AdapttoAI Consulting Intelligence

Venture Signal

Pattern intelligence across all active consulting engagements

GeneratedMarch 27, 2026
Last UpdatedApril 2, 2026
Clients Scanned6 engagements
Confirmed Calls4 of 6
Top Pattern Score23 / 25
Pipeline Status — loading...
See top recommendation →
Client
Status
Next Action
Blocker
Partner Sync — Open Items

One pattern dominates — and YC just funded it

Across 6 consulting clients in 5 countries — spanning manufacturing, automotive, industrial machinery, LED distribution, LED manufacturing, and healthtech — one pattern appears with near-perfect consistency: companies have modern ERP or CRM software (SAP, Salesforce, Odoo, Impulsa) but run their actual day-to-day operations on WhatsApp, email, and Excel.

The most acute version: inbound requests (discount approvals, service inquiries, product matching, purchase authorizations) arrive via email or WhatsApp and require a human to manually validate against ERP/catalog data, then respond — every single time. At Grupo Lamosa: 30–40 emails/day from one salesperson. At Aronlight: 200–300 inbound requests/day. At AutoalDia: 10–15 WhatsApp inquiries daily with no CRM for the service team.

The opportunity: an AI middleware layer that sits between the inbound channel (email / WhatsApp) and the ERP/database, applies business rules, and reduces 80% of decisions to a single click. The Lamosa engagement has already designed this architecture — "El Validador Inteligente." The product question is: can it be packaged for 1,000 companies with the same problem?

Market validation: Mercura (YC W25, $2.1M seed) just launched targeting this exact space for inbound RFQ automation. The outbound proposal generation half of the problem — 700 proposals/week at Aronlight — remains wide open. Composite score: 23/25. Confirmed in 4 of 6 clients called.

Pipeline Status — loading...

6 engagements — 4 confirmed calls

Grupo Lamosa / San Lorenzo
Proposal ↗
Ceramic tile manufacturing · Peru · Part of $1.79B enterprise (BMV: LAMOSA)
Call confirmed
  • 30–40 discount approval emails/day. Each requires: open Excel → find SKUs → compare list price vs. promotional grid → validate margins manually.
  • No system of record for why a discount was approved or rejected. Lost in email threads.
  • After approval, email forwarded to third party for manual SAP entry. 3-step process for a 1-click problem.
  • Expansion surface: same problem confirmed in Argentina, Colombia, Brazil, Spain (Baldocer). Same SAP stack.
AutoalDia / Grupo AGAS
Automotive dealership + workshop + towing · Peru · 20–50 employees
Call confirmed
  • Service team handles 10–15 WhatsApp inquiries/day with zero CRM. Every interaction lives in chat history only.
  • 300 leads/month managed in Excel and Google Drive. No automated follow-up.
  • Impulsa CRM active for sales but entire template infrastructure was lost — single point of failure (Elizabeth).
  • National vehicle logistics (cigüeñas) booked and dispatched completely manually.
Eurostar Machinery
Industrial packaging machinery distribution · Mexico · ~48 employees
Call confirmed
  • 200+ installed machines across Mexico and LatAm tracked manually in Excel. Zero automated PM alerts.
  • Salesforce + SAP paid for but barely used. Real OS: WhatsApp, Excel, and institutional memory.
  • 12-month technician onboarding. Knowledge lives in senior techs' heads — not in any system.
  • Admin team invisible to CEO. "No me interesa lo que hagan los técnicos... Es más los tiempos muertos de admin."
Aronlight
Proposal ↗
LED lighting distribution · EU · CEO: Manuel Vidigal · 1,000+ distributors
Call confirmed
  • Sales proposals generated manually through Odoo — hundreds per week. Engineering team backlogged.
  • 200–300 inbound requests/day from distributors overwhelming team. No triage, no automation.
  • 1,000+ distributors managed on WhatsApp with no automation layer and no self-service.
  • Technical product matching (50,000+ SKU catalog) done manually by engineering team per request.
MAGG / Electro Mag
NEW · MAR 29
LED lighting manufacturing & distribution · Mexico · ~900 employees · 3,000+ distributors · 1,500+ SKUs
Pre-discovery — no call yet
  • 3,000+ authorized distributors placing orders and requesting quotes via WhatsApp/email daily — no automation layer.
  • 1,500+ SKU catalog with variant complexity (wattage, finish, size) — spec matching likely fully manual.
  • Tiered distributor pricing confirmed (gated pricing on distributor portals) — discount approval governance likely manager-bottlenecked.
  • 7 branches with no visible centralized inventory visibility. "Mala organización" flagged in employee reviews.
  • All pains inferred from research. Discovery call needed to confirm.
Futura Labs / Moons
Digital dentistry · Mexico / Colombia / Peru · $27M+ funded · YC W20
Discovery only — no call yet
  • WhatsApp patient support slow — Trustpilot 1.5/5 (84 reviews). Classic scaling-faster-than-ops signal.
  • B2B partner (dentist) outreach likely manual — 1,000+ partners, WhatsApp-driven.
  • 300+ leads/month for B2C appointments — semi-auto at best.
  • No discovery call completed. All pains inferred from research, not confirmed.

Confirmed = pain validated in discovery call or proposal.  Inferred = from pre-call research only. Confirmed carries 3× weight in scoring.

Pattern Frequency Matrix

Recurring pain types — ranked by frequency

Pattern Type Confirmed in Inferred in Score
Manual email/WhatsApp request processing
Request arrives → human manually validates vs. ERP/catalog → replies. Every single time.
Workflow automation Lamosa, AutoalDia, Aronlight Eurostar, MAGG ~ 11
ERP/CRM adoption gap
Paid software (SAP, Salesforce, Odoo, Impulsa) sits unused while ops run on WhatsApp + Excel
Workflow automation Eurostar, Lamosa, AutoalDia Aronlight, MAGG ~ 11
WhatsApp as unstructured B2B channel
Primary channel for service, partner management, or distributor ops — no automation, no logging
Workflow automation AutoalDia, Aronlight, Eurostar Futura Labs, MAGG ~ 11
High-volume proposal/quote generation
Proposals generated manually per request — volume too high for current team structure
Content generation Aronlight, Eurostar AutoalDia, MAGG ~ 8
Single-person knowledge dependency
Critical process owned by one person — no system, no docs, no fallback
Data pipeline AutoalDia (Elizabeth), Eurostar (senior techs) Lamosa 7
No post-decision traceability
Decisions made but not recorded — audit trail lives in email threads or nowhere
Data pipeline Lamosa, AutoalDia Eurostar 7

Frequency score = (confirmed clients × 3) + (inferred clients × 1). Confirmed pains from actual calls or proposals only.

Opportunity Scoring

5 attributes — max 25 composite score

Pattern Recurring
Revenue
Scal-
ability
Buyer
Access
Time to
Value
Expansion
Surface
Total
Manual Email/WhatsApp
Request Processing

5

5

4

5

4
23
WhatsApp as Unstructured
B2B Channel

5

4

4

5

4
22
ERP/CRM Adoption Gap

4

4

3

3

5
19
High-Volume Proposal
Generation

4

3

3

4

3
17
Single-Person Knowledge
Dependency

4

3

2

2

3
14
Recurring Revenue
Monthly SaaS vs. one-time build?
Scalability
Same solution for 100 clients?
Buyer Access
Self-serve vs. 18-month cycle?
Time to Value
Same week vs. 12+ months?
Expansion Surface
Land one team, expand to group?

