Applying best practices in AI/ML to grow in B2B Telecoms

Network X 23 Colt presentation Mizu

On the second day of this year’s Network X there was one presentation which really stood out to me. It was a presentation by Mizu from Colt on their active use of AL/ML to support their sales and marketing efforts to ultimately reduce costs and increase revenue. With B2B being a focus of IFG Consulting Europe and our work in helping companies to grow in telecoms, this area is very much in line with our interests and those of the Operators we are working with.

There’s a lot of hype surrounding AI and how it can solve problems. There are now start-ups that have raised tens of millions just to train their AI, and only then will they become operational. Colt is ahead of the game and is already using a mixture of business analytics, machine learning and generative AI to gain an advantage by gaining a better understanding of the upsides, business risks and intent of their existing customers.

Colt leverages its own unique data set to gain relevant information regarding its customers and then can match this up with external public data sources and private data sources to continue to develop the relevant impulses and what I like to refer to as “compelling events” related to their customers.

During the early part of the pandemic, Colt was able to use AI to influence and help identify new revenue opportunities by identifying a list of companies which had a high intent to purchase more bandwidth from Colt (Mainly Hyperscalers, SaaS companies, large MNC enterprises and media companies etc.). They ended up selling sixty 100Gbps connectivity services across the world during this period.

Customer retention was another positive use case during the pandemic, where Colt was able to identify customers who might have systemic business challenges due to COVID, causing financial stress on their customers who were under normal circumstances viable businesses. They were able to provide relief by proactively offering flexible payment terms. In my view, this was a brilliant strategy not only in helping viable businesses to survive but also generated goodwill and loyalty with the Net Promoter score increasing from a relatively mediocre average of 54 to 74 points.

Furthermore, this forensic approach to understanding their customer is instrumental in identifying market trends and new revenue opportunities and takes the labour-intensive, hard work out of correlating data and events away from the sales team. This thereby allows the sales team to focus on engaging with customers in a more targeted way, reducing the cost of sales, generating more revenue and profit and improving the general morale of the staff.

Given that every sales team has by definition limited bandwidth and resources Colt was able to segment and classify their customer base, both in relation to revenues but also predicted potential and growth. Those customers who had the least potential and growth would be served through digital channels and automation, the next tier of customers would be supported through virtual teams and the most valuable and promising customers would be served by the highly skilled sales team thereby optimising the effectiveness of sales overall.

As part of the data sets both locations including actual buildings were key indicators related to the customer’s potential and intent, with time information sourced through the use of generative AI. Other relevant macroeconomic data points feeding on the model are GDP and growth, existing connectivity infrastructure, regulatory and cost overheads as well as other factors related to their specific business and existing footprint with Colt, culminating in identifying the intent and persona of each customer.

Given that sustainability is becoming increasingly important globally, Colt is also playing its part. Regarding traffic routing, they have also been using machine learning to provide optimised routing not just on cost but the environmental impact therefore giving the customer choice and incentivising suppliers to improve their own environmental credentials.

To conclude, for any other B2B interconnect and enterprise connect providers this is surely a very clear-cut case for adopting best practices in B2B AI/ ML marketing and sales, creating a new baseline for the industry to fuel sustainable and profitable growth.

Mizu, Colt Presentation Slides – Network X 2023

Tools and suppliers used in the project:

AI – GPT4 – public domain

ML – Microsoft Azure – Mauro Technologies

Segmentation – to identify related network spend

This website uses cookies to improve your experience. We’ll assume you’re ok with this, but you can opt-out if you wish. Learn more