← All work

Behavioural AI · Founder & CTPO — Acquired 2025

NOW

Real-time behavioural analytics and an agentic audience engine — applying machine learning to live behavioural data across billions of daily intent signals.

Role

Founder & CTPO (Present.)

Outcome

Sold to a strategic acquirer, 2025

Signal scale

60B daily signals · 320M profiles

Firm

Present. — $5M ARR at peak

NOW intent graph Agentic Segmentation · Intent Signals · Real-time Now

The proprietary analytics platform behind Present., the data-and-analytics brand-strategy consultancy I founded and grew to $5M ARR — sold together, to a strategic acquirer, in 2025.

Context

Present. ran seven concurrent clients at peak — among them Pantone, Genesis Motors, Four Seasons, Cartier, Sequoia Capital and MyMochi (which I partnered with through its growth from $50M to $1B). NOW was the engine underneath: I designed and built it, applied machine learning to live behavioural data, and embedded it into client ecosystems.

What I built

  • A collaborative multi-agent framework that turns an ideal-customer profile into a hyper-targeted audience from a 24k-topic taxonomy.
  • Interactive D3 topic graphs — force-directed, radial and cluster layouts over the full taxonomy, with live search across 24k topics.
  • A multi-step audience builder with boolean (AND/OR) filters and signal-frequency thresholds, plus a demographics and reporting suite.
  • A service layer for 50+ activation integrations across ad networks, CRMs, CDPs and analytics.
  • Built Pantone's entire post-acquisition sales platform as part of the same client work.

How it's built

A React 18 / TypeScript front end on Vite, with Redux Toolkit and Redux-Saga carrying the complex async flows of audience creation and integration sync. A Node/Express backend streams live signal updates over Socket.io, and the agent layer (coordinator, research and topic agents) is wired to the OpenAI SDK. Around a hundred TypeScript components, dark-mode throughout, and a data-visualisation layer built directly on D3 rather than a charting wrapper — because the topic graph is the product.

Outcome

A founder exit: the consultancy and its NOW platform were acquired in 2025. Documented client results ran to 90% reductions in customer-acquisition cost and 34% lifts in conversion.

React 18TypeScriptViteRedux-SagaNode / ExpressSocket.ioD3OpenAIMulti-agent

Core feature

Rebuilt in NOW's own UI

An interactive force-directed graph that maps behavioural-intent topics into category, subcategory and topic nodes — click any node for its growth, sentiment, demographics and related interests.

Topic Node Map — intent graph

Intent Graph

Topic Relationships

Large Language ModelsSentiment AnalysisText-to-ImageChat AssistantsObject DetectionFacial RecognitionTime SeriesChurn PredictionData LakesInteractive DashboardsNatural Language ProcessingComputer VisionGenerative AIPredictive AnalyticsBig DataData VisualizationAI Content CreationAI & Machine LearningData Science & Analytics

AI Content Creation

▲ Trending

Generative tooling that drafts marketing copy, imagery and video from prompts — the fastest-accelerating intent cluster inside Generative AI this quarter.

TOPIC
Growth+182%
Monthly Searches1,284,500
SentimentPositive
Connections7
Age Groups
25–3444%
18–2431%
35–4418%
Top Locations
US58%
Europe27%
Asia11%
Related Interests
technologymarketingbusinessentertainment
Node TypesCategorySubcategoryTopic
IndicatorsGrowthHigh Growth
Drag nodes to rearrange • Click to view details • Search to find specific topics