The market, the gap, and the funded competition

1. Manual Email/WhatsApp Approval Workflow Automation
Inbound approval requests arrive via email or WhatsApp → manager manually cross-references ERP pricing grid → responds. 20–300 per day depending on company size.
Company What they do Funding Pricing Gap
Pipefy
pipefy.com
Low-code workflow platform with approval routing and ERP integrations. Brazilian company, LATAM presence. ~$153M total. Series C $75M Oct 2021, SoftBank Latin America Fund lead. [TechCrunch] From $24/user/mo Requires requests to enter their portal — doesn't intercept and triage the email itself. No AI reasoning layer. Rule-based only.
Kissflow
kissflow.com
Approval workflow builder with multi-level routing, procurement modules. $1M seed, Indian Angel Network, 2012. [Tracxn] Starts at $1,500/mo flat Generic form-based approvals. No ERP data integration for conditional approval validation.
ApproveThis
approvethis.com
Lightweight approval layer running over Slack/Teams/email. Approvers need no license. Unknown Usage-based, undisclosed Structured routing only — no AI understanding of what's being requested, no ERP lookup, no semáforo logic.
SAP AI / S/4HANA Fiori Native SAP approval flows with AI add-ons (Joule, Agentforce). Public company Enterprise licensing, $250K+ implementations Requires SAP to already be the source of truth and 90%+ adopted — useless for companies where the ERP sits idle.
Adjacent Players
ServiceNow — Full workflow suite, but $250K+ enterprise contracts. Out of range for mid-market manufacturing in LatAm.
Monday.com / Asana — Have approval features but are project tools, not email-native approval triaging against live ERP data.
TranscendAP — AP invoice approval automation. Adjacent but finance-only (invoices), not general operational approvals like discount requests or PO exceptions.
The Gap

None of the above products intercept an unstructured email approval request, automatically retrieve the relevant ERP/pricing record, and present a pre-validated recommendation inline so the manager can approve in one click from their inbox — no portal login required. Pipefy and Kissflow require requests to enter their system. ApproveThis has no AI reasoning. SAP tools require full SAP adoption. The missing product is an AI layer that works on top of whatever currently exists (email + Excel + ERP) without requiring behavior change. This is especially acute for LATAM and Southern European mid-market companies where ERP adoption is partial and operational communication is entirely email/WhatsApp. The Lamosa "Validador Inteligente" architecture fills this gap exactly.

Market Size Signal

Workflow automation market: $23.77B in 2025 → $37.45B by 2030 at 9.52% CAGR. [Mordor Intelligence]. No clean TAM exists for "email-native approval triage" specifically — this is a feature gap inside a large market, not a named category yet. SAP Business One alone has ~30,000 mid-market customers globally. Even 0.5% penetration at $800/month = $144M ARR.


2. AI Middleware Between WhatsApp and ERP/CRM
Companies run operations in WhatsApp while SAP, Salesforce, or Odoo sit idle. No live two-way connection between the channel where work happens and the system that holds the data.
Company What they do Funding Pricing Gap
WATI
wati.io
WhatsApp Business API SaaS — team inbox, chatbot builder, CRM sync. $35M total. Series B $23M Oct 2022, Tiger Global lead. [TechCrunch] $79/mo for 5 users UI chatbot builder, no ERP integration. Can't validate requests against catalog or pricing data. Rule-based, not AI-reasoning.
Respond.io
respond.io
Omnichannel messaging hub (WhatsApp + email + Instagram) with AI agents and CRM logging. $8.8M total. Series A $7M Sep 2022, Headline lead. [Respond.io blog] Per-seat + conversation-based CRM-lite for messaging. No domain-specific intelligence. Treats all conversations as support tickets, not business workflows.
Yalo
yalo.ai
Conversational commerce for WhatsApp at enterprise scale (Walmart, Nike, Aeromexico). $97M total. Series C Dec 2023. [Crunchbase] Enterprise, custom Targets large enterprise consumer-facing commerce. Wrong fit for mid-market B2B operational workflows.
Treble.ai
treble.ai
WhatsApp automation for marketing and sales. Bogotá-based, strong LATAM focus. $15M. Series A $15M Jul 2022, Tiger Global + Twilio Ventures. [Tracxn] Undisclosed subscription Marketing automation tool — broadcast, drip sequences. Not B2B operational workflows or ERP query layer.
Adjacent Players
Harmonix AI — Bridges WhatsApp + LinkedIn + calls into Salesforce. Requires Salesforce as the backbone. Doesn't work if the team is Salesforce-light or on SAP/Odoo.
Zapier / Make — Can connect WhatsApp (via Twilio) to Odoo/Salesforce but require technical setup, no conversational intelligence, no natural language querying of ERP data.
The Gap

All current players treat WhatsApp as a customer-facing support or marketing channel. None provide a two-way operational layer where a sales rep can type "¿qué stock tenemos del SKU 45143?" in WhatsApp and have the system fetch it from SAP in real time. The missing product is an AI agent that speaks ERP fluently and lives inside WhatsApp — not a CRM that syncs messages after the fact, but a live query/action layer that makes the ERP accessible without a login. This gap is most acute in mid-market LATAM companies where WhatsApp is the operating system but SAP/Odoo/Salesforce sit idle. None of the funded players (WATI, Respond.io, Treble.ai) have built this bridge.

Market Size Signal

Conversational commerce in LATAM: estimated $18.2B in 2025, 35% YoY growth, ~72% flowing through WhatsApp. [AuroraInbox]. Over 40% of Mexican businesses conduct procurement via digital channels, up from 25% in 2021. B2B e-commerce in Brazil grew 26% in 2025. WhatsApp Business economy estimated at $45B globally. [Invent]


3. High-Volume B2B Proposal Automation for Distributors YC W25 just funded this
B2B distributors generate 100–700 proposals/week manually. Each requires looking up catalog, pricing rules, and client-specific parameters in ERP. Aronlight is the extreme case.
Company What they do Funding Pricing Gap
Mercura YC W25
mercura.ai
AI quote and order automation for distributors and manufacturers. Reads inbound RFQ emails, auto-matches to catalog with pricing and confidence scores. SAP, Oracle, Odoo integrations. $2.1M seed. YC W25 + TQ Ventures + SignalFire. [Startbase] Undisclosed Targets inbound RFQ email automation (manufacturer receiving). The LATAM outbound problem — sales rep proactively generating 700 proposals/week from Odoo + Excel overrides — is not their focus.
PandaDoc
pandadoc.com
Document automation with AI proposal generation, CRM integration, e-sign. ~$106–118M total. [The SaaS News] $35–65/user/mo General sales proposals for agencies and professional services. No ERP integration. Not built for distributor catalogs or SKU-level pricing logic.
PROS CPQ
pros.com
Enterprise CPQ with AI pricing optimization, SAP-native integration. Public company (NYSE: PRO) Enterprise, $100K+/yr Targets large enterprise manufacturing and distribution. Inaccessible and overengineered for mid-market LATAM. 6–12 month implementations.
DealHub
dealhub.io
AI-powered CPQ + subscription billing. $45M total. Last round $25M Jun 2021. [LeadIQ] Enterprise, custom Targets mid-market to enterprise SaaS and tech companies. Not distributor catalogs — wrong use case entirely.
The Gap — and Why Mercura Validates the Opportunity

Mercura raising $2.1M from YC confirms the space is real and early. But their focus is inbound RFQ automation for manufacturers receiving quote requests. The gap they leave open: the mid-market LATAM/Southern European distributor who needs to proactively generate outbound proposals — where the sales rep pulls pricing from Odoo, applies client-specific margin rules stored in Excel, generates the PDF, and sends it. PandaDoc and Proposify have no ERP integration. DealHub targets SaaS companies. PROS targets large enterprise. No current product solves the outbound proposal flow for Odoo mid-market distributors with 500–50,000 SKUs in LATAM. Aronlight is the exact wedge client to prove this architecture.

