Digital Leader - Connecting homes with Hive CTO, Seb Chakraborty

Profile picture for user ddpreez By Derek du Preez May 29, 2017
Hive is automating underlying systems supporting connected homes. Chakraborty explains the complexities behind personalising for different housing needs.

British Gas, owned by multi-national utility company Centrica, launched its connected home division Hive back in 2013 - and is often cited as one of the primary use cases for the Internet-of-Things. Since its launch it has been installing smart devices across the UK, allowing people to control multiple functions across their home via a smartphone application, and has grown its business to over half a million people.

The company, which has since been spun out of British Gas so that it can operate independently as a start-up, is using Salesforce to support many of its underlying processes - and Hive is looking at how it can make use of the cloud company’s Einstein AI platform to automate these processes further.

I got the chance to sit down with Hive CTO Seb Chakraborty at Salesforce’s recent World Tour event in London, where he provided some insight into the complexities behind providing a personalised experience for a very diverse housing market.

The selection of Salesforce came early on after the company was founded to support customer service. Chakraborty explained:

We launched in October 2013 and it was one of the major decisions we took in terms of what we felt we needed to focus on, which was effectively getting our contact centre up and running. And having had experience of Salesforce in previous companies, it just seemed like the perfect fit. It was a very quick and easy decision.

I think one of the killer factors for us was that we knew that in terms of setting the whole thing up very quickly, Salesforce was pretty much there. As soon as we put it in front of customer agents, they loved it.

We didn’t have anything previously, it was all basic spreadsheets - we were a start-up that was pulled out of British Gas. The objective was, ‘let’s get up and running as quickly as possible and let’s make the implementation as fast as possible’. The implementation took about 11 months in total.

Using Einstein

Hive uses a wide variety of Salesforce products, including case management, service, knowledge management, marketing cloud - all “under the covers” and “abstracted away from the agents” - so that from a CRM perspective, customer service doesn’t need to take manual action to fulfil the processes. Chakraborty explained:

I think Salesforce has been able to help us both in terms of how we get a CRM system up and running, but also all of the other components that are useful to help our customers, they’re right across the board. Going through that journey, that fulfilment process, even that these items need to be delivered to someone’s home, and then on-boarded - we want that process to be as slick as possible.

Any agents that service, they want everything to be hidden from them. They don’t want to worry about logistics, or pricing, or booking an engineering. They want the system to do all that for them. Abstracting that whole thing away, making that whole thing automated, is something Salesforce is pretty good at, which makes the whole experience better. What we want to do is make the conversation with the customer as easy as possible. That way they will learn more about what we can do for them.

And although still relatively new to market, Hive is already considering how it can make use of Salesforce’s AI platform, Einstein. This will likely enable further automation for Hive’s services, as well as provide more personalised interactions with customers in the home. Chakraborty said:

We are exploring Einstein, we are looking at some of the use cases. In terms of trying to connect customer data to actually what goes on in the home with devices - for example, how can we quickly alert a customer to the fact that they need to change their batteries? Those kinds of use cases are quite useful.

This could be particularly useful to Hive, given that there are huge complexities in figuring out the use cases and needs of individual homes - of which there is great variety in demand from customers. Not one solution fits all, and so artificial intelligence could be used to help provide a more tailored experience. Chakraborty said:

It actually makes it more interesting. And challenging. And fun. All houses are different, people are different, building materials are different, customers want different things. Trying to tie all those together is why I have a data science team and why algorithms are useful. And why services like Einstein/Watson are useful, because you can then begin to tailor the service to what their interests are.

We are also trying to do that by building actions into our app, so that you can set up recipes of various types, where you walk into a house and you can configure it where motion sensors trigger lights, maybe trigger your sound system. Being able to give that control to a customer because they want a certain mood compared to others, those are the sort of things we are interested in.

Figuring out the need

In a rather refreshing discussion, Chakraborty admitted that Hive - and its customers - haven’t entirely figured out the use case for the connected home yet. That’s not to say that it’s not useful, it’s interaction figures are impressive, but homes have never been connected before in this way and people are still figuring out what works for them. Part of Hive’s job is to meet people in the middle and provide that usefulness via the Hive app.

Chakraborty said:

The whole connected homes piece, even though we have got half a million customers, customers are still learning about the connected home. Getting their head around that whole process, from ‘where do I find these devices, how do I discover what it can do for me, how do I learn how to make my life easier?’. It’s still not totally there and people have got this notion about what it can do for them, but that first step is often the hardest. Once they have got it, they can start to learn more about it.

That’s been our whole story, we want to get our customers excited as much as we can about what the potential of connected home is. Our goal is really around how to make that whole thing slick, easy and give people back time - the time to do the things that really matter to you.

The home is a place where people want to feel comfortable, people love their homes, people want to protect their homes, people want to protect their loved ones - that’s why connected homes are so interesting, it offers all those things. People want all those things, but they don’t know how to make it happen yet. That’s what we do, that’s what our agents do, that’s what all the technologies we put into place do, that’s what our app does.

Figuring this out has partly been made easier by the expertise that Hive has at its fingertips in British Gas’ engineering team. I’ve argued before that the companies that are likely to be successful in the Internet-of-Things space are the ones that already have the experience, skills and scale in a particular domain, in which new IoT business models can be applied. Chakraborty said:

We were spun out of British Gas and because we were trying to do something completely different to the supply of gas and electricity, we are consumer electronics, we needed to curate that whole piece extremely carefully. So we do operate differently. And the talent we have to hire for is different. We need developers, data scientists, people that know how to look after a highly distributed system.

But we are after the same thing, which is connecting back to customers - British Gas and Centrica want to do the same thing, they want to know more about their customers. The partnership is extremely useful, because the engineering talent helped us build the product, because they have been in the home for 200 years. They know what customers want and they can help us design products. We use their experience because if you know your domain as well as you possibly can, you’re going to make a better product.


Chakraborty said that once customers do figure out the use cases for them, that’s when Hive wins ‘fans’. For example, he explained, many Hive customers are using the application more than eight times a day in winter - showing that there is demand.

However, challenges remain for Hive as it continues to scale its business. And again, much of this will rest on figuring out the use cases for customers as it scales internationally. Chakraborty said:

The challenges are all around, how do you get to be a mass market product? And how do you scale your business to enable that? A lot of people don’t realise that Hive is not just for British Gas customers, you can be a non-British Gas customer and get Hive. We are keen to see that mass adoption. And we are keen to grow Hive to be a global brand.

We have launched in the US and we are excited by that. The challenges are all about how do we make what we have done and scale it. Something like Salesforce is really useful for us, because it is already a global player, and therefore it is already localised and already configured for use and purpose.

Also, take the UK population, they’re all different, houses are different. That applies internationally even more so. We can’t take it for granted that we can take a device that works well here and then expect customers in another region in the world will use that same product in the same way. It’s a data challenge, it’s an algorithm challenge.