WOAM Capital · Technical Specification
Compound Model · March 2026

Pipeline
Specification

Seven agents. Compound risk model. Chain mechanism.

Complete technical specification of the seven-agent AI investment pipeline, rebuilt around the compound risk architecture. Each agent is specified with its inputs, processing logic, outputs, and integration with the two-axis scoring model and the investment chain mechanism.

A-01
Intake Agent
WhatsApp conversation · Any language · Profile construction
A-02
Research Agent
Identity · Mobile money · Income baseline · Asset-income fit
A-03
Compound Score Agent
Trust axis · Asset axis · Matrix position · Structure recommendation
A-04
Interview Agent
Video assessment · Peer circle interviews · Consistency scoring
A-05
Decision Agent
Investment structure · Asset catalogue · Chain nominations · Terms
A-06
Agreement Agent
Bilingual legal docs · WhatsApp execution · Payment instruction
A-07
Compound Monitor Agent
Dual-axis tracking · Trust score updates · Chain trigger · Early warning
Architecture

How the pipeline
actually works

The pipeline is not sequential. Agents A-01 and A-02 run in parallel. A-03 runs as soon as A-02 completes. A-04 is triggered conditionally by A-03's score. A-05 waits for both A-03 and A-04. A-06 and A-07 run post-investment. End-to-end latency: under 48 hours for auto-approve candidates; 72–96 hours for candidates requiring A-04 interview.

Inputs received
Output produced
A-01
NGO warm introduction message · WhatsApp number · Language auto-detect
Candidate profile · Business description · Income declaration · Franchise type classification
A-02
[Runs parallel to A-01] · Mobile money API · Identity registries · NGO programme records
Income baseline (3-way triangulated) · Transaction history · Social graph density · Asset-income fit score · Research dossier
A-03
A-01 profile + A-02 dossier · Asset catalogue query · Franchise type lookup
Trust score (0–100) · Asset coverage score (0–100) · Matrix quadrant · Recommended investment structure · Routing decision
A-04
[Conditional: score 60–84 OR trust <75 with asset <50%] · A-03 flags · Peer circle contact list
Video assessment consistency score · Peer circle validation scores · Supplementary trust score delta · Updated compound score
A-05
Final compound score · Recommended structure from A-03 · Asset catalogue selection · Chain nomination slots
Investment decision · Capital structure (grant / asset / cash split) · Revenue share rate · Break-even calculation · Chain nomination request to candidate
A-06
A-05 decision · Candidate language · Mobile money account · Asset procurement instruction
Signed bilingual agreement · Payment instruction executed · Asset delivery order to NGO field officer · Chain nomination intake opened
A-07
[Continuous from investment date] · Revenue share payments · Platform engagement · Video check-ins · Asset condition reports · Formalisation registry
Updated trust score (monthly) · Chain trigger evaluation · Early warning flags · Support module activation · LP-grade KPI report
Agents 01 – 02

Intake & Research.
Parallel. Blind.

A-01 and A-02 run simultaneously and independently. A-02 does not see A-01's outputs before completing its research — this prevents A-01's conversation from anchoring A-02's income estimate. The two outputs are compared in A-03 as independent signals.

A-01
Agent One
Intake Agent
Conversational profile construction via WhatsApp · Any language
Live
What it does

Conducts a 20–35 minute natural language conversation via WhatsApp. Language is auto-detected from the first message and the entire conversation proceeds in that language. The agent does not ask about finances first — it asks about the business, the community, the people. Financial questions emerge from the conversation rather than leading it. Average completion rate: 84%. Average duration: 28 minutes.

The conversation builds a structured profile covering: business type and history, community relationships, group membership, income sources, existing financial relationships (mobile money, savings groups, any MFI history), and the candidate's own assessment of what investment would enable.

Key design principles
  • Trust before finance — financial questions come only after community and business are established
  • Franchise type classification drives subsequent scoring — correct classification is critical
  • Income declaration is recorded as stated — A-02 cross-references independently
  • Barriers to formalisation are probed and recorded — tax anxiety, legal uncertainty, social risk
  • The NGO warm introduction is referenced early — "I'm reaching out on behalf of [NGO name] who suggested I contact you"
  • Chain nomination concept is not introduced at intake — only at the A-05 decision stage
Output schema
{ franchise_type, business_description, income_declared, group_membership, community_embeddedness_signals, formalisation_barriers, payment_rail_confirmed, language, conversation_flags }
A-02
Agent Two
Research Agent
Independent verification · Income triangulation · Asset-income fit scoring
Updated — compound model
What it does

Runs in parallel with A-01, querying four independent data sources to produce a structured research dossier. The income baseline is triangulated from three signals — declared (from A-01 output), transactional (from mobile money history), and peer-validated (from NGO programme records where available). The baseline is the median of the three; significant divergence between signals is flagged.