Market Size Signal

CPQ global market: $3.14B in 2025 → $7.55B by 2031 at 15.74% CAGR. Manufacturing leads adoption at ~31% market share. [Mordor Intelligence]. AI CPQ implementations report 75% reduction in quote generation time and 23% increase in close rates. [MobileForce]. Large enterprises hold ~71% of current CPQ market — mid-market remains underserved and is the primary growth vector.

Cross-Pattern Insight
Pattern 1 (inbound approval bottleneck) and Pattern 3 (outbound proposal bottleneck) are two sides of the same problem at companies like Grupo Lamosa and Aronlight. A single product addressing both — with Odoo/SAP as the data layer and email/WhatsApp as the UX layer — would be differentiated against every competitor listed above, including Mercura.

The Funded Competition
YC W25 · $2.1M SEED

If Mercura called our clients tomorrow — what would happen?

For each client: the Mercura feature that comes closest to their pain, and exactly where the product breaks down. This is the map of where AdapttoAI wins.

Mercura Intelligence Brief — Who They Actually Sell To
Known Clients (15 confirmed · 40+ claimed)
BME Group Bauder GmbH Reisser Gruppe Sanitär-Heinze Stark Deutschland Siteco Hauff-Technik Kessel AG Zander-Gruppe Wiedemann SHK DEWEtech Schönreiter Tork Systems SG (unknown)
87%
Germany
0%
LatAm
SAP
ERP required
Verticals: HVAC/plumbing wholesale (27%) · Building materials (27%) · Industrial/technical (33%) · Lighting (13%). All one use case: inbound RFQ from external customers.
Their ICP

Mid-market to enterprise B2B wholesale distributors and manufacturers in DACH (Germany-first). Minimum bar: SAP ERP + hundreds of daily inbound RFQs. An external customer sends a spec document — Mercura reads it and generates the quote back. 100% Europe. Zero LatAm. Zero US (expansion "planned", not live).

5 Structural Gaps — Where AdapttoAI Wins
BY DESIGN
Geography. Zero LatAm presence. Peru, Mexico, Chile, Colombia are structurally unreachable — their product requires SAP certification + German compliance posture.
BY DESIGN
Internal approval workflows. Mercura solves external customer → company. We solve internal sales rep → manager → ERP. Different workflow, different product.
BY DESIGN
Non-SAP ERPs. Odoo (Aronlight), Impulsa (AutoalDía), NetSuite (MAGG), custom systems — not their target, never will be. Their 3-day implementation claim is SAP-specific.
BY NEGLECT
Track B / Ops intelligence. 100% Track A. Nothing on procurement intelligence, maintenance tracking, or SKU analytics. MAGG's pain doesn't exist in their world.
BY NEGLECT
Outbound proposal generation. Mercura handles what comes IN from customers. Aronlight's 700 proposals/week go the other direction — not their product.
Best Mercura Fit
Eurostar Machinery
Industrial machinery · Mexico · SAP + Salesforce
Eurostar gets inbound quote requests for capital equipment by email. They represent 13+ European brands, each with separate specs and pricing. Mercura's email automation + AI Copilot would directly address this: extract line items from inbound emails, match to catalog, push structured quote to SAP or Salesforce. Both integrations exist. The semantic search handling "synonyms, typos, vague descriptions" is exactly what their team deals with across multilingual European specs.
Where Mercura Falls Short
  • No after-sales or PM automation. Mercura generates quotes — it doesn't track 200+ installed machines, fire PM alerts, or manage service scheduling. That's 80% of Eurostar's pain.
  • No technician knowledge base. Mercura doesn't address the 12-month onboarding problem or knowledge trapped in senior techs' heads.
  • No admin visibility. Ferdinando's core complaint — "los tiempos muertos de admin" — is not a quoting problem. Mercura won't surface what's stalled or invisible to the CEO.
  • DACH focus, no Spanish. No LatAm presence, no Spanish-language product, no context for Mexican mid-market dynamics.
Partial Fit — Missing the Core
Grupo Lamosa / San Lorenzo
Ceramic tile manufacturing · Peru · SAP
Mercura's email automation would correctly identify the inbound discount request emails from Jorge Zapata, extract the SKU, quantity, and requested price, and push a structured record to SAP. The intake and extraction piece works. SAP integration exists.
Where Mercura Falls Short
  • No approval governance. Mercura generates a quote and pushes it to ERP — it does not ask for manager approval first. The entire Lamosa problem is that every discount needs a human decision before it goes to SAP. Mercura skips that step entirely.
  • No Verde/Ámbar/Rojo logic. Mercura has no rules engine for conditional escalation — no concept of "auto-approve if margin > 20%, escalate if margin 15–20%, auto-reject below."
  • No rejection traceability. Mercura doesn't record why a discount was denied — the traceability gap that Lamosa explicitly named as a problem.
  • No Spanish, no LatAm pricing context. Malla Promocional, promotional grids, m² pricing logic — none of this is in Mercura's training data.
Partial Fit — Wrong Direction
Aronlight
LED lighting distribution · EU · Odoo · 1,000+ distributors
The specifications module is the closest fit: Aronlight receives complex product matching requests with vague descriptions across a 50,000+ SKU catalog. Mercura's semantic search — built to handle "synonyms, context, typos" — would help the technical matching team. The AI Copilot for researching products also maps to their engineering team's daily workflow.
Where Mercura Falls Short
  • No Odoo integration. Aronlight runs on Odoo. Mercura's integration list — SAP, NetSuite, Salesforce, Dynamics — does not include Odoo. Non-starter.
  • Inbound only — Aronlight's pain is outbound. Mercura processes inbound requests. Aronlight's 700 proposals/week problem is generating and sending outbound proposals proactively. Opposite direction.
  • No WhatsApp for 1,000+ distributors. Mercura has no WhatsApp channel. Aronlight's distributor network operates entirely on WhatsApp — no automation layer means no self-service for the bulk of their partner base.
  • No client-specific margin rules. Aronlight's pricing involves distributor-level margin overrides stored in Excel. Mercura's pricing engine syncs ERP catalog pricing — it doesn't handle off-catalog margin exceptions.
No Fit
AutoalDia / Grupo AGAS
Automotive dealership + workshop · Peru · Impulsa CRM
No applicable Mercura feature. AutoalDia's core pain is WhatsApp service triage and CRM template reconstruction — neither of which Mercura addresses. Mercura is built for B2B distributors and manufacturers processing multi-line product quotes. An automotive workshop handling 10–15 WhatsApp service inquiries per day is a completely different workflow, channel, and ERP context (Impulsa has no Mercura integration).
Where Mercura Falls Short
  • No Impulsa integration. Their CRM (Impulsa) is not on Mercura's integration list and likely never will be.
  • No WhatsApp channel. 100% of AutoalDia's service ops runs on WhatsApp. Mercura doesn't support it.
  • Wrong industry entirely. Mercura is construction supply chain. Automotive service and towing is not a use case they've built for or positioned toward.
Bottom Line
Eurostar
Mercura could win the quoting piece. AdapttoAI wins everything else: PM automation, technician knowledge, CEO visibility.
Lamosa
Mercura reads the email. AdapttoAI decides what to do with it. The approval layer is ours — Mercura has no concept of it.
Aronlight
Mercura handles inbound. Aronlight's 700 proposals/week goes the other direction. No Odoo. No WhatsApp. Three blockers.
AutoalDia
Zero overlap. Wrong channel, wrong ERP, wrong industry. Not a competitor here in any scenario.

