14 Comments
User's avatar
Mark Palmer's avatar

Congratulations for launching the ECM Substack! Loving the LiDAR analogy, the topic, and the "context" about how to do better AI via rules, entities, and better semantics fed to the model. Good stuff!

Fergus Keenan's avatar

Thanks Mark - really appreciate it. When we stepped back and looked at all of the pieces of the puzzle we were working on, they started to click together - context matters!

Mark Palmer's avatar

Looking forward to more. I’m especially interested in the difference between context engineering, prompt engineering, semantic layers + ECM, use cases, and more about Context Catalogs. Looking forward to it, this is an important area

Mark Palmer's avatar

Adding this link to a related post by AI By Hand about context engineering:

https://open.substack.com/pub/aibyhand/p/context-engineering-by-hand

Good to see the conversation building about this important area!

Heeje's avatar

Clearly, implementing good AI is now a requirement to compete. Thanks for the information!

Shawn fitzgerald's avatar

On point and exactly what companies need to integrate the new technology

Robert Merlicek's avatar

Brilliant. Organisations desperately need context management that over arches the enterprise systems that use AI for automated decision making. It’s like business AI metadata. As we move to multi-modal AI and multi system agent solutions, this will be a necessity

Mark Palmer's avatar

“Business AI Metadata” ==> Yes! Good analogy! There are also parallels with semantic layers and the data prep / wrangling aspects of BI tools from yesteryear :) And Data Catalogs, which are more designed for BI / Human consumption than ECM. It’s cool to see a new category of software emerge.

Fergus Keenan's avatar

Exactly, Robert. And it’s not just multi-modal or multi-agent systems that benefit - our approach gives large organizations control over their business context instead of handing it to the frontier labs. Much like how enterprises adopted multi-cloud strategies to avoid lock-in, we’re helping them achieve the same sovereignty over their AI context, without the pain that came with that first transition.

Ali's avatar

Great article, is there any documentation related to other use cases and technical implementation of ECM?

Fergus Keenan's avatar

Thanks Ali! Watch this space - we are preparing a full technical overview and will be publishing in the coming weeks. Thanks a lot for the interest!

Ali's avatar

Awesome, looking forward to that.

Eric's avatar

The LiDAR analogy really makes this click. A lot of AI projects struggle not because of the models, but because the underlying data isn’t structured or contextualized properly. It also feels like document workflows and access control play a bigger role than people expect here. If the content layer isn’t well managed, the context breaks down quickly. Some tools like Dokmee (dokmee.com) are starting to focus more on that side, which seems aligned with what you’re describing.

Michel's avatar

Interesting take on enterprise context management. A lot of organizations struggle with context simply because their information is scattered across emails, shared drives, and legacy systems.

That’s why document and content management platforms are becoming important infrastructure. Tools like Dokmee help centralize documents and make them searchable, which makes it easier for AI systems to actually access meaningful enterprise context.