New in the compound model: A-02 now also queries the asset catalogue to score asset-income fit for the top 3 most appropriate assets for this candidate's franchise type and geography. This output is passed directly to A-03 for the asset coverage axis.

Data sources queried
  • Mobile money API (MTN MoMo, M-Pesa, Orange Money, Airtel Money) — 6–12 months transaction history with opt-in permission obtained in A-01
  • NGO programme records — WfWI, Saathi, KOMPANION, and other partner databases for registered programme members
  • Cooperative registry — Rwanda RCA, Kenya Cooperative Alliance, and equivalents — checks for existing formal status
  • Social graph analysis — payment counterparty network density from mobile money data, no individual disclosure required
  • Asset catalogue database — franchise type × geography × asset type → income uplift estimate and fit score
Output schema
{ income_baseline, income_confidence, transaction_history_months, social_graph_density, ngo_programme_verified, cooperative_status, asset_recommendations: [{asset_type, cost, fit_score, income_uplift_estimate, enforcement_mechanism}], research_flags }
Compound model addition
Asset-income fit scoring is new in this version. A-02 now queries the asset catalogue and returns the top 3 asset recommendations for this candidate's franchise type, with fit scores (0–1), estimated income uplift, local availability, and enforcement mechanism. This data is the primary input to the asset coverage axis in A-03.
Agents 04 – 05

Interview & Decision.
Conditional depth. Precise output.

A-04
Agent Four · Conditional
Interview Agent
Video assessment · Peer circle interviews · Behavioural consistency scoring
Updated — video + peer circle
Video assessment

A structured 15–20 minute video conversation via WhatsApp or equivalent. Not a form — a dynamic conversation that adjusts based on responses. The AI analyses response latency, narrative consistency, vocal stress patterns, and cross-references statements against A-01 and A-02 data in real time.

A baseline is established in the first session. Subsequent monthly check-ins (A-07) track deviation from this individual's baseline rather than from a population average — controlling for cultural and individual variation in expression style.

Peer circle interviews

Three people nominated by the candidate — savings group members, regular customers, neighbours — each receive a short independent AI interview (8–12 minutes). The AI checks for narrative consistency across peer accounts and against the candidate's own account, without revealing what any other interviewee said.

  • How long have you known her? Describe her business.
  • Has she ever failed to honour a financial commitment to you or your community?
  • What would happen in your community if she received money and did not keep her commitments?
  • Can you describe the people who use her services / participate in her group?
Output schema
{ video_consistency_score, peer_circle_scores: [3 scores], narrative_alignment_score, trust_delta, flags, updated_trust_score }
A-05
Agent Five
Decision Agent
Final investment structure · Chain nomination · Terms communication
Compound model
What it decides

A-05 takes the final compound score (from A-03, updated by A-04 where triggered) and produces the complete investment recommendation. It does not recalculate the score — it structures the investment to the score. The recommended structure from A-03 is confirmed or refined based on any A-04 updates.

Critically: A-05 also activates the chain nomination process. The candidate is informed that she will nominate three people from her community before receiving the working capital tranches — and that those three people's access to investment depends on her completing the revenue share relationship. The nomination request is framed as a partnership, not a condition.

Communication to candidate
  • Decision communicated in candidate's language — plain terms, no jargon
  • Total investment, grant component, and working capital distinguished clearly
  • Revenue share explained: "10% of your income, only until we have received [WC amount] — your existing income is yours from day one"
  • Asset description: what it is, what it does, who it belongs to until break-even, how ownership transfers
  • Chain nomination: "When you reach [break-even threshold], three women you nominate will receive the same opportunity you are receiving now"
  • Timeline: "We expect this to take approximately [N] months based on your income"
Output to A-06
{ approved: true, investment_total, grant_amount, assets: [{type, cost, local_supplier, title_retention_mechanism}], cash_wc, revenue_share_rate: 0.10, breakeven_threshold, chain_nominations_required: 3, chain_release_trust_minimum, monitoring_cadence, agreement_language }
Agents 06 – 07

Agreement & Monitor.
Execution and continuous intelligence.