The Enterprise Incumbent
PRIVATE · EST. $84-100M ARR

Vendavo — what the enterprise already pays for

Vendavo is the $100K+/year pricing optimization platform for companies like Emerson, Ford, and Volvo. They solve the same margin problem AdapttoAI solves — but for enterprises with SAP, a pricing team, and 12 months to implement. They define the ceiling; we define the entry point.

Vendavo Intelligence Brief — Who They Actually Sell To
Known Clients (22 confirmed · enterprise-only)
Emerson Ford Volvo Trucks Cencora Molson Coors Grundfos Huhtamaki Braskem Xylem Eastman Chemical Yamaha Motor GAF DeLaval Fluidra + 7 more
55%
USA
$1B+
Revenue floor
SAP
ERP required
Why This Is Good News for AdapttoAI

Vendavo proves the pricing governance problem is a real, funded category — $84-100M ARR selling it to the enterprise. They don't go below $1B revenue. Their implementation takes 6-18 months with an SI partner. Their LatAm presence = one Brazilian petrochemical company.

BY DESIGN
Zero mid-market. $100K-500K/year contracts + SI partner = automatic filter. Lamosa, AutoalDía, MAGG don't qualify.
BY DESIGN
No WhatsApp/email workflow. Assumes reps are in Salesforce or SAP. No concept of an approval request arriving via WhatsApp.
BY DESIGN
LatAm mid-market = zero. No Spanish materials, no local partners, no LatAm case studies outside Brazil enterprise.
BY DESIGN
6-18 month implementation. No self-serve, no fast deployment. Companies wanting results in 30 days are not their market.
Talking point: "Vendavo is for when you already have SAP and a pricing team. If you're running approvals on WhatsApp today, we are the right next step — not them."

The LatAm WhatsApp CRM Players
LEADSALES $3.7M · WHATICKET BOOTSTRAPPED

Leadsales + Whaticket — the WhatsApp CRM floor

These are the closest geographic competitors — LatAm-native, WhatsApp-first, and growing. But they solve a different layer: shared inbox + lead funnel, not approval governance. Understanding where they stop is understanding where AdapttoAI starts.

Leadsales
$3.2M ARR · 2,500+ clients
WhatsApp shared inbox + visual sales funnel for LatAm SMB sales teams. Meta Business Partner. Mexico-first, expanding across LatAm. Entry: $97/mo for 3 users.
Known clients
Univ. ICEL SCHIRP Asesores Reluvsa Autopartes Family Fitness iGrow Academy albo + 2,494 unnamed
Mexico
Primary market
SMB
<20 person teams
Whaticket
BOOTSTRAPPED · OPEN SOURCE
Multi-agent WhatsApp shared inbox for micro-businesses. Brazilian origin. Open-source community version on GitHub. Lowest price in segment: $39-49/mo. Price-sensitive buyer, 1-10 agents.
Known clients (LIMITED DATA)
Caribe Sur Store Dakati
Brazil
Origin market
$39/mo
Entry price
The Ceiling — Where They Stop, We Start
They validate LatAm WhatsApp willingness to pay. Leadsales hit $3.2M ARR at $97-247/mo — that's 1,300+ paying companies running sales ops on WhatsApp. The behavior is real.
No ERP integration. Neither connects to SAP, Odoo, or NetSuite. A discount request requiring a margin check against ERP data is outside their product surface entirely.
No approval chain. They route conversations and assign agents. There is no concept of "this request requires manager approval before the rep can quote." That governance layer doesn't exist.
SMB ceiling. Leadsales' largest case study has 32-40 sales reps. A mid-market company with regional managers, tiered approval hierarchies, and 200-rep teams outgrows them immediately.
Talking point: "Leadsales organizes your WhatsApp. We make it a decision engine. They give you a shared inbox — we intercept the approval request, check the margin, and enforce the rule. Different layer, not competing tools."

The LatAm WhatsApp Automation Leader
YC S19 · $15M · 2,000+ CLIENTS

Treble.ai — the category leader we're not competing with

Treble.ai is what WhatsApp marketing automation looks like when you build it right: YC-backed, Tiger Global-funded, Meta official partner, 2,000+ clients in LatAm. They dominate the campaign + lead-qualification layer. The gap: they stop exactly where AdapttoAI starts — approval governance, ERP integration, commercial operations.

Treble.ai Intelligence Brief — Who They Actually Sell To
Known Clients (14 confirmed · 2,000+ claimed)
Rappi RappiPay Renault LatAm Platzi EF Education Xiaomi Dentalia Kovi Decreditos Sura Addi Wom SmartBeemo Guru
100%
LatAm
HubSpot
CRM (not ERP)
$5.7M
ARR (2024)
Verticals: Fintech (29%) · Automotive (21%) · Education (21%) · Healthcare (7%) · Other (22%). Colombia-first, expanding across LatAm. Series A led by Tiger Global + Twilio Ventures.
ICP overlap signal: Renault LatAm (dealer network) + Decreditos (automotive credit dealer comms) + Kovi (mobility) are very close analogs to our pipeline companies. They use Treble for outbound campaigns — the approval governance layer doesn't exist there yet.
Where Treble Stops — Where AdapttoAI Starts
BY DESIGN
No ERP integration. Treble connects to HubSpot and Salesforce. No SAP, Odoo, NetSuite. Pricing rules and margin data are invisible to it.
BY DESIGN
No approval chain. Treble automates message flows — campaign → qualify → route. There is no concept of "this request requires manager approval before the rep can quote."
BY DESIGN
Marketing budget, not commercial ops. Treble's ROI is lead conversion rates. Our ROI is manager hours saved + margin points recovered. Different buyer, different budget.
OPPORTUNITY
Treble graduates become AdapttoAI leads. A company that outgrows Treble's informal WhatsApp and needs structured approval governance is the exact conversion moment AdapttoAI should be present for.
Talking point: "Treble automates your WhatsApp campaigns. We govern your WhatsApp decisions. They send the message — we intercept the one that comes back asking for a 18% discount, check the margin against your ERP, and enforce the rule."

