm: at each future state compare passive evolution with eligible attention, propagate advance, close, loss and dwell value, and calculate each current action's marginal dynamic value and per-unit index.
- 3 Solve the current decision epoch as an exact multiple-choice knapsack over opportunity/action pairs; recompute after the next observed event and withhold the policy when evidence, stage support or positive action value fails.
Before you trust it
Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.
Assumptions & guardrails
- States are decision-sufficient at the chosen cadence and transition evidence is recent, identifiable, and not silently pooled across incompatible regimes.
- Historical censoring is honest; stage dynamics are locally stable; action effects are prospective and causal enough for planning; proceeds, loss cost, cash/distraction cost and liquidity value are commensurable; no omitted action shares the constrained attention pool.
- A policy is conditional on the declared state abstraction and transition model; hidden omitted state can invalidate its ranking.
- The epoch knapsack is exact over single-arm dynamic values, while future capacity is an index relaxation and not globally certified. Do not use personal, protected or covertly inferred investor attributes; this is not automated solicitation, securities, legal or financing advice.
Minimum evidence
- historical_stage_episodes: at least 4 rows/items
- current_opportunities: at least 1 rows/items
- stage_order: at least 2 rows/items
- attention_actions: at least 1 rows/items
- attention_capacity_units: required and organization-defined
- horizon_periods: required and organization-defined
How to validate it
Backtest the chosen action against simple feasible baselines on held-out scenarios, sweep costs/constraints/risk tolerance, and require constraint feasibility under adverse inputs.
Calibrating it to your org
Same for everyone
The mathematical kernel, validation rules, method version, and JSON output semantics are organization-independent; no tenant-trained coefficients or company benchmark is embedded in the function.
Specific to you
- governed action-effect registry joined to the same versioned fundraising stage model and current opportunity snapshot, with prospective experimental or otherwise defensible effect provenance
- action executability and lawful perimeter, causal effect validity, stage eligibility, founder attention capacity and distraction cost, primary proceeds, loss cost, liquidity shadow value, discounting, stage/censoring stability, support gate and human activation authority
Calibration workflow
- 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
- 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
- 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
- 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
- 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
- 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.
Call it from your AI
You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.
gitrevio_capabilities_search
{ "q": "allocate the current fundraising attention epoch" }
→ finds "optimize_fundraising_attention_policy"
gitrevio_capability_describe
{ "capability_id": "optimize_fundraising_attention_policy" }
→ returns the input schema and agent guidance shown on this page
gitrevio_capability_run
{ "capability_id": "optimize_fundraising_attention_policy", "arguments": { ... } }
→ returns the result shown aboveWorks in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.
Related tools
Audit fundraising pipeline integrity
Audit a fundraising pipeline as point-in-time evidence rather than CRM theater: reconstruct monotone stage events, terminal status and primary proceeds, retain open opportunities as censored, reject forecasts made after resolution, detect duplicate active investor accounts, and gate the portfolio on mature-forecast support, Brier loss and calibration gap.
Forecast fundraising close and runway
Forecast whether enough primary capital closes before runway pressure by fitting empirical-Bayes age-state competing-risk hazards to advanced, closed, lost and right-censored stage episodes, then simulating every live opportunity under one common market scenario and an explicit burn-before-close cash convention.
Allocate budget with CVAR constraint
Maximize expected portfolio return while keeping probability-weighted loss CVaR below a finance-owned tail-risk ceiling across aligned joint scenarios.
Audit budget constraint binding
Audit whether a claimed budget constraint genuinely blocks value after dependency-feasible portfolio reallocation, separating current-plan inefficiency from scarcity with scenario CVaR and a discrete budget shadow price.
Audit financing term sheet integrity
Audit startup financing terms as exact share, price, proceeds and ownership identities: include pre-money option-pool increases and converting instruments in the pricing denominator, keep secondary purchases out of company cash and post-money share creation, reconcile primary issuance, post-money equity value and reported investor ownership, and retain evidence failures and impossible fees or secondary sales.
Audit technology diligence evidence integrity
Audit a frozen technology diligence case against buyer-declared system/domain/claim scope, management assertions and fresh, rights-cleared, independently reviewed point-in-time evidence.