Validate More, leaning Go

Opportunity Validation: La Cazadora

Discovery call: Apr 21, 2026 · Validation written: Apr 22, 2026

Workflow hypothesis: Planner Co-Pilot (Track B)

Summary (BLUF)

Verdict: Validate More, leaning Go. Strong ICP fit (11/12), strong buyer signal, confirmed market gap, and La Cazadora would be the first Planner Co-Pilot case in the portfolio. Five reasons we are not yet at Go:

Engine architectural divergence. The current platform is a request-processing pipeline. La Cazadora's pain is state-driven scheduling. Needs a decision before scoping.
One-client signal, not a validated pattern. n=1 so far. MAGG is adjacent, not the same.
Volume not yet quantified. Orders per day, schedule regen frequency, lost-customer cost. Without these numbers we cannot price.
Category risk: MMS (Manufacturing Management System). Raffaello flagged a dedicated MMS category may be the textbook answer. Research before we scope, or we risk reinventing packaged software.
Buyer dynamics are split. Ruco (CEO) has budget authority but will not go deep on operational specifics. Daniel (co-CEO, ops) is the actual pain owner. Sell Daniel, close Ruco.

Next action: focused planning call with Daniel, three Excels under NDA, investigate MMS landscape in parallel, validate the 10 questions in section 10, pick between (a) narrow ETA-on-order MVP that force-fits the current engine, or (b) state-driven engine extension that also unlocks MAGG B1/B2/B3 and Eurostar B4.

00 Updates since Call 1 Apr 22

Two new inputs since the Apr 21 discovery: four screenshots from Daniel's Apr 21 presentation (OEE dashboard x2, sábana master schedule, ISIMAT product spec matrix), plus a follow-up transcript between Giuseppe and Raffaello.

Confirmed (moved from hypothesis to fact)

Claim	Evidence	Implication
OEE is measured at machine-level, not plant-level	Dashboard shows per-machine Availability / Performance / Quality, real-time from Pulsar.	We can price against measured baselines, not estimates.
Availability is the bottleneck, not Performance or Quality	OEE 47.74% = Avail 58.60% × Perf 87.15% × Calidad 93.46%. Perf and Calidad already within industry norms.	ROI lever is sequencing and changeover reduction, not speed or scrap. Changes the pitch.
Single planning artifact exists (the sábana)	Screenshot: ~20 columns, one master Excel, one visible owner.	Target system is well-defined. No need to rationalize N competing schedules.
Microsoft GP → Excel planning flow confirmed	Transcript: orders flow from GP into the sábana manually.	GP is the order source of truth. Excel is the planning layer. Two integration targets, not one.
Pulsar is NOT connected to planning today	Transcript: Pulsar feeds the OEE dashboard but the sábana does not ingest Pulsar's downtime signal.	Real-time capacity update is a clean Phase 2. Phase 1 MVP can skip Pulsar integration.
Gemini is already embedded in their Excel stack	Dashboard screenshot shows Gemini prompt UI inside the workbook.	Team is NOT AI-naive. Positioning is "connect your data to AI", not "adopt AI".

New risks surfaced

MMS category risk. Before scoping, research whether a dedicated MMS (Manufacturing Management System) already solves this. If yes, we are competing against packaged software, not filling a gap.
Excel-refresh cadence is manual. Sábana has no visible "last updated" or auto-refresh. Inputs-not-timely is a bigger risk than logic-not-good. The planner is making decisions on stale data.
Buyer-champion split. Ruco (money) does not go deep. Daniel (pain) is not the signer. Every deliverable must speak to both in different languages.

New accelerators surfaced

Excel logic is fine — replicate, don't rebuild. The sábana's planning logic works. The gap is process-manual + inputs-not-timely, the cheaper half of the problem. MVP scope drops, build risk drops, Phase 1 ships faster.
ROI handle is now in MXN. Raffaello: 10% OEE improvement = real money; efficiency gains, not headcount replacement. Pricing anchor shifts from "cost to build" to "value of 1 point of Availability per month".
Sector strategy aligned with Contru model. Contru specializes by sector with multiple modules; our thesis is "manufacturing distributors" as the broader segment with commercial + operational suite. La Cazadora becomes the lead proof case for a "Suite for Manufacturing Distributors" positioning.

