

  All Discussions  |  

  Hello Post  |  

  Direct Messages  |   

  Categories  |  

  How-To(Forum)

Data Integration

DD

Dear all

We have a historic data model which provides and example to underpin a BIG scenario, and some generated sample data.

While it is always possible to debate fields, the key challenges as I’ve seen them are:

getting enterprise reference data into source systems so that the same elements can be referenced commonly across different systems
data quality and completeness
extracting for a common moment of relevance. While everyone lines the idea of “live” data - it also has to be good and to the same reference point (e.g. are actuals up to the same time point as forecast continues?)#
operating the machinery of integration - set up, capture, validation, transformation (of needed) - speed and flexibility of periodic recapture

…we all probably have views on this.

Do we ned a data integration partner in our group, or should we stay focused on the logic - and get integration partners for implementation?

It strikes me that to use prioritisation solutions, or visualise using what Eli has, or put together a shared information platform - data integration is vital….

@TonyS @ChrisB @EliO

TonyS

@DD

David - I think you have nailed the real issue.
Everything you have listed is exactly why most initiatives stall:
• reference data fragmentation
• inconsistent quality
• misaligned time horizons
• heavy integration overhead
That is the machinery problem - and it is real.
My view is we should not try to solve that inside this group.
There are already strong integration players who do:
• ingestion
• alignment
• transformation
• orchestration
We risk diluting the proposition if we go there.
Where we stay focused is the logic layer above that:
• how data becomes usable structure
• how obligations, constraints, and dependencies are derived
• how that turns into decision insight (BIG / x42 / Integreli)
Because even with perfect integration today:
organisations still cannot answer
“what depends on what, and what breaks if this changes?”
That is the gap we are filling.
So, I would frame it as:
• Integration partners > make data usable
• Our layer > make the business understandable
On your last point - fully agree:
Yes, integration is vital.
But it is an enabler, not the differentiator.

If anything, the model becomes:
• plug into existing data platforms where they exist
• use partners where they do not
• keep our focus on the reasoning and decision layer
Feels like a cleaner separation - and easier to scale.

DD

I agree with you Tony - data ingestion is a material issue.

It is an enabler, not the substance of BIG

I would like to find partners who can support delivery in this domain.

Laters,
D

TonyS

agree @DD in my opinion either the business have an agreed common data model and have delivered or on the process of delivering it. If they have not than at least BIG is a starter for ten albeit only for BIG outcomes and not necessarily sorting out operational data exchange between applications. But many businesses do not have this - even billion pound companies I have engaged with do not have this - and these are now the companies deploying AI thinking it will solve all their problems. It will actually make their problems bigger and quicker resulting in the business saying this AI is rubbish - back to the crap-in crap-out with everyone going back to using Excel



  All Discussions  |  

  Hello Post  |  

  Direct Messages  |   

  Categories  |  

  How-To(Forum)