A-06
Agent Six
Agreement Agent
Bilingual legal documentation · WhatsApp execution · Asset and payment instruction
Live
Legal document generation

Generates a bilingual Stage 1 Pre-Formalisation Agreement (English + candidate's language) from the A-05 output. The agreement is plain-language by design — comprehensible to someone who has never seen a legal document. Title retention language is explicit: assets listed by name and description, ownership conditions stated clearly, transfer trigger defined as the break-even threshold amount.

WhatsApp delivery. Candidate acceptance recorded with timestamp and message hash as execution evidence — legally valid in Rwanda, Kenya, Uganda, Tanzania, and Ghana without physical signature.

Post-agreement actions
  • Payment instruction generated: formalisation grant disbursed immediately to confirmed mobile money account
  • Asset procurement instruction sent to NGO field officer with delivery confirmation required before Tranche 2 release
  • Chain nomination intake opened: candidate receives link to nominate three candidates; pipeline opens intake conversations with each nominee
  • Stage 2 (equity agreement) template held pending formalisation — auto-triggered when cooperative registry confirms registration
  • Monitoring baseline initialised in A-07: income baseline, asset baseline, payment schedule, chain nomination status
A-07
Agent Seven · Always On
Compound Monitor Agent
Dual-axis tracking · Trust score updates · Chain trigger · Early warning · LP reporting
Compound model — rebuilt
Continuous monitoring signals
  • Revenue share payments — platform collection at source (where MNO API live) or voluntary remittance; cumulative tracked against break-even threshold
  • Mobile money inflows — monthly income vs income baseline; growth consistent with franchise type; irregularities flagged
  • Platform engagement — modules accessed, formalisation progress tools used, network activity; strong predictor of completion
  • Video check-ins — monthly 10-minute conversation; consistency tracked against individual baseline from A-04; stress signals cross-referenced with financial data
  • Asset condition — NGO field officer confirmation (monthly for enhanced, quarterly for standard); photo upload to platform; MDM status for smartphones
  • Formalisation registry — cooperative registry queries; Stage 2 trigger when registration confirmed
  • Peer network health — payment counterparty activity; community relationship changes visible in transaction data
Compound monitoring logic

A-07 does not produce a single risk flag. It produces a compound risk profile update — both trust score and asset status updated simultaneously. A missed payment is interpreted differently depending on the current trust score and asset coverage status (see Appendix A for three worked examples).

Chain trigger evaluation runs monthly: when cumulative revenue share reaches the break-even threshold AND the current trust score meets the minimum set at investment, the chain is released — three nominee intake conversations opened automatically, no human intervention required.

LP-grade KPI reports generated automatically as a byproduct of monitoring. No additional data collection required. Reports delivered to management company dashboard continuously; LP-formatted reports generated quarterly.

Compound model — key change from previous version
Previous A-07 tracked revenue share payments and flagged anomalies. Rebuilt A-07 tracks two independent dimensions simultaneously and routes the same event to different response protocols based on the compound risk profile. A missed payment from a high-trust, high-asset portfolio company routes to grace period. The same event from a medium-trust, zero-asset portfolio company routes to urgent video check-in. The same event from a low-trust, high-asset portfolio company routes to asset condition verification. See Appendix A, Section 7 for full worked examples.
Asset Catalogue

Title-retained assets.
The structural protection layer.

The asset catalogue is a structured database of available productive assets by franchise type and geography. A-02 queries it during research. A-03 selects from it during structuring. A-06 instructs procurement. A-07 monitors condition. The catalogue is updated quarterly as new geographies and franchise types are activated.

Digital · Communications
Smartphone
$80–$200 · Tecno/Itel/Infinix

Essential for savings group coordinators, market traders, and any franchise requiring mobile money at full functionality. MDM profile installed at disbursement.

Enforcement: IMEI registered · MDM deactivation if abandoned · Bulk supply from local distributor at 15–20% below retail
Energy · Productive
Solar home system
$150–$350 · BBOXX/SolarNow

High fit for community kitchens (refrigeration, lighting), learning companions (evening sessions), and any franchise requiring reliable evening operation. Eliminates kerosene cost immediately.

