Automation is a very broad subject. In its simplest form, automation is a tool that takes away some of the manual labor that we’d normally have to do, and that’s not to say that automation is new.
Humans have been automating for thousands of years in different forms and, over time, that has evolved. Take a ﬂour mill for example – that automates the job of grinding grain into ﬂour, historically by using wind or water. It’s a big part of the reason that humanity has had the fuel needed to drive evolution and progress.
Now we’re talking about digitized automation and, speciﬁcally within the marketing sphere, how it helps us communicate with our clients and customers right through the different stages of the conversion process.
For me, automation is there to carry out the manual tasks that slow humanity down. We’re currently taking steps into it replacing more human activities, such as through AI-driven chatbots, or through marketing automation.
But there’s a problem: the majority of people do not yet trust AI, and that runs into automation too.
A study by Pega in 2019 showed that only 40% of respondents agreed that AI has the potential to improve the customer service of a business they interact with, and around 30% felt comfortable with the idea of a business using AI to interact with them. Even more damning, only 9% were very comfortable with the idea.
As a more ﬂippant example, The Terminator was released nearly 40 years ago, and you still hear jokes about AI becoming self-aware and taking over the world even now. Maybe they’re right and the movie was an amazing prediction of the future. They’re probably just jokes created by a fear of the unknown – it’s how some people rationalize the unfamiliar.
A big part of what we do at Hallam is using and developing our own automation to carry out manual tasks that can be easily done by a human, and even then I can see a lack of trust in colleagues when using automation. We use it but we still don’t completely trust it. We develop platforms to automate our internal process, and we still don’t have full trust when they ﬁrst come into use.
It all comes back to the same thing: trust has to be earned.
No doubt this trust will build over time, but we can’t wait for it to build without effort. The technology is here to be used now and that trust will only be earned through usage, as trust must be earned and not expected.
If I was to summarize the following few paragraphs in a single statement: be empathetic and respect this lack of trust.
Don’t try and fool people into thinking your automation is a real human when it isn’t. People aren’t stupid and, after all, we are evolved enough to develop automation and AI, right? I’m sure we’ve all used a website with a chatbot that pretends to be a human, but they’re clearly not. This is a big no as you’re trying to fool the user, and all you’ll do is erode trust in your brand. In the modern age, brands need trust to thrive.
If it’s a bot, in the extreme example be honest that it’s a bot and that it has ﬂaws but it will help as best it can. At the very least, don’t pretend it’s a human. This is being human.
Test and learn
I’m a big believer in testing and learning. By testing ideas in a very small, non-risky way, you get to learn faster and apply that learning quicker to your next test.
I work with a colleague who has written his own monitoring bots for monitoring share and cryptocurrency prices. He wrote them for fun and to experiment (a sign of a great engineer) and certainly doesn’t trust them to manage his ﬁnancial movements, but does trust them enough to get the data so they can make the decisions themselves.
This approach is a perfect demonstration of test and learn. Start with something non-risky that does limited or no damage if it does go wrong. Once you’ve done that a few times, you’ll be able to earn the trust to try the next thing.
If we bring that back to marketing automation, test using automated follow-up emails on a smaller group of customers and measure the interaction and impact, then develop a theory about how it could be improved and run the test again.
We’ve applied the same thinking to some of the automation we’re lucky to have access to with our partners such as Google, and that’s helped us become one of the highest agency adopters of value-based bidding in Europe.
By running small, iterative tests and measuring their efﬁcacy, it becomes easier to understand what’s working and what isn’t, and ultimately build that trust in yourself.
Be clear on risks
If you’re creating your own automation around a speciﬁc business process, there will be varying degrees of risks with the work you’re doing. For example, if you’re developing an automation that automatically emails customers with their invoices, there’s a higher risk of damage than if you’re just generating invoices to send internally.
All software has bugs unless you have the same budgets as Nasa, so expect these and plan around them. Ask yourself:
Do you need that extra feature that’s going to increase the chance of failure?
Am I testing more fundamental parts of this automation before I start messing with the bigger, riskier features?
Can you do a smaller release next time that allows you to test the system is a less risky way?
Don’t be afraid
Lastly, it’s right to have a certain level of healthy fear of new technology you haven’t worked with. This caution is paramount to understanding risk and planning around it. But don’t be afraid and fail to embrace automation. You already have embraced automation in many ways and probably have realized it (think bread) – it’s the next type of automation that needs healthy caution, testing and learning to get right.
Jon Martin is technical director at Hallam.