Now you’re talking

x Sponsored content on Research Live and in Impact magazine is editorially independent.
Find out more about advertising and sponsorship.

Artificial intelligence (AI) now underpins many of our daily interactions, whether we realise it or not, and brands are increasingly tapping into AI to improve their services and communicate with consumers.

The global chatbot market is expected to reach $1.23bn by 2025, according to a September 2017 report by Grand View Research. Chatbots – which allow companies to converse with individuals via online platforms, such as Facebook Messenger, or within a website directly – are also broadening the rules of engagement for research. 

As people become increasingly familiar with communicating with brands via chatbots, marketers are turning to these interfaces to conduct surveys and gain insights from interactions with consumers. More than just a gimmick, chatbots, when used in the right way, have the potential to transform how research is conducted in future – from improving participant engagement and informing the ‘System 1’ approach to research, to reducing costs. 

“We’re seeing organisations tapping into Facebook Messenger to run extensive quantitative surveys, but this could only be the start of what’s possible,” says Catherine Crump, chief executive at insight consultancy FreshMinds (now rebranded as Decidedly). “As chatbots become more sophisticated – moving beyond fixed responses to a limited range of questions and, instead, using AI to learn intelligently and respond to the inputs they’re given – we see their applications extending to qualitative research.”

Audience interaction

Querlo, an AI start-up that uses IBM’s cognitive intelligence, Watson, has moved neatly into this space. It has a platform that allows brands to use chatbots to interact with audiences for customer experience and insights purposes, and is being used for B2B and B2C research. 

The company builds custom chatbots and conversational AI solutions that are designed to be embedded within websites. Text analysis pinpoints specific words in the conversation, so marketers can gauge the individual’s sentiment and emotion. Gamification can offer users incentives, such as rewards.

To date, the technology has been used by companies such as HSBC, Unilever, Pepsi, Tetra Pak and Lavazza.

“Brands need to take control of market research in an intelligent way,” says entrepreneur and philanthropist Francesco Rulli, Querlo’s founder. He says “nimble” strategies, such as chatbots, could be used more to help marketers compare and bolster research findings, particularly when the costs are high in terms of time and spend to build panels. 

“If you just spent $100,000 on a research project that took you two and a half months, why not spend a couple of thousand dollars doing a side-by-side evaluation of what people are talking about – so that, if you see major discrepancies, you can sit down with the research company and analyse why they are there.”

Honest Conversation

With heightened awareness around personal data after the recent furore involving Facebook, Rulli believes brands could use chatbots to establish an honest conversation with consumers about the value of their data and how it will be used. “Companies can have a very direct conversation and say, ‘if you answer my questions, I’m going to use this data in a certain way’; or ‘I’m going to reward you for the information you share with me’.” He suggests developing an approach that is akin to the way airlines reward frequent flyers, with the number of points clearly visible when making a booking. 

One problem is the sense that chatbots are preprogrammed with answers, rather than having a genuine conversation, but – as AI evolves – platforms are becoming more sophisticated in mimicking language and responding intelligently. 

Khaled Ismail, vice-president of communications for Europe, Central Asia, Middle East and Africa at Tetra Pak – which worked with Querlo to build chatbots for B2B research – says: “It was a novelty in the beginning, especially when you’re dealing with customers and people whose opinion you want. But, with AI included in it, you have something so smart it makes it intriguing to see what the next response will be.

“We’re trying to use it to capture information in an interactive way, so that the person on the receiving end feels they are not just responding to a survey.” But, Ismail adds: “It’s all to do with how much you invest in it – it can be basic, or you can make it really witty and interesting.”

A 2017 study by Michigan State University and YouGov found that participants rated an overall survey experience more highly when they took it via a Facebook Messenger chatbot, on their smartphones, compared with the typical YouGov, web-based survey platform. 

Surveybot, a platform that allows businesses to create their own chatbot surveys for Facebook Messenger, has benefited from the behavioural shift towards chat apps among young people, according to its founder Liam Houlahan. “People spend more time on messaging apps than on social media, and email response rates are no longer what they used to be, especially for younger demographics.”

One example of this is Surveybot’s work with a media company to survey its radio station audiences about song choice each week. “One of the stations at this company had a huge problem with their audience no longer responding to their email-based surveys. This station has a very young target demographic [ 18- to 25-year-olds].” Houlahan claims the brand fulfilled its research quota within the first week, going on to test 40 songs every week using the platform.

Participant Diversity

Other Surveybot clients range from large corporates – which use the technology for internal human resources surveys and event registrations – to research agencies conducting field work and retailers capturing feedback data while shoppers are in store. According to Houlahan, the chatbots also achieve a greater diversity of participants, by reaching people in a space where they are already spending time. 

Beyond respondent engagement, chatbots could also have bigger implications for the way the insight function works within organisations. “Imagine how powerful it would be for employees across an organisation to have access to an insight chatbot, able to scour previous research studies and give instant answers to stakeholders’ most pressing questions,” says Crump. 

“It would help to raise the profile of the insight function and ease the administrative burden, allowing client-side insight professionals to focus on the context, commercial knowledge and counsel that’s needed to help senior stakeholders make key decisions.”