Volume dimension

1 → 2 / 3

Machine-level OEE granularity confirmed.

ROI handle

1 → 2 / 3

Sequencing thesis grounded, MXN anchor identified.

New composite

12 / 15

Up from 10 / 15.

01 Opportunity Clarity Score

Scored 0 to 3 per dimension. Anything under 2 needs validation before proposing.

Dimension	Score	Evidence	Gap to close
Workflow scoped	2/3	Planner Co-Pilot hypothesis named. Bounds fuzzy: schedule regen vs ETA-per-order vs both. Sales UI ownership TBD.	Narrow to ONE sub-workflow before pricing.
Owner named	3/3	Daniel owns the pain (co-CEO, ops). Brandon is internal ally. Ruco champions.	None.
Volume quantified	2/3	120M tubes/year, 120 customers, 2-month lead time. Machine-level OEE granularity confirmed via dashboard (Avail / Perf / Calidad per machine, real-time from Pulsar). Orders per day and schedule changes per day still NOT confirmed.	Get sábana Excel + daily and weekly volumes of planner output.
ROI handle	2/3	OEE 47.74% measured. Availability 58.60% is the bottleneck (Perf 87.15%, Calidad 93.46% are fine) — sequencing is the lever. Raffaello: 10% OEE improvement = real money in MXN. No $ per point of Availability yet.	1 point of Availability = how many MXN/month? 1 missed delivery = how many MXN lost?
Signal quality	3/3	Ruco: "ya llego Claude, tenemos que si o si investigar." Daniel volunteered the planning framing unprompted. Brothers aligned. Short decision chain.	None.

Composite score

12 / 15

Up from 10 / 15 after Apr 22 updates.

Priceable today?

Not yet

Workflow scope still needs narrowing (ETA vs sábana regen).

02 ICP Fit

From docs/icp.md V3. A lead needs 7+ to pursue, 10+ for strong fit.

Condition	Score	Evidence
C1. ERP invested but underused	3/3	Microsoft GP confirmed. Daily workflows bypass it (Excel sabana, Pulsar, manual cotizador external).
C2. Back-office processing complexity	3/3	100% made-to-order. Complex per-machine rules (diameter x threading x cap, color sequencing). Rules undocumented.
C3. High-volume manual workflows	2/3	120M tubes/year at 500 employees is industrial scale. BUT the high-volume pain is operational (scheduling), not commercial. Daniel: "el cotizador funciona bien."
C4. Multi-party coordination	3/3	Planner + sales + plant operators + ERP admin + customer. Explicit multi-party visibility gap.

Total

11 / 12

Strong fit.

Cluster

Industrial Supply Chain (packaging for CPG).

ICP rubric gap flagged. C3 scored 2 not 3 only because the rubric assumes commercial transaction volume drives ROI. Here the ROI driver is operational volume (scheduling decisions, plant hours). ICP V4 should include operational-intensity clients. Log against docs/icp.md.

03 Engine Fit CRITICAL

Read product/architecture-spec.md v1.0 and product/unified-product-mvp.md. The platform is a request-processing pipeline:

INGEST → PARSE → EVALUATE → ROUTE → EXECUTE

Every module currently in the registry (A1, A2, A3, A4, A5, A6, B1, B2, B3, B4) fits this shape: something arrives, parse it, apply rules, route, act.

La Cazadora's Planner Co-Pilot does not fit cleanly

It is state-driven and cadence-based, not request-driven.

Inputs: order book + machine capacity + current inventory + business rules + Pulsar data
Trigger: daily clock, or state change (new order confirmed, machine status change)
Output: proposed 15-day schedule + ETAs per order, surfaced to sales and planner

There is no incoming "request" to ingest. There is no single sender. The trigger is the passage of time or a change in state.

Three options

Option A: Force-fit the current engine Recommended for MVP

Treat "new confirmed order" as the INGEST event. PARSE the order, EVALUATE against current schedule and capacity, ROUTE to planner or auto-send ETA to sales, EXECUTE = publish ETA + update schedule proposal.

