Marketing in the AI-era of search: your guide to creating exceptional customer experiences
Purna Virji, Senior Management of Global Engagement for Microsoft, gave her session on how to improve the customer experience in the AI-era.
Research conducted by Microsoft found that over 70% of people have never used voice search to make a purchase. Furthermore, when a voice app acquires a user, there is a 6% chance that the user will be active in the second week. When voice search is being regarded as the next big thing, companies need to realise that consumers will only want to use voice if the customer experience (CX) is better than other existing options.
Virji gave the audience an example that it is the norm for families in the US to want to travel somewhere during the Thanksgiving holidays. The normal protocol for American consumers would be to search “Philadelphia flights to Dublin”, “Philadelphia to Paris” or “Philadelphia to London” and so on. She argued that Skyscanner would already know all the information about this for consumers and for AI to be chosen first by consumers in a situation like this, why can’t websites as such implement voice search to increase conversions.
The science of CX and why it matters
Laziness is entrenched deep in human nature. Daniel Kahneman once said that “The law asserts that if there are several ways of achieving the same goal, people will eventually gravitate to the least demanding course of action.”
Virji argues that there are two theoretical systems govern how we make choices. System 1 is fast, emotional and intuitive. Whilst System 2 is logical, deliberate and slower. She stated that in order to increase conversions, the cx must be easy, therefore companies should aim to keep their customers in System 1 as it is seamless, frictionless and easy. In contrast, System 2 brings scepticism and counter thinking which can stop consumers from making a purchase. Brands that gain success through AI and voice search keep the conversation with consumers moving smoothly and efficiently – They resolve glitches and problems that are encountered quickly to win loyalty and profit.
The principles of conversational design
These are the 4 C principles that Virji detailed to increase conversions and customer loyalty through voice search:
- Clarity: In order to increase repeat activity with consumers, companies should focus on ensuring their technology allows them less time to figure out what their bots mean, and give them more time to do the things they want. Clear and concise questions are ideal in order to keep their concentration. They should aim to kick it up another notch by adding another question at the end of the options so the customer continues using the interface. Virji recommended that companies should write sample dialogue and roleplay interactions before even considering coding.
- Character: AI needs to have a distinct personality. Companies should create a persona for them and picture it when writing so they can plan how they will react in certain situations. They should be friendly, engaging, seamless and not distracting. However, it should remain clear to consumers that they are speaking to a bot. Research at Microsoft and Facebook show that you should not trick customers into thinking they are talking to humans as they tend to be more understanding when communicating with a machine with limitations. Whilst not a lot of companies build a tone of voice guide for their bots, their AI’s voice should be professional, friendly or irreverent – this personality will depend on company branding.
- Compassion: The skills of a bot must always be considered a brand representative of the company. Therefore, companies should aim to make their bots better understand and resonate with humans. Virji gave the example that bank or financial apps should give their consumers a warning if they use voice search to pay for something or transfer money if they are about to go into their overdraft. On the other hand, encountering small talk is pretty common for a bot and it is where a lot of bots struggle. Small talk scenarios should be planned, perhaps by using a spreadsheet that categorises phrases by user intent, to program responses.
- Correction: Much like a sales assistant in a traditional retail sense, when a bot apologies too much it ruins the cx, as it shuts down the conversation. Instead, companies should try to resolve the problem and offer an alternative. For example, if a consumer tries to buy something through voice search that is sold out, the bot should offer an alternative to increase the chance of conversion. Furthermore, in order for humans to not dwell on the glitches that AI make, bots should confirm that they have made a mistake (“Got it”) and move forward.
In short, companies that want to capitalise on the rise of AI and voice search should do the following:
- Focus on clarity
- Add suitable character
- Embed empathy and compassion
- Have guardrails for correction
Virji concluded her session by reminding us that the early days of the internet were a slow, clunky and overall poor experience. If companies strive to resolve their problems and glitches quickly through the principles that she mentions above, then we will soon remember AI and voice search in a similar way.
