Yagrumo · a short film
The harvest in your data
What a Revenue X-Ray finds, and who it's for.
Data
The market today
AI assessments all look the same.
AI Opportunity Assessmentinterviews → roadmap · cost savings
AI Readiness Auditworkshops → ranked ideas · cost savings
AI ROI Assessmentprocess map → roadmap · time saved
AI Workflow Auditinterviews → deck · automation
AI Opportunity Auditworkshops → roadmap · productivity
AI Strategy Sprintinterviews → matrix · efficiency
Indistinguishable
Fixed-fee offers reviewed, 2026: Acropolium, Digital Scientists, GetWidget, Hubspire, Upwork listings
Data
The island's twist
0%
of companies in Puerto Rico already use AI
Main blocker to AI adoption
Security · the top blocker in Latin America
Experience · also a regional blocker
Data structure · the top blocker in Puerto Rico
IDC Latin America (Oscar García), survey from late 2025, reported by El Vocero
Conclusion
How it works
Don't ask for trust. Show the money.
customers.csvIDs only, no names
transactions.csv24+ months
products.csvcatalog + margin
Revenue X-Ray
$ at stake
one figure for the revenue already sitting in their own data
+ 50 named accounts to act on Monday
From the Revenue X-Ray research doc, Oct 2026
Conclusion
What the X-Ray reads
KeepWho is quietly leaving?At-risk accounts ranked by revenue at stake
GrowWhat do similar customers buy?Cross-sell and reorder gaps per account
PriceWhere do discounts leak?Margin leakage by rep, customer, product
CreateWhat new offer is hiding?Bundles, service tiers, unserved segments
Revenue levers, not cost levers: the brief asked for better products, more customers, better experience, new offerings
Conclusion
Who fits
The filter isn't an industry.
Repeat customers we can identify
More than one product line
Two or more years of transaction history
An owner who feels growth pressure
Same three-file schema for every match, which is what makes it a product
DataConclusion
Segment 1 · B2B distributors
+$0M
in 12 months from AI recommendations at a $400M distributor
+0 bps
gross margin from AI pricing at a distributor (McKinsey)
0+
establishments served by Ballester Hermanos0+
customers served by Méndez & Co.Distribution is among the business functions least likely to use AI today. Pharma and medical distribution excluded.
Distribution Strategy Group (proton.ai case); McKinsey, Pricing in distribution; Unisco; Méndez & Co.; US Census BTOS AI supplement, 2026
Data
Segment 2 · Cooperativas de ahorro y crédito
$0B
in assets across 0 credit unions, Q1 2026
$0M
in certificates opened by members a next-best-product model picked (160 CDs, one US credit union)Regulated by COSSEC. Slower to buy, but every member's full product history sits in one place.
NASCUS, Q1 2026; Estudios Técnicos via News is my Business; Trellance + Dataiku (MSUFCU case)
Data
Also a fit
03Specialty practices35% of recall texts booked within a month
04Dealer service departmentsOut-of-warranty retention 8.5% → 50%+
05Boutique hotelsRecord $1.99B lodging revenue in 2025
06Consumer retailPersonalization lifts revenue 10–15%
Curogram; Kenect (Austin Subaru); Discover Puerto Rico, 2026; McKinsey via Starr Conspiracy benchmarks
DataConclusion
The biggest sector, left out
Manufacturing is the economy. Not the buyer.
Manufacturing 44.2% of Puerto Rico GDP, FY2024100%
Revenue decisions sit at US headquarters, not in the plant
DDEC, Puerto Rico Manufacturing Profile 2025. Keep manufacturers for an operations offer.
Conclusion
The product is a ladder
FREELeak Calculator5 numbers → a rough range
FREELive demoa fictional island distributor
ENTRYRevenue X-Ray3 files → $ + named accounts
PROOFRevenue Sprint90 days vs. a holdout group
RECURRINGRevenue Enginerefreshed monthly
Trust is built before the data ask: show the output on fake data first
Conclusion
How we work with your data
Customer IDs only. No names, no contact detailsAnonymized
You see the output on demo data before you share anythingTrust first
Lift is measured against a holdout group, not guessedProof
You pay for proof before any buildNo lock-in
Every X-Ray ends with a dollar figure, the accounts behind it, and the evidence for both.
Know what to build.
Then build it.
Yagrumo · Data & AI