+ No engine change. Fastest to MVP. Clear pricing.

− Only covers ETA-per-order. Misses the 15-day schedule regen (which is the bigger pain).

Option B: Extend the engine to support state-driven workflows Concurrent, strategic

Add a second workflow class: cadence triggers (cron), state-change triggers (ERP event, inventory threshold), aggregate-input pipelines.

+ Unlocks MAGG B1 (daily procurement), MAGG B2 (preventive maintenance thresholds), MAGG B3 (cadence dashboard), Eurostar B4 (SLA escalations). Opens a full Track B class.

− Engine work before client work. Risk of over-building before signal is validated.

Option C: Separate product stack for Planner Co-Pilot Reject

Build it outside the unified engine.

+ No engine coupling risk.

− Diverges the product. Negates the "one platform" positioning.

Recommendation: A first, B concurrent

Sell La Cazadora the ETA-per-order slice as the MVP (force-fits current engine, clear scope, clear pricing). In parallel, start the state-driven engine extension work (partially justified by MAGG already, La Cazadora makes it binding). Full-schedule regen ships in phase 2 once the engine supports it.

Engine extension is a Track B platform decision, not a La Cazadora decision. Validate with Raffaello this week.

04 Portfolio Pattern Match

Subagent scanned all 12 client folders.

Client	Match	Evidence
MAGG	Adjacent	"Production runs approved based only on historical sales, no forward demand signal" + "Production at 100% capacity." Procurement + maintenance pain. Not Planner Co-Pilot shape.
Eurostar	Adjacent, inferred	Reactive PM + technician knowledge in heads. After-sales, not factory scheduling.
Lamosa, Aronlight, AutoalDia, BCP	No match	All Track A commercial workflows. No scheduling pain.
Agro Super, Eurocomponents, PepsiCo, Nedelko, Futura Labs	Early stage	Not enough data.

Confirmed matches

1 / 12

La Cazadora itself. n=1.

Adjacent

2 / 12

MAGG + Eurostar.

La Cazadora is the first confirmed Planner Co-Pilot signal. Not yet a pattern. If it closes and MAGG maintenance ships, pattern becomes n=2 and worth cataloging in Venture Signal.

05 Existing Offer Overlap

Scanned all live proposals.

Client	Modules pitched	Overlap with Planner Co-Pilot
Lamosa	A1 + A6	None
Aronlight	A2 + A5	None
MAGG	B1 + B2 + B3	None (procurement/maintenance, not scheduling)
Eurostar	B2 + B4	None
AutoalDia	A4	None
BCP	AI Readiness HR	None

Zero overlap. Planner Co-Pilot is a net-new workflow offering. No existing proposal touches production planning, capacity, ETA, or plant-to-sales visibility.

06 Competitor Coverage

Scanned docs/competitors/*.md (12 files).

Competitor	What they do	Addresses LaCaz pain?
Smartbase	Inbound PO entry to ERP	No
Mercura	Outbound RFQ / quote automation (YC W25)	No
Handle	Commercial WhatsApp / email for insurance	No
Tenex	AI strategy consulting	No
Vendavo	Pricing governance	No
Clarinet, Faction, LeadSales, Torrenegra, Treble, AI Readiness Ent	Commercial / workforce / strategy	No

Gap confirmed. No documented competitor in our tracked set solves production planning + plant-to-sales visibility for mid-market LATAM manufacturers. Enterprise tools exist (SAP IBP, Oracle SCM, o9) but none in our price band or timelines. The mid-market LATAM band is effectively open.

07 Venture Signal Status

Scanned venture-signal/2026-03-27-report.html and index.html. Catalogued patterns: A1 through A6 (commercial) and B1 through B4 (ops intelligence: procurement, maintenance, SKU dashboard, after-sales).

No Planner Co-Pilot / production planning pattern catalogued. If La Cazadora proceeds to proposal, add Track B, Production Planning & Plant Visibility. Until then, leave as "candidate pattern, n=1, under validation."

