Skip to content

How to Use AI as a Butler, Not a Stalker

6 minute read

For AI to cement its place within consumers’ lives over the next 10 years, companies need to employ the technology as a ‘butler, not a stalker’. This analogy, used by Dr Nicola Millard, BT’s Principal Innovation Partner, should be remembered by customer experience experts as a guiding philosophy to ensure that their services remain trusted as they hand over responsibilities from humans to machines.

“Butlers are nice to have because they anticipate your every need,” she explains. “They predict things based on what they know about you. Stalkers inherently feel creepy. If a customer starts to feel that you are behaving like the latter, then they are probably not going to let you have their data.”

The businesses that will succeed in the years to come will be the ones who understand how to tread that line – they will employ AI technologies to maximise a personalised customer experience, but they will never cross the creepy line.

In this article, we’ve outlined a step-by-step guide that brands can use to get their customers used to being spoken to by AI.



The first step in communicating with your customers using AI is personalisation. After all, offering a tailored experience based on what you know about someone is exactly what a butler does.

But time is running out for brands to assemble the tools to do this well. Google’s plans to phase out third-party cookies means the race is truly on for companies to get their first-party data in order. While this means personalisation is about to get harder for many, the move will only boost companies that are doing the right thing.

Consumers don’t like it when they can’t see the link between the information about them being used and an action a company has taken. According to the 2022 Acquia CX report, 76% of consumers become frustrated when brands suddenly appear to them because they’ve searched for them online. 83% believe their data will be more private when web browsers stop using tracking cookies.

All this points to the return of an environment that’s a bit more familiar – companies that know their customers the best will be the ones that perform the best. If you can use what your customers tell you about themselves to offer a better experience, then you can lead the way in customer service in your sector.

For example, in the UK, food and clothes retailer M&S recently acquired the fashion app Thread, which uses AI to suggest items to its customers. This creates the experience of a personal shopper. The move will enable M&S to add Thread’s technology to its website. A customer will be able to join the dots from the products they’ve shown interest in and the ones they are recommended. This will help them to offer a butler-level personalised service, not the feeling of a stalker, breaking in and looking through a person’s wardrobe to find what they like.



Once your customers are used to receiving a personalised experience when they come to your site, you can then move onto the second step – proactively communicating with them. As Nicola explains: “This is where you have the relationship with your customer that is: ‘Tell me what I want to know, when I want to know it, and on the right channel”.

If you’re proactively reaching out to customers, entering their private messages, emails, or any other channel that you know they prefer communicating with, then you need to make sure that your content still verges on the side of a butler.

Brands should start by using AI to communicate messages that are of value to the customer. Rather than offers and recommendations, consider using AI for information about deliveries or changes to their order.

“Proactive has to work for the business and the customer,” says Dr Mark K. Smith, General Manager, CXone – ContactEngine. “We typically begin this process by looking at where our clients are wasting the most money and analyse whether we can improve the experience and/or reduce cost. A lot of the communication we start with is around location-based events, so for example the arrival of an engineer at someone’s address. If you can put a number on what that costs you – most companies can – then you can measure the reduction in your failure rate. Start where you’ve got a win for yourself and for your customers.”

But beyond the content of a message, a great butler doesn’t just understand what you want to hear and when, they also understand how you like to hear it.

A recent study from Intercom found 38% of customers would leave a company if they were not treated as an individual. But how can you expect your contact centre to understand the nuances and needs of millions of customers? Take emojis. 58% of respondents said they found it acceptable for businesses to use emojis – younger generations are four times more likely than older ones – but 20% would leave a business if they used too many. 55% of customers prefer casual language – but that still means 45% prefer a professional tone. A one-size-fits-all approach doesn’t work. An approach based on blocky ‘customer types’ isn’t much better.

The only way to ensure your communications match the tone every single one of your customers expects is through an advanced AI that can adapt to these different expectations.



By this point, you’ve got customers trusting you with enough data to create a personalised experience both on your site and proactively. The next step – and one that will take your toes right up to the line of being a stalker – is prediction.

This means different things for different sectors. For example, a bank could use AI to predict what transactions are fraudulent and use that to protect their customers quicker. In retail, it might be using AI to find the customer who buys all your new products on their first week in market and making sure they don’t miss a new launch.

This area will be all to play for as the internet of things spreads and creates new possibilities. Consider a cruise ship, where an individual’s food allergies and preferences are shared throughout the different dining areas. Within a few days of being on board, individual eating habits could be picked up and prompted. There are also opportunities within renewals for companies such as Telcos. If a customer has upgraded every time, might they be well-placed to receive offers in the month before their contract is up?

This area needs to be the focus for all brands who are looking to become the leading butler in their sector.



Being a butler, not a stalker, is essential for AI to succeed in commerce. The brands that do the right thing will earn their customers’ spend, while the others will lose customers and sales. By following these simple rules, you can make sure you stay on the right side of the line:

  1. Only use first-party data – information about a customer that they have knowingly given you.
  2. Ensure you do not appear to be using AI solely to get more money from consumers. Let them feel a benefit.
  3. Consider a personalised tone and medium for your message, as well as the content.
  4. Review and analyse each AI initiative you introduce to make sure customers remain comfortable.
  5. If something feels like it might be too creepy, then it probably is.

By Ian Haddon, COO, ContactEngine


Keep up to date with the latest events, resources and articles.

Sign-up for the Engage Customer Newsletter.