Chatbot design for telco users
Context
Fastforward.ai is a Dutch start-up platform that built chatbots for Telco enterprises used in social media platforms. People would use them to check their balances, buy bundles, get e-sims or reach out to support. The experience was a mix of native and web functionality, but after scaling, it become scattered in terms of experience, functionality and UI. The platform launched in South East Asia first, then in the Caribbeans and US; from low to high complexity telco operators.
My contribution
At my time at FF.ai (a bit ahead of its time with the name), I’ve increased the UX team’s delivery speed from days to hours by documenting user journeys and UI elements in a design system. I’ve lead remote user research and improved the external mobile web part of the experience. Ensuring consistency and increasing quality was a critical step to scale the product even further, to 30+ markets, on Messenger, Whatsapp, Viber, Instagram.
The existing experience extended the native chat experience to the web, trying to mimic UI elements and interactions; in attempt to naturally extend native elements. However, this wasn’t executed precisely and wasn’t maintained properly, so it degraded. All these were created in isolation, by a few different teams. My job was to bring these together and take the product to the next step.
First, I’ve mapped all common and uncommon journeys, to spot inconsistencies, patterns and ask questions.
In parallel, I’ve ran remote user research with users from South East Asia in Maldives and Burma. This was challenging to set up in the beginning, but interviews were extremely useful.
We learned more about the way they use their phones, their perception of telco operators and their needs in terms of social media chatbots: personalised recommendations, discounts on data bundles, a way to quickly extend from these chats. We also discovered elements that social media channels were adding, that the engineering team wasn’t aware of.
Screenshots from participant’s screenshots during the research. Overall, the experience was too complex, unclear and even messy.
Main research insights, categorised.
What really stood out was the lack of ecosystem thinking in building this product. We, as a start-up were working in parallel with telco’s internal teams on similar products, but across different channel. I’ve highlighted that our product needs to be more clear on the value it provides, differently than the rest.
After mapping out usability challenges as well, I’ve started exploring visual flows too. Had fun with some empty and error states too. Some of the illustrations below are done together with another designer (Silviu).
After testing these new flows, and getting internal feedback and around 3 major iterations, these we gradually launched, starting with low to high traffic markets and following feedback. Then as we documented it all as a design system, before continuing.
The interesting part of this design system was the combination of social media chat native elements, and web based elements. A large part of this system become patterns of conversations, guidelines and predefined responses. Localisation across 30+ markets was a strong consideration too.
Things I learned
- Start-up work is chaotic, but I loved the bonds.
- User insights were a powerful driver within the company. People starting using them as real arguments in discussions.
- A user flow become successful (based on conversion, drop off rate, time on task and reviews) after at least 3 iterations of it.
- It took me a while to realise that the flows we were shipping were subjected to the telco’s brand perception too.