08 Pain Point Analysis

Scanned docs/client-painpoint-analysis.html. Production planning / scheduling / plant-sales visibility is not listed as a cross-client pain. MAGG capacity and Eurostar PM are the closest adjacencies, neither maps.

09 Proposed B5 Module Registry Entry

Draft only. Do not promote into module-registry.md until Call 2 answers the scope questions.

### B5, Production Planning + Plant-to-Sales Visibility ("Planner Co-Pilot")

What it does:
Ingests the production schedule source of truth (Excel sabana or equivalent),
machine-level business rules (compatibility, sequencing, setup time),
real-time machine state (Pulsar-like systems or ERP), and the live order
book. Outputs: (a) a proposed 15-day production schedule for the planner to
approve or edit, and (b) realistic ETA per order, surfaced to sales before
they commit a date to the customer.

Input channel: ERP order feed (state-driven) + scheduler approval UI +
machine monitoring API (Pulsar or equivalent)
ERP dependency: Microsoft Dynamics GP (La Cazadora), extensible to SAP, NetSuite, Odoo
Output: schedule proposal (planner UI) + ETA per order (sales UI + ERP write-back)

Key features:
- Business rule engine per machine (compatibility, color sequencing, setup cost)
- Capacity model per production line (extruders, printers, decorators)
- Schedule proposal with planner approval + manual override
- ETA computation exposed to sales at quote-acceptance time
- Plant-to-sales status sync (live plant state, not 2-3 day lag)
- Key-person risk mitigation: rules moved from planner's head into the system

Delivery complexity: High (10 to 14 weeks) if state-driven engine
extension first, or Medium (6 to 8 weeks) if MVP is scoped as ETA-per-order
only to force-fit current engine.

Client pipeline:
| Client       | Stage         | Phase                                  |
| La Cazadora  | DISCOVERY DONE | ETA-per-order MVP scope under validation |

Scope questions that must be answered before pricing:
- How many orders per day flow through the planner?
- How many schedule re-plans per day (not per week)?
- Microsoft GP: API access path? Read-only or read/write?
- Pulsar: what data does it expose (availability only, or piece counts)?
- Does sales UI live in our product, or write back to GP?
- Is the MVP the ETA-per-order slice, the 15-day schedule regen, or both?
- What is 1 point of OEE worth in MXN?
- What is the cost of a missed delivery date today?

10 Open Validation Questions for Call 2

Reset after Apr 22 updates. Grouped by block. Top 5 (Proceso) must get answers before pricing.

Proceso — how a decision gets made

#	Question	Why it matters	Owner
1	Walk us through the sábana end-to-end: who updates it, how often, what triggers an update, what happens when machine downtime hits mid-shift.	Exposes where stale data is hurting decisions, which is the actual MVP surface.	Daniel
2	When Pulsar shows a machine went down 2 hours ago, how long until the sábana reflects it? And how long until sales finds out?	Quantifies the inputs-not-timely gap we keep hypothesizing.	Daniel / Brandon
3	What is the worst decision you made in the last 30 days because the sábana was stale?	Converts abstract pain into a concrete $ / lost-customer anecdote.	Daniel
4	How many orders per day flow through the planner? How many schedule re-plans per day (not per week)?	Drives pricing volume tier. Distinguishes ETA-per-order scope from full sábana regen.	Daniel
5	Are the machine business rules captured in the sábana formulas, or in the planner's head?	Build effort. If formulas, we replicate. If in head, we extract.	Daniel / Brandon

Sistemas — integration surface

#	Question	Why it matters	Owner
6	Microsoft GP: API (read-only or read-write)? If none, export path (CSV, ODBC)?	Determines whether MVP writes ETA back to GP or lives as a parallel layer.	Brandon
7	Pulsar: what does it expose (availability only, piece counts, downtime reasons)? API or dashboard-only?	Drives Phase 2 scope and whether real-time capacity is a 2026 or 2027 problem.	Brandon
8	Has the team already evaluated any MMS (Manufacturing Management System) platform? If yes, what failed?	Category risk check. Are we competing against a package they already dismissed?	Daniel or Ruco

Economía — the ROI anchor

#	Question	Why it matters	Owner
9	1 point of Availability = how many MXN/month at current volume? Or: EBITDA impact of Availability 58% → 68%?	The pricing anchor. Without a MXN number we cannot quote value.	Ruco
10	Cost of a missed delivery date today (churn, penalty, rework)? Last concrete case.	Second ROI handle, for the sales UI / ETA-per-order slice.	Ruco

Non-negotiable before Call 2: the sábana Excel + OEE dashboard export + ISIMAT spec matrix, all under NDA with customer data anonymized. Without these the call is another discovery, not a scoping session.