The opportunity

Single Best Opportunity — Composite Score 23/25
UPDATED · MAR 28
AI Approval Middleware for Companies
That Have ERP but Run on Email

The scene, at every client:

A salesperson sends a WhatsApp message or emails an Excel attachment: "Can I give this client 18% on SKU 45143?" The manager opens the file, finds the right tab, eyeballs the margin column, checks the promotional grid in a second spreadsheet, and types back "OK" — or doesn't respond for three hours because they're in a meeting. The salesperson follows up. The client is waiting. This happens 30 to 40 times a day.

Why the ERP doesn't solve it:

Every one of these companies has an ERP. SAP, Salesforce, Odoo, Impulsa. The ERP has the pricing rules, the stock levels, the margin data. But no one opens the ERP to answer a discount email. It would require logging in, navigating to the right module, looking up the SKU, and cross-referencing three screens. By the time a manager does all that, it's faster to just remember the rule from experience. So the ERP becomes a system of record for what happened — not a tool that shapes what's happening. The actual decisions run on institutional memory, Excel tabs, and WhatsApp.

The architecture — already proven with Lamosa:

Email or WhatsApp arrives
  ↓
AI reads it → extracts SKU, quantity, requested discount, requester
  ↓
ERP queried automatically → margin, stock level, promotional grid retrieved
  ↓
Verde / Ámbar / Rojo classification applied
    🟢 Verde: within rules → auto-approved, ERP updated
    🟡 Ámbar: borderline → manager gets a 1-line summary + one-click Approve/Reject
    🔴 Rojo: outside all parameters → auto-rejected, salesperson notified with reason
  ↓
Every decision logged: who requested, what rule triggered, what the manager decided

The manager never opens the ERP. The salesperson gets a faster answer. Every decision is traceable. The tool fits the way people already work — no behavior change required. First results are visible in week 1.

Why Mercura validates this — and where the gap is:

Y Combinator just backed Mercura ($2.1M, W25) to solve a related but different problem. In Mercura's world, an external customer sends a request for quote (RFQ) to the company. The AI reads it and generates the quote document that gets sent back. It's customer-to-company communication, automated.

What we're solving is different. Here, the salesperson already has a customer on the line and needs permission from their own manager before they can close. The request travels internally — from sales rep to commercial manager — and it needs a decision, not a document. Mercura has no answer for this.

What internal approval governance actually means:

Right now, "approval governance" at these companies is: the manager's judgment, stored in their head, delivered by WhatsApp. No rules written down. No audit trail. No way to know whether the same discount request gets a different answer on a Tuesday versus a Friday. Approval governance means turning that into a system: define the rules once (margin floors, stock thresholds, exception SKUs), enforce them automatically, and log every decision with the reason. The manager stops being a bottleneck and becomes a reviewer — only called in when the situation is genuinely ambiguous. That layer, built on top of the ERP data companies already have but never expose, is what no one has productized for the mid-market.

Build Roadmap

Productization path

Productization Path
Stage 1 — Now
Consulting Anchor

Build for Lamosa, Eurostar, Aronlight as custom engagements. $15–30K per build. Extract the reusable components. Lamosa IS the MVP client — architecture already scoped.

Stage 2 — 6–12 months
One Platform, Two Tracks

Package as a single platform with one ERP connector built once and reused across all modules. Track A (Commercial): email/WhatsApp intake + rules engine + approval UI. Track B (Ops): procurement intelligence + maintenance tracking + SKU performance. The connector is the moat — once plugged into a client's SAP or Odoo, each new module is a sales conversation, not a new integration. Three tiers, billed annually: Core $18K/yr (1 module, up to 500 req/month) · Growth $42K/yr (multi-module, 2,000 req/month) · Enterprise $78K/yr (multi-subsidiary, ERP write-back, unlimited). A manager spending 3+ hours/day on approvals costs $20–32K/year in labor — Core pays for itself in weeks.

Stage 3 — 18–36 months
Scale

Land in one company on one module, expand across both tracks and all subsidiaries. MAGG is the first client where both tracks apply simultaneously. Lamosa Peru → Argentina → Spain. Franchised rollout through SAP/Odoo/NetSuite implementation partners — they do the connector, we do the modules.

Interactive Prototype — Live Now
Raffa — Italy review needed
3-Workflow MVP
Discount Approval · Quote Request · Spec Matching. Notification inbox, dashboard, rules config, and system architecture overview. Shows exactly what the product looks like before a line of production code is written.

What to build and how to price it

Architecture Principle — One Platform

Track A and Track B are not two products. They are two surfaces of one platform, sharing a single ERP connector. The connector is built once per client — every additional module is a sales conversation, not a new integration project. That is the moat.

Input channels
Email · WhatsApp · Voice
Where requests arrive
ERP Connector — built once
SAP · Odoo · NetSuite · Impulsa
The moat. Reused across all modules.
Track A — Commercial
Approval routing · Quote gen · Spec matching
Track B — Ops
Procurement intel · Maintenance · SKU performance
Entry on Track A (Lamosa, Aronlight)

Quick win on commercial approvals. ERP connector built. Track B (ops intelligence) becomes a natural upsell 3 months later.

Entry on Track B (MAGG)

Ops pain is the door opener — NetSuite timing creates urgency. ERP connector built. Track A (distributor approvals) is Phase 2 once inside.

Pricing Benchmark
NEW · MAR 28

No competitor publishes pricing. Estimates below are derived from job postings, ARR/customer count back-calculation, and category benchmarks.

Player What they solve Motion Est. ACV Market
Mercura External RFQ automation (customer sends quote request, AI responds) Sales-led, demo-gated $25K–60K/yr DACH, German-speaking
Vendavo / Pricefx Enterprise CPQ and pricing optimization Enterprise, 6-18 month cycles $100K–150K/yr US/EU enterprise
WATI / Respond.io WhatsApp messaging platform (no ERP logic, no approval rules) Self-serve / PLG $1.2K–6K/yr Global SMB
Our target position Internal approval governance — salesperson asks manager, AI pre-filters, manager one-clicks Consulting-led → SaaS, annual billing $18K–78K/yr
Core $18K · Growth $42K · Enterprise $78K
LatAm + EMEA mid-market

The pricing gap is the opportunity. Mercura starts at ~$25K/year and requires a 3-6 month enterprise sales cycle with German-speaking AEs. At $18–78K/year (annual contract, three tiers), we price on ROI — not on what feels safe. A manager spending 3+ hours/day on approvals costs $20–32K/year in labor alone; the Core tier pays back in weeks. The ceiling of our range ($78K) sits just above Mercura's entry point, creating room to grow upmarket without conceding the mid-market. All tiers are billed annually — standard for ERP-integrated SaaS with meaningful onboarding. The first 3 consulting clients fund the build; the fourth client pays for the product.

