Keep Your Brand Safe: A Controlled Leap into AI Technology
To a reputable business, image is everything. It is something you have spent many years building. It is something you have crafted to almost perfection.
It is who you are.
The idea of bringing new, somewhat unknown, potentially scary (?) AI technology into the mix can give some executives sleepless nights.
Recently, one of our blog subscribers reached out to me personally and asked:
Personal information redacted
In this article, I want to expand on the answer I gave him. How can a large, conservative brand conquer the fear and take the leap into AI technology? Here's what we will cover:
Is the impact of AI on branding a legitimate concern?
Before we dive in, it is worth we explore this question. Can a chatbot truly impact brand in a negative way?
Research we carried out in 2017 showed quite the opposite. 43% of respondents labelled companies implementing chatbot technology as innovative. A true positive impact on brand.
But we are no dupe.
We know this is when things go right. When all goes well, implementing new technology turns out to be amazing -- no big surprise there.
What happens when things go wrong, though?
In 2016, Microsoft released Tay onto the world wide web. What ensued can be described as a branding exec's worst nightmare: racism, fowl language, drugs (and, sadly, no rock'n'roll).
It is important to be aware of the two sides of the coin. As we live and breath AI technology, we also need to be aware of the potential negative impact it can have on business -- if only to know how to prevent it.
Our reader's client is right to be wary. And our reader was right to reach out to us for advice. Let's explore the many ways we make sure our clients never have to suffer a Tay-gate.
1. Make branding plays the exception
I will open this list with a somewhat controversial opinion.
Branding plays are a fantastic way for companies to innovate. Some of the largest companies in the world rely on massive brand plays to make huge profits every year. Think of Coca Cola and their Santa-driving-a-truck-with-music-that-drives-you-nuts-on-Oct-10th commercial.
Equally, diving into AI technology to pull a branding play can provide huge returns. It's also a tough one to pull off.
This was very evident when I worked on Unilever's chatbot last year. Turning one of their most well-known brand's most beloved mascot into a chatbot? Huge brand play.
For a large company like Unilever, it is exponentially risky and damaging if anything were to go wrong. Imagine if our build pulled a Tay and Monkey started cursing at everyone? Yikes.
(Thankfully, it didn't. In fact, it earned us worldwide recognition. Yay ubisend!)
My first piece of advice: try not to leap into AI with a brand play. Make that your third, forth, tenth step. Unless you have a specific project in mind and the right partner to do it with, you're better off focusing on internal process first.
Important note: when a brand play works, it really pays off. This year, we worked on several projects that started as a brand play. For example, we worked with a new political party that used chatbot technology to raise awareness on local issues in the community, open the doors to communication, and made government accessible. It worked (and still does) perfectly.
My point is, I'm not saying don't do it. I'm saying if your company is very brand-aware and slightly worried about things going south, your best bet may be to focus on a something more process-oriented. Onto my next point.
2. Focus on internal processes
What is the opposite of a public, customer-facing, super-scary-if-it-goes-wrong AI project? An internal one!
The principle behind it is simple, and one we hammer on in many different articles. As companies grow, internal process tend to become more complex. What used to take just a quick shout across the office has now become a convoluted series of emails and forms.
Whatever it is that is inefficient, in any organisation of any size can most likely be improved using AI technology.
The good news is, tackling those issues have virtually zero impact on your brand. Because the changes are internal, no one outside of your employees really experiences them. This means you have more freedom to experiment.
Some of the best ways to implement AI into internal processes:
- Turning HR documents into an easy-to-use chatbot (removes the burden of emailing back and forth with the HR staff).
- Use AI-driven algorithms to organise and prioritise job applications (removes the burden of looking through, for some of our clients, literally thousands of applications).
- Build a ticket resolution recommendation tool for your customer service team (fastens ticket solving, makes customers happier without interacting with them, teaches your team).
And many, many more.
My second piece of advice is, look internally. If you're trying to implement new technology, one that some of your board members may be wary of, think of using it to solve long term internal issues. It is much easier to build a business case for it, your business can earn (or save) a lot of money from it, and your board members will sleep better at night.
3. Get your brand people involved ASAP
Depending on the size of your company, your 'brand people' may be John in marketing or a team of 25 that includes designers, UX, logo owners, and more.
Whatever it is, and whichever AI project you decide to embark on, get them involved. From the start. As in, now.
There are a few reasons we found this to be useful.
The first is, we're real believers in the idea of involving all the key groups within a company. A chatbot project is 99.9% of the time something that is going to impact everyone, to a certain degree, within your organisation. It's good to make everyone feel involved.
Get your brand people involved at the conversational flow meetings
The second is more particular to the nature of a chatbot project. Chatbots speak. People interact with them. Whether it is an internal or external solution, real humans will engage with it. Thus, you need to make sure your brand people are there to tell you how it should engage back.
Finally, your brand team might be the best source of inspiration. They may have loads of amazing ideas on how the bot should respond to specific things, what it should do when talking with the CEO vs. talking with an intern, that sort of things.
Those ideas are what take a simple bot into a project everyone can get behind. Don't lose out on the inspiration.
4. Express your risk aversion
I started this article talking about innovation and how people love companies that use chatbots. I then talked about the danger of chatbots for a company's image, mentioning the Tay disaster.
Clearly, we need to find the middle ground. Thankfully, that is entirely possible. All you need to do is talk to the AI company you are working with and express your risk aversion.
We could plot chatbots on a scale like such:
On the far left, we have the boring and barely innovative flow chatbots. Not really risky (aside from the risk of annoying people with how bad they are, yikes), not innovative (since, like, 2013).
On the far right, we have Tay-like builds. Incredibly innovative, superbly interesting, but, of course, very risky.
What you want is to find a suitable, comfortable place within that range:
How do you do that? By communicating with the team that is going to help you make the AI leap. Any company worth their salt will understand the situation and not push you into an uncomfortable position. In fact, in my experience, it is usually our team that has to calm our clients down and rein the project in a little bit.
Sit with the company's key stakeholders and your AI team (or provider) and share your thoughts. Together, find where you fit in the diagram above.
A controlled approach to brand and AI technology
As we've seen, making sure your brand shines in all its glory is more than feasible. Implementing AI technology (whether chatbot or other) should not be done at the detriment of something you have taken years to build.
There are two keys to keeping your brand intact during this process:
Make sure what you're building makes sense from your company's point of view. A fun brand chatbot could be fun, but it might make more sense to start with a smaller, internal, easy-win approach.
Take the right approach to innovation. Where, on the approximate scale presented above, does your company stand? Innovate at a level that makes you feel comfortable.
Finally, communicate. Get the right people (arguably, all the people) involved early on. Communicate with your staff, communicate with your stakeholders, and communicate with your AI company.
Now let's bring that business into the next century!