11 Recommendation

Validate More, leaning Go

Execute this plan:

Giuseppe + Raffaello send shareable one-pager to Daniel + Ruco this week, requesting the three Excels (sábana + OEE dashboard + ISIMAT spec matrix) under NDA + async answers to the 10 questions. Propose Call 2 date.
Call 2 target: 45 min, Daniel primary, Brandon optional. Goal: walk through the sábana end-to-end, close on ETA-per-order MVP vs full sábana regen scope, get the MXN number for 1 point of Availability.
In parallel (before Call 2), two internal workstreams:
- Engine path decision: force-fit current engine vs extend to state-driven. Giuseppe + Raffaello. Product decision, not client.
- MMS category research: scan the Manufacturing Management System landscape (Epicor Kinetic, Plex, Infor CloudSuite Industrial, LATAM-native players). Are any in La Cazadora's price band? Do any solve the sábana problem out of the box? If yes, we need a positioning answer before scoping.
After Call 2: if volumes, engine path, and MMS positioning are resolved, run /pricing-workflow la-cazadora. If any unresolved, Call 3.
Do NOT send any proposal or pre-proposal until engine path + MMS positioning are closed.

Blockers to Go

If orders per day < 20, revenue math collapses. Pass.
If Microsoft GP has no API and no roadmap to one, technical risk too high. Reconsider.
If the planner refuses to codify his rules (ego or political), build is impossible. Pass.
If an MMS already covers 80% of this workflow off-the-shelf in their price band, we need to respect the category and either partner or walk. Research this before scoping.

Accelerators

Brandon is already building Python + Excel automation. We can skip the "convince them automation works" stage entirely.
Short decision chain (brothers aligned, family-run).
Ruco is the buyer. Daniel is the champion. Shortens the signed-deal timeline but forces deliverables to speak two languages.
Excel logic is fine — we replicate, we don't rebuild. The sábana's planning logic works. The gap is process-manual + inputs-not-timely, the cheaper half of the problem.
Gemini is already in their Excel workbook. Adoption friction on "use AI" is zero. Positioning is "connect your data to AI", not "try AI".
Availability is the ROI lever, not Performance or Quality. 58.60% Availability has a concrete ceiling (best-in-class ~85%) and sequencing is a software-solvable problem, not a hardware one.

12 Appendix: Skill Template Proposal

Running this analysis by hand surfaced a clean template for a repeatable skill. If we build /validate-opportunity [client], it should enforce these sections in this order:

Summary (BLUF: Go / No-Go / Validate More with one-paragraph reason)
Opportunity Clarity Score (5 dimensions, 0-3 each, 15 total)
ICP Fit (from docs/icp.md)
Engine Fit (against product/architecture-spec.md). The highest-value step.
Portfolio Pattern Match (grep all clients/*/call-notes.md)
Existing Offer Overlap (scan live proposals)
Competitor Coverage (scan docs/competitors/)
Venture Signal Status (check latest report)
Pain Point Analysis cross-ref
Proposed Module Registry Entry (draft inline, do not promote)
Open Validation Questions (ranked)
Recommendation (Go / No-Go / Validate More with blockers and accelerators)

Two novel findings worth encoding in the skill:

Engine Fit check: the architecture-spec readthrough produced the single most strategic finding (state-driven vs request-driven divergence). Without it, we would have scoped and priced the wrong thing.
Capability-gaps FEAT-2026-04-08-001 resurfaces: the /cross-check idea (call-notes vs proposal, PASS / GAP table) is a sibling skill. Both belong in the same post-discovery toolkit.

Skill build estimate: 1.5 to 2 hours. T2. No new infra.