Enforcement: Serial registration · Installed infrastructure (not easily moved) · Primary incentive is ownership desire, not recovery threat
Energy · Productive Use
Solar-powered appliance
$200–$500 · Sector-specific

Sewing machine (tailoring), grain mill (food processing), refrigerator (community kitchen / pharmacy). Direct income generation. Asset-income fit typically 0.90+.

Enforcement: Serial number · Asset tag · NGO field officer quarterly check · Highest fit score assets in catalogue
Transportation
Cargo bicycle
$150–$320 · Local market

Market traders, agricultural collectors, delivery service operators. Replaces 2–4 hours daily walking. Directly increases daily trading capacity and geographic reach.

Enforcement: Serial registration · WOAM Capital title on local registry · Inspection at monthly NGO contact
Transportation
Motorcycle
$800–$1,200 · Local market

Repair collectives, health workers, agricultural collectors in large-radius geographies. Highest value single asset in catalogue. Vehicle registry title retention is legally cleanest enforcement mechanism.

Enforcement: Vehicle registry registration in WOAM Capital name · Strongest legal title in catalogue · Cannot be sold without title transfer
Tools · Equipment
Franchise-specific tools
$100–$500 · Franchise-dependent

Repair tools (diagnostic equipment, hand tools), healthcare equipment (TENS machine, treatment table), documentation equipment (camera, audio recorder for cultural memory keepers). Directly enables the income-generating activity.

Enforcement: Serial number registration · WOAM Capital asset tag · NGO field officer verification · Tamper-evident marking
Chain Mechanism

How the chain
works technically

The investment chain is not managed by a separate system. It is embedded in the pipeline — each stage has a specific role in the chain mechanism. A-05 activates nominations. A-06 opens nominee intake. A-07 monitors the chain trigger. The chain grows through the same pipeline that processes every investment.

1
Nomination
A-05 activates nomination request
After investment decision, A-05 communicates the chain mechanism to the portfolio company and requests three nominations. Nominations are not a condition of receiving the formalisation grant (Tranche 1 — unconditional). They are a condition of receiving the working capital (Tranches 2 and 3). The portfolio company has 14 days to submit nominations before a grace period extension is offered.
2
Pipeline
A-01 opens intake for each nominee simultaneously
Each nominee receives a WhatsApp message: "[Referring portfolio company name] suggested I contact you. She has received support from WOAM Capital and thought you might also be interested. Would you like to have a conversation about your business?" The intake conversation runs through the standard pipeline. The nomination is recorded but the nominee's assessment is independent — the referrer's endorsement is a trust input, not a bypass of due process.
3
Hold
Nominees receive conditional approval — funds held
Nominees who pass the pipeline scoring receive conditional approval. The investment decision is made and communicated. But funds are not disbursed — the conditional approval is held pending the referring portfolio company's break-even trigger. Nominees are informed of this clearly: their investment is committed and waiting for the referring portfolio company to reach her break-even. Typical wait: 17–23 months at standard income growth.
4
Trigger
A-07 evaluates chain trigger monthly
Each month, A-07 checks: (a) has cumulative revenue share reached the break-even threshold? (b) is the current trust score at or above the minimum set at investment? If both conditions are met, the chain is released automatically — no human approval required. The three nominees' investments proceed to A-06 (agreement) simultaneously. The releasing portfolio company is notified that she has unlocked capital for her three nominees.
5
Circuit
Circuit breaker: if chain fails, nominees released via standard pipeline
If the referring portfolio company fails to reach break-even within a defined window (currently 36 months from investment — approximately 2× the expected break-even timing), the circuit breaker fires. Nominees whose conditional approvals remain valid are released into the standard pipeline — their investment proceeds without the chain trigger, funded from the general investment pool. WOAM Capital absorbs the loss on the broken chain link. The nominees are not penalised for someone else's failure.

"The chain is not a referral programme. Every component runs through the standard pipeline. The only difference is that the disbursal of working capital tranches is conditional on nomination, and the release of nominees' investments is conditional on the referrer's break-even. Trust flows through the same system that processes every investment."

← Series A Narrative
WOAM Capital · Pipeline Specification · March 2026
Appendix A →