Want AI in Your Marketing Department? Relieve Developers First

I happen to believe that this is a litmus paper for what marketers or growth hackers want to read, explore, and eventually apply in their businesses nowadays.

Coming from a tech background, I’m kinda surprised this is becoming mainstream these days. Although fat cats like Netflix or Google clearly benefit from using advanced algorithms such as ML, I believe there’s a limited set of problems you can solve with machine learning in the vast majority of marketing departments. And having been a part of software development teams building tools for marketing departments for many years now, I can’t help but think that introducing AI should not be (generally speaking) a marketers’ priority at the moment.

Anyway, if you or your boss still wants to experiment with some AI ideas, I have a “one surefire piece of advice that can [put your desired AI outcome]” kind of tip for you; Artificial Intelligence is all about software. So, it is a software development team that should be involved in this issue in your company. But developers are scarce and always busy, so if you really want to take that first step towards having AI supporting your marketing efforts, you should figure out how to give your software developers more time to wrap their heads around machine learning and the AI family.

And believe me, developers would love to jump into AI right away. Developer-oriented websites like reddit, Hacker News, or HackerNoon are full of articles and comments presenting open source tools, strategies, and use cases from top engineering teams around the world. Your developers probably can’t wait to take all those shiny toys and use them in their own playground. What stops them is actually, amongst others things, your department - through countless little requests. So, here’s what you can do:

• ‍Listen to Scott’s Brinker’s now 7-years old advice and learn something about programming - I’m not saying you need to become a full-stack app developer, but a dash of SQL skills to analyze data on your own can be useful and it’s really not that hard. Likewise, you can try understanding how web applications and servers work (a nice guide), so that when you create an error ticket, your message has more context and more details, and in effect your issue can be fixed more quickly.  
•  Meet your devs and learn how you can help them get their job done. Perhaps a small change in the way your department works can result in huge savings in engineering and maintenance time. Or, try to prioritize and drill down requests before you put them into the development backlog; there’s nothing worse for a dev team than discrepant and vague requirements.
• ‍Partner up with developers to figure out which software to buy and which to build. Specifically, make sure to take the hybrid approach into account. So-called developer-first API platforms hit the market a couple of years ago and are now becoming an essential tool for modern companies. Why? Because they satisfy both marketer’s and developer’s needs. What they offer is not a ready-to-use product, but a set of developer-friendly building blocks. Your dev team can then use them to  build your applications, adapted 100% to your process and in no time at all. This is the opposite of the out-of-the-box software which is usually rigid and makes you adapt your processes to the tool. This is also the opposite of implementing features from scratch - which sometimes takes months. If you want to learn more about researching these sorts of tools, I encourage you to take a look at this article, meanwhile a handful of examples to let you understand the context:

• Your content doesn’t look the same across various devices – consider replacing your legacy CMS with Contentful.
• Your customers can’t easily find your products in the catalog – how about applying a lightning-fast search which gives you typo-tolerance and synonyms out of the box, like Algolia?
• Want to run more A/B tests but the development is slow and the experiments are hard to manage – tap into LaunchDarkly and roll-out experiments with ease.
• Malicious users find more and more security issues – take a look at Snyk and include a vulnerabilities scan in every feature release.

There’s many more. You can also seek developer-oriented upgrades within the software you already use. For example, if your Salesforce workflows are not up-to-date with your processes, allocate your budget for migration to the recently announced Salesforce DX to make SF development way faster. Just talk to your CTO, the possibilities are endless.

In summary, don’t start your AI journey by following the big cats and their super-effective AI transformation presentations. There’s much more relevant knowledge already in your developer team. Ask them. I bet they will provide you with a handful of low-hanging-fruit solutions to your problems you will want to employ AI for without tapping into AI at all. So, instead of reading yet another article about how a new chatbot revolutionized X or how the Y company tripled their revenue with a brand-new Sales AI, start small, with a problem-oriented meeting with your CTO. And make sure you don’t steal developers’ space for creative thinking.