MarketPulse
All chapters

02 — Assumptions & Data Sources (the audit trail)

Read this first if you are evaluating the numbers. Ai Palette is a private company. There are no SEC filings, no 10-Ks, no audited financials, no published revenue. Every financial figure in this strategy is a labeled estimate built with one of three methods and anchored to a public benchmark. The judging criterion is defensibility of logic, not access to non-existent internal data.

The three estimation methods we use

MethodWhen we use itExample
Top-down (TAM→SAM→SOM)Sizing market opportunityAI-in-F&B market → CPG insights slice → reachable enterprise accounts
Bottoms-upSizing revenue / ARR# target accounts × win-rate × ACV per tier
Analog benchmarkingPricing, NRR, CAC, motion thresholds"comparable B2B SaaS median per-seat is ~$45/mo, enterprise insights ACV $25K–$120K"

Every estimate below is tagged with its method [TD] / [BU] / [AB] and a confidence level.


A. Facts about Ai Palette (sourced, not estimated)

ItemValueSource
HQ / foundedSingapore, 2018Tracxn, CB Insights
Total funding~$13.46MTracxn
Languages / countries covered18 languages / 24 countriesAgFunderNews; aipalette.com
Core productsForesight Engine, Concept Genie, Screen Winner, FoodGPT, Brand SAYaipalette.com
Named customersNestlé, Danone, Kellogg's, Cargill, OlamAgFunderNews
Main competitorsTastewise, Black Swan Data, Spate, Mintel, Datassential, Native AI, TechnomicTastewise blog, CB Insights

B. Market sizing inputs (public reports → our TAM/SAM/SOM)

InputValueSource
AI in Food & Beverages market (2025)~$16.36B, ~39% CAGRPrecedence/GlobeNewswire
Generative AI in CPG~9.5% CAGR segment, F&B ~36% sharemarket.us
North America share of AI-in-F&B~32% (2023)grandviewresearch
Value at stake for a $10B CPG from AI digital transformation$810M–$1.6BMcKinsey, State of Food & Beverage

Our derived sizing [TD], medium confidenceillustrative: We treat consumer-insights

  • NPD software as a thin slice (1–2%) of the broad "AI in F&B" number, because most of that market is manufacturing/quality/supply-chain AI, not insights SaaS. Resulting serviceable software market for trend-intelligence/NPD tooling is estimated at **$1.5–3B today**, growing double-digit. SOM is bounded by reachable mid-to-large CPG accounts per region (see doc 07).

C. SaaS commercial benchmarks (anchors for pricing & GTM) [AB]

Benchmark2025 valueUse in this strategySource
Per-seat as primary model57% (down from 64% in '24); median ~$45/seat/moWhy we don't lead with pure per-seatMonetizely / SaaS CFO
Usage-based adoption61% of B2B SaaS use some consumption pricingJustifies credits/usage layerMaxio
Hybrid pricing growthHighest median growth (~21%)Why we recommend hybridMonetizely
NRRmedian ~101%; top performers ≥111%Retention targets per tierBenchmarkit / Growth Unhinged
CAC paybackmedian ~20 mo; "good" ≤12 moMotion choice & efficiency guardrailFirst Page Sage
PLG vs SLG thresholdPLG works <$10K ACV; SLG for >$25K ACV + committee buyingMotion-by-tier mappingThoughtlytics / ProductLed

D. Assumptions we are making explicitly (challenge these!)

#AssumptionMethodConfidenceRationale
1Ai Palette ACV today sits ~$30K–$120K (enterprise, multi-market deals)ABMedEnterprise insights SaaS with named global CPGs; multi-seat, multi-market
2~80% of the codebase can stay global; ~20% is the localizable edgeAB/expertMedTypical for data-platform SaaS where the ML core is shared and only data sources/NLP/taxonomy/compliance differ
3Willingness-to-pay: US > W.Europe > Japan/Korea > SEA > IndiaABMed-HighStandard enterprise-software WTP ordering by GDP/IT-budget density
4A new "Essentials" self-serve tier can reach <$10K ACV and be PLG-assistedABMedAligns to PLG ACV threshold; widens funnel in price-sensitive markets
5Localization cost is dominated by data acquisition + taxonomy, not UI translationexpertMedTrend data depth is the product's value driver
6Target blended NRR 110%+ via usage expansion + multi-market land-and-expandABMedAchievable for enterprise data SaaS with consumption upsell

E. What we deliberately did NOT do

  • We did not fabricate Ai Palette revenue, headcount-by-region, or churn. Where those would normally come from internal data, we used ranges + benchmarks and flagged them.
  • We did not rely on any single market report as ground truth; sizing is order-of-magnitude.

Sources