Hi guys.
Are you implementing AI based solutions for Network Performance and Optimization.
Is this a reality?
Do you know the main solutions / vendors?
1 Like
AIOps is what is called now.
Hi there,
I have worked with a vendor who purchased a performance tool that uses AI to detect network anomalies.
It makes correlations between KPIs to (predict) if a certain KPI will be impacted for a certain (cell/site/subscriber).
It’s not yet mature enough. Such tools require tremendous data sources, experts to handle models training, and validations.
The tool I mentioned called (Agility) a product of B-yond.
It’s a good tool as of now, and it has good detection capabilities.
So it’s mainly for performance and analysis.
For AI based (optimization) usually most of them are classified tools (for vendors) they spent millions on them, to give you reliable results, Genex Cloud/WinSpace are examples of these, which are Huawei products. You might find some companies started making such tools for public.
All in all, AI is the new beast, which will come at last to solve many problems faster, not all of them, of course.
Viavi AIOPS, Cellwize, Spectrum effect… some example of companies that use this.
1 Like
Thanks guys.
I was thinking that it would be hard and take time to have a vendor agnostic solution.
Since, to have a strong model you require a huge amount of data from different sources (Operators) which I think is not easy to develop.
Data is gold. Operators will not provide so easy.
That’s correct, I think companies are making them as (case by case) tools.
From your data, they train models to be your future tools.
Huawei will be used only for the Huawei network, Ericsson the same etc. At least that’s normal.
Thanks for the feedback. I will check more about the solutions.
There is little value in using AI to improve network performance.
The most important step for telco is to combine radio, switch, transmission, customer experience, sales, marketing, finance data into one tool,
As long as data remains in silos… with RF team unable to get marketing, sales and finance data, and switch team refusing to share data with radio etc…
I have been working in data science industry for the last six years.
I can tell you. AI is not sexy.
The sexy thing is to combine data.
Combine data from all sources into one tool.
Telcos are fighting a war of survival.
It’s now or never.
If they do not consolidate data, and do not use data to increase sales and cut costs, they will all die a collective death.
There is no winner in this war. Only losers.
Telcos must change their style. Their ego is gonna kill them.
There is no point improving KPIs of a given customer if he only pays 2$ a month.
On the other hand, the customer which pays 2000$ a month, he simply churns without even complaining.
It’s only when we combine sales, marketing, finance, complaints, DT, KPIs from RF/switch/trans etc into one tool that we can identify what to optimize and what to ignore.
2 Likes
In US ATT charges me 100 a month😅