Fractional AI Product Lead Service
Make your product, marketing, and GTM teams AI native.
I help startups build a closed-loop growth engine, from experiments to data to insights, and ship the AI workflows that make it run 10x faster.
Book an intro call
Previously / Currently: Zynga | Glu Mobile (acquired by EA) | Supernatural (acquired by Meta) | 1 patent | 2 exits
The Engine: Four Hubs, One Learning Loop
A growth engine isn't a dashboard. It's the system your team uses to form hypotheses, run experiments, and convert raw events into compounding product knowledge — instrumented, automated, and AI-augmented end to end.
01 — Analytics Hub
PostHog + warehouse instrumentation. DAU, retention, funnels, lifecycle, LLM cost and quality. Stripe joined to product events.
Stack: PostHog, BigQuery, Stripe, dbt. Learn more about Product Analytics
02 — Experimentation Hub
The learning engine of the company. A/B and prompt experiments, holdouts, sequential analysis, automated readouts.
Learn more about experimentation for AI startups
03 — Internal Apps Hub
AI copilots for PMs, support, and analysts. Cursor/Claude/Codex workflows wired into your data and your ticketing.
04 — GTM Hub
B2B go-to-market: signal-driven outbound, account scoring, ICP refinement, marketing-to-sales handoff wired to your data.
Learn more about B2B GTM
The Growth Engine: Two Loops, One Engine
The Knowledge loop converts raw events into insight. The Experimentation loop converts insight into the next hypothesis — and the next hypothesis produces new data that re-enters the engine. Insight is the bridge. The whole thing compounds.
Knowledge Loop
Data leads to Information, which builds Knowledge, which sharpens into Insight that changes a decision.
Experimentation Loop
Insight becomes a sharp hypothesis. The hypothesis becomes a designed experiment. The test produces new data — and the engine spins again.
Engagement: How We Work Together
Embedded as a fractional AI product lead service — 12 weeks, 10 hours a week. We instrument, we experiment, and your team owns the engine — plus a delivered library of agents, skills, and apps — when I leave.
Week 1-2: Audit
Map the existing stack, events, dashboards, decisions. Find the 3 biggest leaks in the loop.
Week 3-6: Instrument
Stand up Analytics + Experimentation hubs. Define activation, retention, and the north-star event.
Week 7-12: Run and Build
4-6 experiments in flight. Weekly readouts. Ship agents, skills, and internal apps wired into PM, marketing, and support.
Week 12+: Hand Off
Your team owns the engine and the delivered library of agents, skills, and apps. Optional monthly review.
What You Get: Agents, Skills, and Apps
Agents: Production agents wired to your data, your tickets, your warehouse. PM copilot, support triage, analytics agent, growth-experiment agent.
Skills: Reusable, prompted workflows your team invokes from Cursor, Claude, or Codex. Experiment readouts, funnel analysis, creative variants, ad copy generation.
Apps: Internal apps for the workflows that don't fit a chat. Experiment tracker, insight library, AI cost monitor.
About Anshul Dhawan
15+ years in growth and product. Led growth and analytics at Zynga. Senior Director of Product Management and BI at Glu Mobile (acquired by EA). VP Product and Analytics at Supernatural VR (acquired by Meta). One US patent on experimentation methodology. 100M+ players analyzed.
Read full bio | LinkedIn | anshuldhawan.com
Tell me where you are — pre-instrumentation, mid-build, or stuck on the loop. I'll come back within 24 hours with a take.
Email: anshuldhawan@gmail.com
Based in San Francisco. Available worldwide.
Frequently Asked Questions
What is a fractional AI product lead?
A fractional AI product lead embeds into your startup part-time (typically 10 hours per week for 12 weeks) to build your growth engine — analytics instrumentation, experimentation platform, AI agents, and GTM workflows — then hands off ownership to your team.
What does the AI Powered Growth Engine include?
Four hubs: an Analytics Hub (PostHog + warehouse instrumentation), an Experimentation Hub (A/B and prompt experiments with automated readouts), an Internal Apps Hub (AI copilots for PMs, support, and analysts), and a GTM Hub (signal-driven outbound and pipeline management). Plus a delivered library of agents, skills, and apps your team keeps.
How long is a typical engagement?
The core engagement is 12 weeks at 10 hours per week. It starts with a 2-week audit, followed by instrumentation and build, then handoff. Optional monthly reviews are available after the engagement ends.
Who is Anshul Dhawan?
Anshul Dhawan has 15+ years in growth and product, with leadership roles at Zynga (acquired by Take-Two for $12.7B), Glu Mobile (acquired by EA for $2.4B), and Supernatural VR (acquired by Meta for ~$400M). He holds a US patent on experimentation methodology and has worked on products reaching 130M+ MAUs.