Sean Maddison

Energy contract renewal & switch

Letting customers quickly & compliantly switch to a new energy plan through voice, chat and messaging.

Project overview

Background

A large retail energy supplier, taking over 200,000 voice calls and 60,000 chats every week. Around 5,000 of these are tariff renewal / switching queries with an increase of up to 50,000 a week during busy periods when fixed-length tariffs end.

Opportunities

Solution

I implemented a voice version of this journey first, as this project was delivered before I integrated our digital channels into our conversational offer. Chat & messaging channels were added in version 2.

The bot allowed customers to hear a high level list of all tariffs they’re eligible for. Selecting a tariff then gave further detail of features, principal terms and pricing. The customer could go back and review any tariff. Once they’d decided they can proceed to confirm the contract, record an audio signature, and complete the journey. If required, they would then go on to setup a Direct Debit and/or book a smart metering appointment in the same journey.

"So we've got FixedPlan version 1 at around £95 a month, next there's SuperSaver version 5 at £80 a month, and SaveOnline version 12 at £78 a month. And we've always got our StandardPlan at £90 a month.

Which one would you like to hear more about?"

Once this was up and running, I then added in a proactive tariff offer into the value-add stage of the automation orchestration: if a customer was recorded a vulnerable, or on SVT, we’d proactively offer tariff switch if we had plans available for them. This helped us towards demonstrating to the regulator that we were trying to help customers move off the default plan.

Design approach

This was delivered in parallel alongside three other initiates: Direct Debit setup, Direct Debit manager and Smart appointment booking. This was key to run in parallel as DD setup is required for some tariffs, as is agreeing to a smart meter appointment!

For all of our bot development, I agreed a set of simple principles for me and my team to follow during design and implementation.

Slide from the principles pack
The principles we agreed early on for all new bots, these would evolve over time with customer experience always being the highest priority

This customer journey & business processes were already extensively mapped out as part of regulatory compliance. We also had a native web application that was well documented, which we could crib some design inspiration from. Taking these inputs I designed and ran some tight workshops with key colleagues from various business areas to map out a conversational-style flow for automated renewals through multi-channel bots.

Early stage diagram for the journey
Early diagram mapping out the core journey stages

Leveraging existing (complex) APIs

We had a universal quote engine that already drives the agent desktop, the native online app, outbound system communications and third-party sales systems. The API implementation was then just a new set of endpoints into existing systems which was relatively simple.

I had the interesting job of writing new XML handler scripts, the average tariff data output from our quote engine was a 3,600-line XML document! I dealt with this by writing a groovy script to parse the output, and extract only the data items required for my journey—this was the most time consuming part to both develop and unit test.

Working in DevOps style with the back-end team allowed us to deliver in parallel, reducing the cost and length of the project

Regulatory & compliance oversight

With most of my initiatives for this organisation, I had full autonomy to design and deliver almost anything. With this solution, we had a bit more oversight than usual. Sales-related journeys were a regulatory hotspot at the time and getting this journey wrong can result in significant negative repercussions for the whole company.

I consulted with our regulatory and process compliance teams at every stage of the design and deliver process—with a content review/sign-off process and regular demonstrations of the new journey. My colleagues in these areas were extremely helpful in getting us to the best design, and our new bot was fully compliant on day 1 of release

Challenges

Benefits

For this and the three other initiatives mentioned earlier, I worked closely with our planning & performance team to calculate the expected benefits for each new service. We came up with a three year projection, and I’m happy to say we exceeded these all within year 1.

We used Full-Time Equivalent (FTE) reduction as the main benefit-driving KPI for this suite of bots.

Benefits chart showing FTE saving over one year for four self-service bots
Year 1 projected benefits for the new suite of bots (voice only)