Build vs. Buy: Optimise AI by Fixing the Data Foundation

Summary

This article argues that the biggest barrier to useful AI is not models but the quality of customer data feeding them. Poor, fragmented or duplicate customer records undermine AI accuracy, personalisation and predictive insights. The piece recommends moving beyond a binary “build vs buy” decision towards a hybrid, composable strategy: buy proven components for speed and stability, build differentiating capabilities, and prioritise identity resolution and governance as the bedrock of AI success.

Key Points

AI effectiveness depends on high-quality, unified customer data — garbage in, garbage out.
Identity resolution (creating one trusted customer profile) is foundational; without it, segmentation, targeting and recommendations fail.
Building a custom platform gives control but is costly and time-consuming; buying offers speed and tested features but can lack flexibility.
A hybrid, composable approach is recommended: choose best-of-breed tools for specific functions (identity, unification, governance, activation) and integrate them into a flexible stack.
Best practices: prioritise identity first, buy for speed and build for differentiation, treat data infrastructure as modular, and regard data as a key business asset, not just an IT problem.

Context and relevance

As enterprises race to deploy generative AI and intelligent agents, many are discovering that strategic benefit hinges on a reliable data foundation. Industries like retail and travel are already seeing limits to model performance due to fragmented customer records. The article is relevant to C-suite leaders, data architects and programme sponsors who must balance time-to-value against long-term differentiation and compliance needs. It ties into ongoing trends around composable architectures, data governance, and identity-first approaches that enable scalable, trustworthy AI.

Author style

Punchy: the author is direct and pragmatic — urging leaders to stop treating build vs buy as an either/or and to focus on practical steps that unlock AI value. The tone stresses urgency for organisations serious about AI outcomes.

Why should I read this?

Short version — if you’re planning or running AI projects, skip the fluff: this tells you where the real work is (your data). It gives a clear, no-nonsense roadmap: sort identity, use proven tools for the basics, build what sets you apart, and keep your stack modular so you can adapt fast. Saves you time and costly mistakes.

Source

Source: https://ceoworld.biz/2025/10/20/build-vs-buy-optimize-ai-by-fixing-the-data-foundation/