Our Blog



You don't need a bot!

AI and bots are two of the hottest buzzwords in tech at the moment. It seems like every week you meet someone with a great story about how AI transformed their business and they’re completely disrupting the market.

Research houses are warning everyone that companies now reaping the benefits of AI invested long before their competitors. “Wow!” you say. “That’s one bandwagon we should definitely jump on!” Right? Maybe. But maybe not.

Don’t create a bot for bot’s sake! A strange thing to say given we build bots for our clients, but it’s all too easy to get carried away in the latest hype and promises of a better world. Before investing in any form of AI it’s critical to understand what benefits or outcomes its introduction could provide the people who will use it and your business. If you don’t you risk budget, reputation and trust in AI technology that will be hard to win back.

This post examines what areas you should consider before walking into your boss’s office with your gleaming bot proposal in hand. Let’s start with the basics.

What are AI and bots?

Bots and AI are included in the same conversation so much, people sometimes think they are the same thing. They aren’t.

AI is an intelligent machine or program that understands human intelligence and imitates human behaviour. Bots are programs that interact with people through a text or voice interface. Basically we use AI to make bots smarter.

The simplest analogy I’ve seen is this: give a three year-old an iPad. Over time the iPad, through its ability to react to and learn from the toddler’s responses, can teach him or her to say words, to speak sentences, and help achieve their goals. But the iPad cannot teach the three year-old how to decide right from wrong. This would require AI.

Now you know this, it becomes apparent just how far away from pure AI we actually are. It’s not around the corner…but we’ve moving a little closer every day. So far, through the development of smarter AI, bots have evolved from simple transactions (“Hey Google, what’s the weather like today?”) to human-assisted engagements such as booking a hair appointment using Google Duplex.

The leap to full AI is huge and for the moment anyway, we’re nowhere near replacing direct human interaction.

The leap to full AI is huge and for the moment anyway, we’re nowhere near replacing direct human interaction. So, employees shouldn’t shudder at the thought of introducing AI into the user’s experience.

What we will start to see is a shift in responsibilities as bots take over basic and standardised tasks currently carried out by people. Businesses should prepare for retraining as teams up or cross-skill into the grey areas of human engagement where a black and white answer isn’t available or appropriate.

What are bots good at?

As you would expect of a digital interaction, a bot’s strengths mirror the weaknesses of using a person to perform a task.

When evaluating AI look at your business and where adding the above capabilities could increase revenue, reduce costs, or improve the user experience. As you identify places don’t forget to also identify where human intervention is preferable, and the interplay along the user’s journey between person and machine (journey maps are great for this – more on these in our next post on bot design).

For instance, a bot could direct a customer to the appropriate sales rep to deal with a complaint. But until AI is able to demonstrate empathy a human is better suited to resolve the complaint. The customer could then be passed back to the bot to complete any refunds or transactional tasks.

This scenario uses the best aspects of each resource and optimises the complaints process. Staff don’t need training in refunds and the team size is smaller as each rep can resolve more complaints per day.




Available 24/7    

We typically work 9 – 5 and out of hours support is expensive 

Always on brand 

We put extensive measures in place to ensure humans always stay on brand (templated assets, training, guidelines, brand policing)

Cheap and quick to scale 

Extending teams to service increased demand is time consuming and expensive. Also, a degree of redundant capacity is also typically included to allow for flex as demand varies over time 

Data quick to gather and distribute to and from Business Intelligence platforms for reporting

Data analysis is manual, much slower slow and inefficient (humans spend time collating data rather than analysing and identifying insights)

Low running costs

High running and support costs

Machine learning can steer continuous micro-enhancements that can be rolled out immediately

Improvements slowed by traditional, manual decision-making processes


Where is AI being used?

Currently chatbots are the poster children of the AI world but consider that AI is simply the brain, it doesn’t need to have a face in the form of a user interface (UI). AI could be used to create more accurate insurance premium algorithms or predict medical outcomes for patients (Google boasts its AI can predict when you will die with 95% accuracy).  

