AWS and HighByte Maturity Model: Is Your Factory Data Ready for AI?
AWS and HighByte have published the white paper "A Maturity Model for Industrial Data Management", which proposes three maturity stages and a self-assessment questionnaire. A useful reminder for anyone investing in AI: data comes before models.
According to IIoT World on 24 September 2026, AWS and HighByte have published a 25-page white paper titled "A Maturity Model for Industrial Data Management". The document proposes a model for assessing whether production data is truly ready to support advanced analytics, artificial intelligence and autonomous systems.
A note on the source: the IIoT World article is labelled "Sponsored by HighByte | Editorially Independent", so it should be read as vendor-driven content. It is also a short summary with no results data, and this commentary is not based on a reading of the full white paper.
Who wrote it and the premise it starts from
The authors are Ashtad Engineer, Worldwide Head of Manufacturing Solutions at AWS, and John Harrington, Chief Product Officer at HighByte. The document also covers the AWS Industrial Data Fabric architecture and how HighByte Intelligence Hub handles data integration from edge to cloud.
The premise will be familiar to anyone working on the plant floor: many manufacturers investing in AI, advanced analytics and autonomous systems discover that their underlying data cannot support them.
The three stages of the model
The model is organized into three maturity levels:
- Foundation: basic data collection and governance;
- Intelligence: analysis and optimization;
- Transformation: autonomous operations and AI integration.
For each stage, the document outlines use cases for both discrete and process manufacturing, technology stack requirements, best practices and a generative AI readiness assessment.
How the assessment works
The assessment has four steps and more than 50 questions. It starts at Stage 1, checking whether basic data collection and governance are in place; if so, you move on to Stage 2 and then Stage 3. The outcome is a low, medium or high maturity level, indicating where to invest next.
Questions are tailored by role:
- Production: how metrics are monitored and bottlenecks identified;
- Maintenance: equipment monitoring and failure prediction;
- Quality: tracing defects back to process parameters;
- IT/OT: data architecture and integration strategy.
Why it matters for plant leaders
Its main value is practical: a checklist to work through before buying AI tools. Each function can assess itself independently and then compare results with the others, bringing out differences in perception between production, maintenance, quality and IT/OT.
Before AI comes reliable connectivity, data collection and context at the MES level.
The model reinforces this thesis: without reliably connected machines, continuously collected data and contextualized information (order, batch, shift, process parameters), analytics and AI are unlikely to deliver usable results.
Using it with a critical eye
Since this is a framework that promotes its sponsors' products, it makes sense to adopt its structure but not necessarily its technology recommendations. A reasonable approach:
- run the self-assessment by role, without anticipating the desired answer;
- compare results across functions and identify shared gaps;
- invest first in the lowest level found lacking;
- evaluate technical solutions against your own architecture, not that of the model's vendor.
In short, the white paper says nothing radically new, but it offers a structured method for a concrete need: understanding, before spending, whether your factory data is up to your AI ambitions.