Tonight’s Peter Ndegwa’s ‘Dear diary’ entry…

Posted by

[Peter Ndegwa is the incoming Safaricom CEO whose tenure starts today 1st April 2020]

Dear diary,

So, today I started my new job as the Safaricom CEO!! Sema excitement gallore. Hio kampuni sini kubwa! wah. Haidhuru, we will ride this wave crest one day at at time. On day one, I got the privilege of meeting many friendly and awesome people in the company. I look forward to working with and for them. There is a lot I can tell you about today but I need to jolt down the below ndio nisisahau. To make sure I have an easy time and achieve my objectives and the expectations of the public at large, I need to do the below asap.

Improve Customer Experience

Nikubaya. Everyone seems to think our customer service levels have gone south. This is a very serious issue because Safaricom’s customer service is a key differentiator to competition. Me thinks we’ve reached the limit in offering human centered customer service. The Zuri chatbot is a good start but we need to make people start using it en-masse, Vernacular stations wamesikia hii story ya Zuri kweli? I will ask the marketing people tomorrow. There is need to divert traffic from 100 to Zuri with priority. I think with Zuri, we can scale customer service big time. This will not only improve our customer satisfaction and experience, but it will also reduce our customer care operating costs because we will not need as many live persons to answer o customer queries and fulfill their requests. I think we should also explore voice like Google assistant or Alexa. With personalization at the core of Zuri, we can not only scale but also work toward creating personalized experiences tailored to the end users needs. for example, someone should set from within Zuri or WhatsApp a recurrent M-PESA standing order or a reminder to carry out a task by use of natural language chat. With Google soon adding more local languages to their NLP engines, our customer should soon get customer support using their local dialect if supported. Hapa iko kazi but the opportunity is immense. We need to leverage and partner with Google here.

Reduce Network Operating Costs

We need to reduce our operating costs like yesterday. Two low hanging fruits hapa ni outsourcing the management of all our towers to a third party and going OpenRAN. This combined can reduce our current CAPEX by up to 60% and OPEX by 65%. This will also help us transition from hii story ya Huawei. The fact that our parent Vodafone is converting all its 100,000 base stations in the EU to openRAN means we also need to realize the cost benefits of openRAN mapema mapema before the big boys eat our lunch. Out towers need to be outsourced to enable us focus on delivering value and not running infrastructure which is a low value activity. Outsourcing this immediately converts most of our fixed costs into variable costs, giving us flexibility both in the books and on how fast we can scale our services.

Move Safaricom from being a Telco to an AI and Edge Computing Power House

With the expected disruption of mobile money by Facebook/WhatsApp and Wechat, add to this the decline in voice and data revenues, it will be important I start positioning this company to survive the future. This future might be sooner than later na hii mambo ya COVID19 kwanza.
We can partner with Amazon in better ways than just being their reseller. Our base station locations countrywide can be used to offer localized machine learning cloud services at the edge. The traditional cloud is too far for most of the coming machine learning and AI systems and we stand a great opportunity to re-purpose ourselves as edge computing providers. I think it would be easier to station a natural language processing ML system at a base station for near real-time responses at low latencies. With the coming 5G base stations, we should consider equipping them with mini data centers to provide cloud computing at the edge. Amazon can lease our space for their Wavelength service or Alibaba with their Mobile Edge Computing Service while at it. These base stations need to evolve from being telecommunication hosting sites into computing and AI delivery infrastructure sites.

We also need to build our AI capabilities pronto. Safaricom should be AI driven if we are to scale the ever increasing demand for personalized services by our customers. Our network operations need to be AI driven to improve uptime and quality of service. These many network downtime’s are partly as a result of our inability to scale the management of such an organic network.

Ethiopia expansion

Hapa iko kazi pia. But at the bare minimum we should go into Ethiopia and build our our network from scratch, hakuna cha ku partner with people running nearly obsolete networks. Initial perception is everything when entering a new market. There will be a lot at stake in this venture and many lessons along the way. we need to learn the market fast and build a modern network that will provide services that should create network effects to gain market share. I read somewhere that KETRACO is partnering with Liquid Telecom to provide high capacity fiber into Ethiopia from Kenya, I need to speak to CTO to see if they are already looking at this in their plan. This way, we can lower our network costs drastically by using these fiber cables to provide the Ethiopia network on C-RAN from Kenyan side or the cloud. Our network roll out costs can drop by up to 40% and speed of roll out can increase if we go this route.

M-PESA is a golden egg laying dinosaur.

Truth only. I need to see now to bring M-PESA to the 4IR. One way is to think outside the box on this. M-PESA should be a one stop money and an online payments portal. Its current user interface is very limiting in making M-PESA online transactions. Conversational AI is one way we can enhance and scale the use of M-PESA. We have to create a platform around M-PESA and not the other way round. We should have an M-PESA platform that is domiciled in WhatsApp. You should order online via M-PESA and not use M-PESA to pay online as is the case. we are missing a big piece of the online transactions world by simply being payment merchants and not commission merchants. Masoko failed because we tried to copy existing two-sided transaction platform models as opposed to thinking outside the box. Easier way would have to have the ability to login to the M-PESA portal via whatsApp/Zuri, place orders for whatever you need. On the back end, we can get vendors and suppliers to respond to conversations on Zuri’s whatsApp based on what has been ordered. ‘Zuri order me pizza’ should have Pizza Inn and Domino’s paying us to respond to this based on some metrics like this user prefers ordering from Dominos or vice versa, location etc and have little cab/boda do the delivery all from within M-PESA. Hapa tutapita na wengi ….. but that’s why they hired me..

Yawn… dear diary, I am tired and need to rest. Day one FTW!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s