Workflow Catalog
NEW · MAR 28

What to build, for whom, in what order, and how to price it. ✓ = confirmed in call or proposal  ·  ✗ = wants it but blocked from using Mercura  ·  ~ = inferred, not yet confirmed

Client Demand & Competitor Coverage
Workflow Lamosa Aronlight Eurostar AutoalDia MAGG ~ Mercura Priority
Inbound Request → Quote
Aronlight Module 1 — confirmed in proposal
HIGH ✓ MOD HIGH ~ Partial* MVP — Aronlight
Spec / Product Matching
Aronlight Module 2 — confirmed in proposal
HIGH ✓ MOD HIGH ~ Partial* MVP — Aronlight
Discount / Price Approval
Lamosa design partner track
HIGH ✓ MOD ~ MOD HIGH ~ No MVP — Lamosa
Escalation Logic (add-on: Naranja/Rojo path for Lamosa) MOD LOW MOD MOD ~ No Phase 2
PO / Supplier Approval MOD MOD HIGH ✓ MOD ~ No Phase 2
Credit / Terms Approval MOD HIGH MOD ~ No Phase 2
Returns / Credit Note MOD MOD MOD MOD ~ No Phase 3
AI Copilot (Quote Building) MOD ✗ MOD MOD ~ Partial Phase 3
Voice Processing Yes Potential ~

* Mercura covers Inbound Request → Quote and Spec Matching but has no Odoo integration and no WhatsApp — making their product unusable for Aronlight. We are building these workflows for Aronlight directly. "Partial" = Mercura covers the concept but not for this client. ~ = guessed, no confirmed client use case yet.

Track B — Separate Product Ops Intelligence — confirmed at MAGG, not part of the approval middleware

Different buyer (ops/manufacturing director vs. commercial manager), different data sources (production systems, machinery logs vs. ERP/CRM), different ROI story. Keep separate from Track A.

Feature Status Confirmed at Priority Notes
Procurement Intelligence CONFIRMED MAGG — plant visit Mar 16 HIGH 17,000 raw material SKUs in Excel, no ordering rules. AI layer to flag reorder points, over-stock, waste.
Maintenance Tracking CONFIRMED MAGG — plant visit Mar 16 HIGH All maintenance knowledge in technicians' heads. Digitize manuals + preventive schedule + machine history.
SKU Performance Dashboard CONFIRMED MAGG — plant visit Mar 16 MED Currently a manual Google Sheets semáforo fed daily. Replace with live dashboard: velocity, backorder, zombie SKU detection, negative sales tracking.
Voice Processing POTENTIAL ~ No client yet TBD Mercura has this. No confirmed client use case yet — watch for voice-first approval or voice note intake in field operations.
Build Effort & Pricing Model
Workflow Build Time Impl. Fee (one-time) Annual Base What's Included (flat, unlimited use)
Inbound Request → Quote 5–6 weeks $10,000 $12,000/yr Unlimited requests · Odoo integration · email + WhatsApp channels · quote log in ERP
Spec / Product Matching 4–5 weeks $8,000 $8,000/yr Unlimited catalog queries · semantic search across full SKU catalog · Odoo catalog sync included
Discount / Price Approval 3–4 weeks $5,000 $8,000/yr Unlimited approvals · Verde/Ámbar/Rojo rules engine · decision log · WhatsApp/Teams/email delivery
Escalation Logic (add-on to Discount Approval — Naranja/Rojo escalation path) +1–2 weeks $2,000 +$4,000/yr Unlimited escalations · precedent log · decision brief generation
PO / Supplier Approval 3–4 weeks $5,000 $6,000/yr Unlimited PO approvals · budget check · approved supplier validation · ERP write-back
Credit / Terms Approval 2–3 weeks $4,000 $6,000/yr Unlimited credit requests · payment history query · aging report · ERP terms write-back
Returns / Credit Note 3–4 weeks $6,000 $6,000/yr Unlimited return requests · parallel quality + finance routing · credit note generation
AI Copilot 6–8 weeks $15,000 $10,000/yr Unlimited queries · per-seat license for all sales reps using the tool

Implementation is quoted as a bundled package when 2+ modules are built simultaneously — bundled pricing is ~30% below per-module sum. Billed in 3 milestones: 40% at signing · 40% at staging complete · 20% at go-live. Annual subscription starts Month 2 — first month free. All annual fees billed upfront, no volume counting, no overages.

Annual Subscription — Three Tiers
Starter
Up to 500 req/month
50–200 distributors
Minimum
$14K/yr
Ideal
$20K/yr
Growth
Up to 2,000 req/month
200–600 distributors · Lamosa here
Minimum
$24K/yr
Ideal
$32K/yr
ARONLIGHT
Scale
Up to 8,000 req/month
600+ distributors
Minimum
$40K/yr
Ideal
$55K/yr

Tiers apply per module. A client using Module 1 + Module 2 pays per-module annual fees at their tier. Lamosa (Discount/Price Approval, ~1,200 approvals/month) sits in Growth.

What Each Workflow Actually Does
Inbound Request → Quote MVP
Input: External customer emails or WhatsApps: "I need 300m² of a porcelain tile similar to your SKU X, but matte finish, for a commercial lobby."
ERP query: Catalog search for matching products, stock levels, client-specific pricing if on file.
Logic: AI matches request to closest SKUs, generates a draft quote with alternatives ranked by fit. Flags if none match well.
Output: Sales rep receives a ready-to-send quote draft. Reviews and sends — no ERP navigation required.
Write-back: Quote logged in ERP as a pending opportunity. Follow-up triggered if no response in X days.
Note: Mercura covers this — but requires SAP/NetSuite/Salesforce. No Odoo. No WhatsApp.
Spec / Product Matching MVP
Input: Customer or inside sales describes a product vaguely — typos, synonyms, dimensions in different units, cross-language terms ("azulejo porcelánico rectificado 60x60 mate").
ERP query: Semantic search across full catalog (50,000+ SKUs), ranked by relevance with spec comparison.
Logic: Returns top 3 matches with specs, stock, price, and a confidence score. Flags if best match is below threshold.
Output: Sales rep sees: "Best match: SKU 45143 (94% fit). Alternatives: SKU 45199 (matte, 81%), SKU 44820 (similar spec, 76%)."
Write-back: Selected product added to quote or order draft.
Note: Mercura covers this — but no Odoo integration. Aronlight's entire catalog is in Odoo.
Discount / Price Approval MVP
Input: Salesperson emails or WhatsApps: "Client X wants 18% on SKU 45143, 500m²."
ERP query: Margin at requested price, stock level, promotional grid for that SKU.
Logic: Verde → auto-approve if margin above floor and stock healthy. Ámbar → send to manager. Rojo → auto-reject with reason.
Output: Manager gets a WhatsApp/Teams message: "18% on SKU 45143. Margin: 22%. Stock: 800m². Suggested: approve. [Approve] [Reject]"
Write-back: Approved decision updates order price in ERP. Every decision logged with rule triggered and manager response.
Exception Pricing
Input: Salesperson flags a deal that breaks all standard rules: "Client Y wants $19/m² on 5,000m², list is $28. Annual contract worth $95K."
ERP query: Client's 12-month purchase history, last 3 comparable exceptions and their outcomes, current margin at requested price.
Logic: Always Rojo — no auto-approval. Instead of blocking, escalates with a structured decision brief.
Output: Director receives a 5-line memo: deal size, margin impact, client history, comparable precedents. One-click decision with full context.
Write-back: Approved exceptions logged as precedent, surfaced automatically in future similar requests.
PO / Supplier Approval
Input: Procurement sends email/WhatsApp with supplier quote: "Need to approve $12K in parts from Voltex for Project Norte."
ERP query: Remaining budget on that cost center, approved supplier list, requester's spending authority limit.
Logic: Verde → within budget + approved supplier → auto-approve. Ámbar → over budget by <15% → escalate. Rojo → unapproved supplier or over budget → route to finance director.
Output: Approver gets: "Project Norte has $18K remaining. This purchase = 67% of remainder. Supplier approved. [Approve] [Reject]"
Write-back: Approved PO created in ERP, supplier notified automatically.
Credit / Terms Approval
Input: Salesperson sends: "Client Z is asking for 60-day terms, currently on 30. They have $8K outstanding."
ERP query: Client's payment history (on-time %), outstanding balance, current credit utilization vs. limit, aging report.
Logic: Verde → history >95% on time, utilization <60% → auto-approve. Ámbar → good history but high balance → escalate to credit manager. Rojo → past-due history → escalate to GM.
Output: Credit manager receives 4-line client snapshot: payment track record, current exposure, request, recommendation.
Write-back: Approved terms updated in ERP customer record.
Returns / Credit Note
Input: Salesperson or client sends: "Client wants to return 120m² of lot 2024-089 — enamel defect."
ERP query: Original order, invoice, delivery record, return policy per product category, defect history on that lot.
Logic: Validates order exists and is within return window. Routes simultaneously to quality (defect check) and finance (credit note authorization). Auto-rejects if outside window.
Output: Two parallel notifications — quality manager and finance manager, each with their specific action.
Write-back: Credit note generated in ERP once both approvals received. Salesperson and client notified.
AI Copilot (Quote Building)
Input: Sales rep on a call with a customer — opens Copilot, describes what the client needs in plain language.
ERP query: Real-time product availability, pricing, recent orders for this client, comparable configurations used in past deals.
Logic: Surfaces product options, checks availability, flags conflicts (lead time, stock), suggests bundle pricing if applicable.
Output: Multi-line quote draft ready to export — no tab-switching, no ERP navigation during the call.
Write-back: Quote saved to ERP as draft, flagged for pricing approval if it includes discounts.
Note: Mercura has partial coverage — product research only, no ERP governance layer.
Client Reality Check

