in this week’s long read, we’re taking a close look at the emergence of bots and ais in digital marketing. what are the implications for the digital marketing industry, and what methods can digital marketers use when creating scripts for their own brand’s bot?
digital marketing ‘bots’ and ‘ai’s are just scripts, integrated into an app interface. they rapidly scan the user’s text input and deliver a response, pre-designed and pre-loaded by their creators.
bots and ais have typically performed a customer service role, answering queries and directing users to helpful contacts and resources. now, increasingly, they are also used to participate in or even initiate sales conversations. the biggest existing platform for sales-focused implementation of bots and ais is the facebook messenger app, which announced the launch of a sales bot development “ecosystem” for ad clients in april 2016. if you want to sell your product in messenger, you can now make a bot to do it for you.
5 facebook messenger bots you can try right now!
you can try interacting with some facebook bots in real-time right now! click on the links below to go straight to facebook messenger to interact now (click on ‘get started!’ once loaded). you’ll see they are fairly limited right now, and you need to understand the context of what the bot is supposed to help with (buying flowers vs. buying jeans for example) but it does give you a glimpse into the future of artificial intelligence interactions:
klm airlines facebook messenger bot – a bot to help you book fights.
healthtap facebook messenger bot – a bot to help you answer health relates questions.
the wall street journal facebook messenger bot – a bot that give you the latest financial news and
fynd’s facebook messeger bot – a bot to help you buy clothes.
1-800 flowers facebook messenger bot – a bot to help you buy flowers in the us.
facebook bot in action: the wall street post facebook messenger bot
of course, there are still situations in which digital marketing bots cannot answer a customer’s question – their programmers simply cannot second guess any question the public might pose. standard practice in these scenarios is for the chat feed to refer the customer to a help page, or a customer support agent of their own species. think of a marketing chat bot as a highly evolved email autoresponder: though the variables are far more numerous, its job is also to deliver a certain response to an online conversant, based on a certain input.
the case for bots in digital marketing
the bot and ai scripts you’ll find in commercial applications don’t strictly invent their own solutions to problems – they are not intelligent in that sense. rather, they are meticulously programmed to understand a vast range of words and phrases, and to know how best to respond to each query.
perhaps the greatest benefit of customer-facing chat bots is their potential to rapidly absorb quotidian customer enquiries and deliver adequate responses that progress the conversation. why waste a human being’s time on replying to the same bread-and-butter questions every single day when a machine can shoulder the same task to similar effect.
from the employer perspective, there’s also benefit in the fact that a bot doesn’t require a paycheck at the end of the month – though they do require maintenance and development.
of course, chat bot apps aren’t just a tool for satisfying the customer’s curiosity – they’re a fresh marketing channel. clinching sales through digital chat channels is going to become an industry unto itself over the years to come.
where do human digital marketers fit into the picture?
the great controversy surrounding bots and ais – and, indeed, surrounding the wider automation of the workforce – is the human cost, as measured in job losses. whilst development and maintenance of advanced b2c bots will certainly create work for a number of developers and higher-level digital marketers, these direct beneficiaries of the innovation are likely to be outnumbered significantly by the many customer support agents whose jobs are put at risk.
there is a way forward for digital marketers working in the field of digital b2c chat: namely, studying and mastering the art of scripting for bots and ais. see the section below for a few tips on how to get started – though we must stress, they’re just the tip of the iceberg and don’t even touch on the vital arts of channelling customers and segmenting your audience.
advanced tips: elocution for chat bots
when you converse with an artificial intelligence, you’re still receiving human idioms, human voices. a chat bot can articulate itself like virginia woolf, shakespeare, lauren goodger, batman… perhaps even all of the aforementioned in nightmarish combination. the crucial point is that your options are plentiful and your mind should be open. here are our top tips on making your chat bot a fine conversationalist.
- people love to personify things – take a quick look at the twitter account for the european space agency’s philae probe (from the rosetta mission). notice something… charming? the probe ‘refers to itself’ in the first person, and the denizens of the twittersphere seem to get right on board with the act. imitating this effective tactic in your chat bot app is easy – just write in the first person (where appropriate) and give it a name! emotive signals like humorous, positive or negative language can also work like a charm.
- getting the right tone of voice – chat bots and ais are such a new phenomenon that the jury’s still out on how the tone of voice used should relate to the brand. one school of thought says that a chat bot’s voice should be a watered-down approximation of the brand’s own – and that certainly sounds like a safe bet to us. on the other hand, chat bots can represent an opportunity for typically straight-and-to-the-point brands to cut loose and show their fun side (we’re looking at you, slack & slackbot). a/b testing would be the surest way to determine the right path for your chat bot.
- gunning for conversions – if your chat bot or ai is sales-focused, you’re going to want to gently steer the conversation ‘onto the rails’ – i.e. you’re asking product-related questions that steer the customer through the sales process, whether you’re selling them a movie, a t-shirt, or whatever else. chat bots designed to sell products will always differ from those used simply to support or engage the customer – the focus here is centred on clearly and concisely relaying product information, with just a few nods to brand and voice.
- making a chat bot is an iterative process – it’s impossible to program a chat bot with a conclusive response for every conceivable input – for one thing the possibilities number beyond any marketing team’s capacity, and for another, clarity of written and verbal communication varies from person to person. when maintaining a chat bot, regularly assess the inputs that led to an inconclusive response, and devise catch-all solutions to stop the same thing happening in the future.