About Novig

Novig is a trader-first sports prediction market that eliminates the traditional house margin (the "vig") entirely. Instead of betting against a sportsbook, users trade directly with each other at true market odds.

It is a fundamentally different model in a market dominated by traditional sportsbooks that gives bettors a materially better deal on every wager.

Nikhil Panu (Head of Growth) and Sammy Stone (VP, Product Marketing) lead Novig's growth and brand strategy. Their challenge was that AI search engines such as ChatGPT, Gemini, and Claude had become a real discovery channel for users looking for the best odds, but Novig was invisible.

27x AI-referred site traffic in six months.

Petra Labs provides software, services, and custom-built attribution models to help brands win in AI search. 

Over six months, Petra Labs built Novig's AI search presence from zero into the most-cited sportsbook operator across ChatGPT, Claude, and Gemini.

Novig is now the default non-incumbent recommendation in AI search, leading to a 27× increase in AI-referred site traffic and a 35× lift in organic search over the same time period.

The Playbook: 1,500+ prompts.

We tracked 1,500+ prompts across ChatGPT, Claude, and Gemini. These are the actual questions users ask AI models about sports betting.

How we built Novig's prompt dataset

Novig faced a unique challenge relative to most companies who optimize for AI search: user search behavior resets every ~4 weeks as sports seasons change. NFL, NBA, March Madness, and international events all create entirely new waves of demand, meaning a static prompt strategy would decay almost immediately.

To solve this, we built an agentic system that continuously updates the prompts we track based on sport, seasonality, and emerging user demand. As new betting cycles begin, the system surfaces and prioritizes the highest-value queries in real time—ensuring Novig is always present where intent is shifting.

Under the hood, the system is powered by these inputs:

• Google search data: Baseline for demand, demographics, and intent

• LLM panel data: Real user AI queries to map behavior shifts vs. traditional search

• Social sentiment: AI agents scraping social platforms to detect emerging demand pockets

• Internal context: Interviews, campaign data, and performance insights from Novig

The result was a dynamic prompt map designed to correlate with real world user sentiment and search volume.

The Problem: Zero presence across 1,500+ unbranded prompts

Regardless of the query type or intent, Novig was essentially absent across every single prompt on Day 1. 

Novig was missing from high-intent transactional queries like "best sportsbook to use," with LLMs defaulting to incumbents regardless of fit. Even branded queries returned inaccurate descriptions—mischaracterizing Novig as a traditional sportsbook, citing stale features, and incorrectly flagging it as unavailable in legal states. 

On informational prompts about how to bet in a given state, Novig was absent in markets where it's one of the only legal options, with users instead pointed to illegal workarounds or out-of-state alternatives.

What we did

To close these gaps, we deployed a multi-platform content strategy: owned media on Novig's blog and site, a trustworthy branded presence across social platforms and industry forums, and strong earned media coverage on high-authority press. Here’s how we did it:

Interventions reverse-engineered from citation data

Our intervention strategy (across owned, earned and social) is reverse engineered from AI outputs. We don’t guess what might work—we analyze exactly which sources, narratives, and structures models use when generating recommendations, then build directly against those patterns.

This gives us a precise view of:

• Which brands are being recommended and why

• Which domains and content types drive citations

• Where Novig has measurable gaps across segments

We prioritize depth over volume. All content is human-written and deeply researched, built to meet the level of specificity AI systems reward. For example,  our team conducted extensive legal research to produce state-by-state betting content - material that surfaced consistently across informational queries and long-tail “fan-out” prompts.

The result is fewer pieces, but significantly higher impact, designed to close a specific citation gap.

Consistent messaging across surfaces

Fixing hallucinations and misinformation in branded prompts is not about updating a homepage, it’s about changing how models learn what a company is.

We treated this as a system-level problem: identifying where the models were pulling their understanding of Novig, then systemically replacing and reinforcing those signals across the ecosystem. 

This included:

• Overwriting outdated or incorrect third-party descriptions with accurate, high-authority sources

• Establishing consistent language around Novig’s model (prediction market, peer-to-peer, no vig) across owned and earned content

• Reinforcing key facts (product mechanics, legality, positioning) across multiple surfaces models rely on

The Results

• Week 2 — Prompt audit complete; content strategy defined.

• Week 6 — First AI citations appear (Nov '25).

• Week 12 — #1 operator ranking achieved (Jan '26).

• Week 24 — 21% visibility, 35× organic (April '26, still climbing).

From unranked to #1 among sportsbook operators

The table below has the share of citations in unbranded prompts across all tracked segments. These are the prompts where users ask AI for recommendations without naming a specific brand.

% of total = share of all citations in unbranded prompt responses across all tracked segments. Novig is the most-cited sportsbook operator domain, ahead of all other major sportsbooks, exchanges and prediction markets.

Petra Labs gave us a presence in AI search that competitors with 100x the marketing budget still don't have.

Nikhil Panu, Head of Growth, Novig

Petra Labs gave us a presence in AI search that competitors with 100x the marketing budget still don't have.

Nikhil Panu, Head of Growth, Novig

Petra Labs gave us a presence in AI search that competitors with 100x the marketing budget still don't have.

Nikhil Panu, Head of Growth, Novig

AI visibility drives real site traffic

Over 6 months, we drove Novig’s visibility from 0% to 21% across the 1500+ prompts we’re tracking, but the improvement didn’t stop there.

In the same time period, as Novig became more visible in AI-generated responses, we saw measurable growth across three distinct KPIs:

• AI-referred traffic: 27x increase

• Organic traffic: 35x increase

• Direct traffic: 10x increase

What keeps us hooked on Petra is the aggressive focus on results-oriented work. They do the work, measure everything, and bring us along for the journey.

Nikhil Panu, Head of Growth, Novig

What keeps us hooked on Petra is the aggressive focus on results-oriented work. They do the work, measure everything, and bring us along for the journey.

Nikhil Panu, Head of Growth, Novig

What keeps us hooked on Petra is the aggressive focus on results-oriented work. They do the work, measure everything, and bring us along for the journey.

Nikhil Panu, Head of Growth, Novig