Below are areas where our poster children are making an impact and some real-life use cases.


Where is AI being used?

 Amtrak’s ‘Julie’

  • Replaced their old telephone-based customer service agent
  • Julie answers 5 million customer enquiries each year
  • Bookings are up 25%
  • $1m annual saving in customer service emails
  • 50% year-on-year growth in user engagement with Julie 

1800-Flowers ‘GYWN’

  • Gywn helps customers search for and place orders
  • She becomes smarter over time as she learns more about your behaviour
  • 2017 revenues up 6.3% to $165m
  • 70% of people ordering through the chatbot were new customers 

Hilton Hotels Connie

  • Standing almost two feet high, Connie is a bipedal robot which interacts with arriving guests
  • She advises guests on local attractions and interesting sites
  • Connie fine-tunes her responses by learning from frequent requests 

Your AI should connect with your ‘Why’

Does a bot fit with what you believe as a brand? There are deeper, more philosophical concepts in play here that relate to your organisation’s ‘Why’. It would be unwise to move forward with any AI development if it doesn’t fit your brand ideals and the DNA that drives your business.

AI might seem inappropriate given your products and services, but it may make perfect sense based on your organisational goals. Take Amtrak’s Julie as an example. Amtrak provides medium and long-distance intercity rail services across the USA and Canada. On this basis it might seem Julie had a relatively small role to play. But if you consider Amtrak’s mission to connect people with their destination, she makes much more sense and delivers a huge impact on their operations.

Look at best practice not industry practice

Organisations like Apple don’t look at what their competitors are doing. They look at companies solving challenges similar to theirs.

Look further than your own industry. Look for examples of how bots or AI are being used to solve a problem or achieve an outcome that your business is also facing. It might be in an area completely outside your expertise. That doesn’t matter. Don’t be afraid to borrow ideas from areas outside your normal comfort zone.

Decide who will use it and what it will do for them

Remember that AI and bots can benefit customers and internal teams so it’s important to understand who the user is up front. This will guide scope, capabilities and personality (if creating AI with an outward-facing persona).

Ask yourself; are your current customers already adopting this technology? Could they? Would new customers get value from it?

The more you understand about a customer’s current experience the greater value your AI will deliver. AI isn’t free, and your budget isn’t limitless so it’s critical to get the most bang for your buck. Your AI or bot needs to start paying for itself in terms of revenue or lower costs as soon as possible. When this happens getting stakeholder buy-in for future enhancements is easier.

A quick and dirty way to get a handle on where AI could add value is a user journey map. They give decision makers an easy to understand, simplistic view of the bigger picture, highlighting which areas of the current experience need improving (there are always multiple places). These can be prioritised and aligned with existing projects and budgets. The second post in this series covers bot design and looks at how user journey mapping can be used to craft the AI experience.

Last…but not least

Get your ducks in a row on how you will report performance. Before you start knowing what needles you are trying to push, by how far, and what your baseline stats are. If appropriate (or available) you should also look at benchmarking yourself within your industry or the domain you are challenging/disrupting.

Because we don’t know how people will use technology, be open minded about the benefits your AI is expected to provide. If anticipated KPIs aren’t moving after launch but you’re seeing positives elsewhere, make sure you get to the bottom of why. Data is your friend and bots can produce a lot of it so your digital listening posts need to include the AI.

Hopefully you have enough insights to make an informed decision on whether AI or a bot is the way to go. Now for the fun bit: designing it. In our second bot post we’ll talk through the key aspects of bot design and how to set your AI up for success.

This blog is part of the #datareimagine series. For more experts' insights, clients' experiences and to download our datasheets, click the banner.

For more experts' insights, clients' experience and to download our datasheets, click the banner #datareimagine

Posted by: Adam Ford, Practice Lead, Digital & Experience Design | 03 July 2018

Tags: AI, bot, customer experience, Artificial Intelligence, #CXreimagine, #datareimagine, #AIReimagine

Top Rated Posts

Blog archive

Stay up to date with all insights from the Empired blog