What most determines the quality of outcomes from applied AI initiatives: the breadth and quality of data used, or the underlying model?

Tomer Cohen · Dec 20, 2023

For any product leader, designer or engineer, understanding data collection and data quality is the second most important part of the job, after defining the objective of the algorithm

If AI is at the center, data is the oxygen that brings it to life — the algorithm eats what you feed it — yet people gravitate to the UI and individual elements and assume someone else will collect the data

27:04 20VC Roundtable: Spotify, Adobe & Linkedin CPOs on How AI Changes The Future of Product, Why AI is Now the Product, How TikTok Changed Product, Why Cost is the Biggest Barrier to LLM Usage & Why Incumbents Can Adopt AI Faster Than Any Prior Innovation Cyc

Aidan Gomez · Aug 19, 2024

He dramatically underrated the importance of data quality — he thought it was just scale, and models are surreally sensitive to their data, with even a single bad example among billions mattering

Proof points internally at Cohere transformed his understanding of what matters in building the technology

Scope: change of mind over the last twelve months

49:56 20VC: Chips, Models or Applications; Where is the Value in AI | Is Compute the Answer to All Model Performance Questions | Why Open AI Shelved AGI & Is There Any Value in Models with OpenAI Price Dumping with Aidan, Gomez, Co-Founder @ Cohere

Kieran Flanagan · Jul 11, 2025

The quality and breadth of your data is the dominant driver of AI results: combining more internal and external data sources reliably improves any AI initiative

Every leap in their chat sales agent, support deflection rate and email personalization came from adding or refining data sources, e.g. loading developer docs into the support agent

Scope: described as mundane but consistently true

17:01 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025

Kieran Flanagan · Jul 11, 2025

The data layer is the main thing that determines AI results, and most companies lack clean internal and external data; the differentiator is being creative about external data sources rather than relying on traditional B2B data

Their micro-audience work only produced novel insight because they combined internal product-champion lists with external job-ad data and a KPI seed list to surface rising KPIs they would never have marketed to otherwise

59:52 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025