MVP workflows — per client

MVP Workflows — Client Reality Check
NEW · MAR 28

For each MVP workflow: why each client is rated HIGH or MOD, whether the workflow is already in a proposal we sent, and what action is needed.

MVP — BUILD FIRST
Workflow 1 — Discount / Price Approval

A salesperson asks their manager for permission to give a client a below-list price. The AI reads the request, queries the ERP for margin and stock, applies Verde / Ámbar / Rojo rules, and sends a one-click approval to the manager's channel. No ERP login required. Every decision logged.

Grupo Lamosa / San Lorenzo HIGH ✓

Manager (Doménico Casaretto) receives 30–40 emails/day from Jorge Zapata requesting discounts. Each requires opening Excel, finding the SKU, comparing list price vs. malla promocional, and validating margins manually. No traceability on rejections — they're lost in email threads. Confirmed in discovery call.

Proposal Status
✅ Sent — Fully Covers This
El Validador Inteligente is the entire proposal. All 3 phases map directly to this workflow.
Action: No proposal changes needed. Close 8 open questions with Diego → confirm ERP (Phase 3 depends on it) → finalize ROI numbers → start Phase 1.
Aronlight MOD ~

Discount approval governance almost certainly exists at Aronlight's scale (700 proposals/week, 1,000+ distributors), but it was not their stated pain. Their confirmed problems are proposal generation speed (Module 1) and technical product matching (Module 2). The approval governance layer is inferred — logical, but not validated in either discovery call or the v2 proposal. Don't lead with it.

Proposal Status
⚠️ v2 Sent — Different Focus
Proposal covers Module 1 (Sales Proposal Automation) + Module 2 (Spec Matching). Discount approval not in scope.
Action: Stay on Track A (proposal + spec matching). Once Module 1 is live and delivering value, surface approval governance as a natural follow-on. Do not introduce it now — it competes for attention with the pain they already said yes to.
Eurostar Machinery MOD

Complex capital equipment quotes require Ferdinando personally and take weeks — quote turnaround was named as a lost-deal reason. But this is low-frequency / high-value (one $500K machine deal, not 40 daily discount emails). The confirmed pains in the discovery call were after-sales automation, technician knowledge management, and admin visibility — not approval governance.

Proposal Status
⚪ Not in Proposal
Eurostar proposal covers after-sales, technician app, admin ops. Different engagement entirely.
Action: No changes to current proposal. If Eurostar closes on the after-sales module and a follow-up conversation surfaces quote approval as a pain, revisit then. Don't introduce it unprompted — it wasn't their stated problem.
MAGG / Electro Mag HIGH PROPOSAL: SENT

Plant visit done Mar 16. Ops pains confirmed in person: 17,000 procurement SKUs tracked in Excel with no rules, maintenance entirely in technicians' heads (no system), inventory stockouts as #1 client complaint, SKU performance tracked on manual Google Sheets daily. Production at 100% capacity. Commercial pains (distributor quote/approval) still to confirm with commercial team. Proposal sent by Raffaello to Omar (Apr 2026). NetSuite starts April — if Omar responds before kickoff, AI layer ships with the rollout.

Status
🟢 Proposal Sent
Sent by Raffaello to Omar (Apr 2026). Awaiting reply.
Status: Proposal sent. Entry: Track B (ops) — procurement rationalization + maintenance + SKU intelligence. Track A (distributor approvals) positioned as Phase 2. Next: Omar replies → if positive, kick off before NetSuite scope locks. If pricing pushback, align with Raffaello on retainer vs. module-based model before responding.
Phase 2 Add-on to Workflow 1 — same ERP connector, +1–2 weeks
Workflow 2 — Exception Pricing

A deal that breaks all standard rules — large volume, deep discount, or strategic account. Rather than auto-rejecting, the AI builds a decision brief for the director: deal size, margin at requested price, client's purchase history over 12 months, and the last 3 comparable exceptions that were approved or rejected. The director decides with full context. Approved exceptions are logged as precedent and surfaced in future similar requests.

Aronlight LOW

Strategic account exception pricing was not raised in either Aronlight discovery call or in the v2 proposal. At 700 proposals/week it's plausible that some exceptions exist, but this is speculative — there's no direct evidence it's a pain point. Their confirmed problems are speed (proposal generation) and accuracy (spec matching), not governance.

Proposal Status
⚪ Not in Scope
v2 proposal covers proposal generation + spec matching only. Exception pricing not discussed.
Action: No action. If Track A (Modules 1 + 2) succeeds and Aronlight expands into approval governance, Exception Pricing would be a natural Phase 3 add-on. Do not introduce it in the current engagement.
Grupo Lamosa / San Lorenzo MOD

The current proposal handles exceptions as Rojo auto-rejects — Phase 1 includes building an "exception SKU list" that defines which SKUs can never receive further discounts. That handles the clear-cut cases. What's not in scope: the strategic deal escalation flow for large accounts (e.g., a construction company asking for annual contract pricing well outside the grid). That would require a director-level decision brief, which isn't part of the current proposal.

Proposal Status
⚠️ Partially Covered
Exception SKUs → auto-reject only. Director escalation flow not in scope.
Action: Don't change the current proposal. After Phase 1 is live and exception SKUs are defined, ask Diego: "Are there deals that fall outside all rules but still get approved by Doménico personally?" If yes, scope Exception Pricing as a Phase 2 addition.
Eurostar Machinery MOD

Large accounts like L'Oreal almost certainly involve director-level pricing decisions on custom machine configurations. But this is 2–3 deals per year, not 40 per day. The ROI case for building dedicated exception pricing logic for Eurostar is weak at this stage.

Proposal Status
⚪ Not in Proposal
Different engagement focus. Not worth introducing now.
The Platform Has Two Parallel Tracks — One Per Design Partner
Track A — Aronlight (Most Likely First)
Proposal Generation + Spec Matching
Two discovery calls done. Pain questionnaire done. Proposal v2 sent. Waiting on Loom walkthroughs before development starts. Odoo integration confirmed as the technical layer.
Next action: Get Loom walkthroughs (Sales Director flow + Technical Director flow) → start Module 1 development.
Track B — Lamosa (Strong Second)
Discount / Price Approval (El Validador Inteligente)
Discovery call done. Proposal v1 sent. Architecture fully scoped (3 phases, Verde/Ámbar/Rojo). Waiting on 8 open questions from Diego before confirming ROI and Phase 3 ERP scope.
Next action: Close 8 open questions with Diego → update ROI numbers → start Phase 1.
The same underlying platform — ERP-connected AI middleware on top of email and WhatsApp — serves both tracks. Track A produces documents (proposals, spec matches). Track B produces decisions (approve/reject with rules and audit trail). Both share the same ERP connector architecture, the same channel integrations, and the same rules engine. Building for both in parallel means each client funds a reusable component the other client also needs.
Action Summary
Aronlight — proposal sent, architecture confirmed. Most likely first design partner. Two discovery calls done, v2 proposal in their hands. Unblock: get the two Loom walkthroughs and schedule the Sales Director + Technical Director calls. Development can start immediately after.
Lamosa — all 8 questions answered, proposal updated, ready for kick-off. ERP confirmed (SAP). 4-level approval flow defined: Verde (auto, UB ≥ 20% + min price per SKU or malla promocional) → Ámbar (Jefe Canal) → Naranja (Ger. Ventas/MKT) → Rojo (Ger. General). No exception SKUs. Cycle time: same day / max 24 hrs. Next: schedule kick-off with Antu and Andres — access to Jorge Zapata's last 200 emails + 2hr session to validate semáforo rules.
Eurostar — different engagement, don't conflate. After-sales, knowledge management, admin visibility. Neither MVP track applies. Proceed on the existing proposal scope.
🔵 MAGG — proposal sent by Raffaello to Omar (Apr 2026). Ops entry confirmed: procurement + maintenance + SKU intelligence. Track A (distributor approvals) as Phase 2. Waiting for reply. If Omar responds before NetSuite kickoff, AI layer ships with the rollout. If pricing pushback, align on retainer vs. module model before responding.

The commercial structure

MVP Scope — 4 weeks
  • Email listener that reads inbound requests and extracts: SKU/product, quantity, requested price/discount, requester name
  • ERP connector that queries pricing rules and stock levels (start with CSV/Excel import — add SAP/Odoo API in v2)
  • Rules engine: Verde/Ámbar/Rojo classification with configurable thresholds per client
  • Approval UI: one-line summary + single Approve/Reject button delivered via Teams, WhatsApp, or email
  • Decision log: every approval/rejection recorded with timestamp, requester, rule triggered, and manager decision
  • NOT in MVP: full ERP write-back, multi-language, mobile app, self-serve onboarding
Ideal First 3 Paying Customers
  • Aronlight (EU) — Track A — Most advanced: 2 discovery calls done, questionnaire complete, v2 proposal sent. Module 1 (Sales Proposal Automation, 700/week) + Module 2 (Technical Spec Matching). Blocking item: Loom walkthroughs from Sales Director + Technical Director → start dev immediately after.
  • Grupo Lamosa / San Lorenzo (Peru) — Track B — All 8 discovery questions answered (Mar 30). SAP confirmed. 4-level semáforo scoped (UB ≥ 20% threshold). Proposal updated. Next: kick-off with Antu and Andres → fastest path to a signed engagement in LatAm.
  • A second Lamosa subsidiary — Track B expansion — Once Peru is live, approach Argentina or Spain (Baldocer). Same SAP stack, same approval logic, 80% of the build reused. Land-and-expand within the Lamosa group is the fastest path to logo #3 at near-zero acquisition cost.
Aronlight — ROI Case
Their current cost
$300K
7–8 FTEs at €40K fully loaded to process 700 proposals/week manually
Annual savings
$245K
Eliminates 6 FTEs. Keeps 1–2 for review. AI handles drafting end-to-end.
Payback at ideal price
<5 mo
Year 1 net: $155K saved after $90K spend. Year 2+: $190K/yr net.
Our cost to operate / month
~$700
API $225 · infra $170 · dev support $100 · account mgmt $200
Our margin / month at ideal
$3,883 · 85%
At $55K/yr. Minimum ($40K/yr): $2,633/mo · 79% margin.
Commercial Terms
Implementation — Milestone Payments (total $35,000)
Week 1 — Signing
$14,000
Kickoff · requirements · Odoo connector started
Week 4 — Staging
$14,000
System live in staging · client testing begins
Week 6–7 — Go-live
$7,000
Production deploy · handoff · first month free

Subscription starts Month 2. First month on us while the team gets up to speed.

Year 2+ Loyalty
  • → 10% off annual renewal (~$4K–5.5K saved)
  • → Or: one add-on module at no extra impl. fee
  • → Second Lamosa subsidiary: 50% off implementation
Referral Credits
  • → $5,000 credit per referred client who signs
  • → Applied to next annual renewal — not cash
  • → Refer 2 clients → $10K off Year 2 (~18% discount)
  • → Aronlight has 1,000+ distributor contacts
Why Now
LLMs crossed the capability threshold for document understanding and structured data extraction in 2023. Prior to this, "email-to-ERP automation" required brittle regex rules and broke constantly. Now a single Claude API call can read an Excel attachment, extract 10 data fields, compare against a pricing grid, and return a structured decision in under 2 seconds. The timing is right because the capability is new — but the problem is 20 years old. The window to become the default solution in this LatAm/EMEA segment is open right now — Mercura just confirmed the category from the US/EU manufacturing side.
Next Actions

3 concrete steps

Action 1 — This week
Schedule Lamosa kick-off with Antu and Andres
All 8 questions answered (Mar 30). ERP: SAP. UB threshold: 20%. 4-level approval flow confirmed. Proposal updated and live. Next step: get access to Jorge Zapata's last 200 discount emails + schedule 2-hour session with Antu and Andres to validate semáforo rules with real cases. Demo at day 14.
Action 2 — Parallel track
Get Aronlight Loom recordings
Follow up on the two Loom requests (Sales Director flow + Technical Director flow). Aronlight's 200–300 requests/day is the scale proof-of-concept. Validating at that volume creates the high-volume case study for the eventual SaaS pitch — and is worth $30–50K in consulting revenue on its own.
Action 3 — After first contract
Re-run /venture-signal with revenue
Re-run once either Lamosa or Aronlight signs. Pattern confidence moves from "4 confirmed calls" to "1 active paid build" — that changes the productization narrative from hypothesis to traction, and makes the pitch to future clients significantly stronger.
Post-report options:
/red-team "AI Approval Middleware" — stress-test before committing   /market-scan "approval automation" "LatAm" — validate demand by geography   /evaluate-opportunity — full business case with financials