our views and our knowledge in analytics and other releveant topics

our blogs

Is AI the future of data implementation?


Data is crucial for understanding your website and your customer. It tells you who’s visiting your site, the content that keeps them engaged, and the reasons they leave. However, to be able to gather reliable and actionable analytics, you have to start with the implementation.

Any implementation requires a data strategy and planning. It needs to be aligned with your business goals so that you’re measuring the most relevant actions and customer behaviour to gain actionable insights. It requires an understanding of the site’s technical structure and the tools and technologies that fit your business needs. It is arguably one of the most critical but time-consuming tasks. Therefore, it’s no surprise that the presentation of "Launch It" at this year’s Adobe Summit caused much excitement and left me wondering whether I might soon be out of a job!

So, what is Launch It? In short, Launch It uses the power of AI, natural language processing and image recognition to scan a site, identify key pages and actions, create a strategy document (or SDR) and automatically tag your site with the appropriate JavaScript. Turning a task that would usually take days, or even weeks, into a 5-minute job. Sounds great right? Sure, if you’re working with a very simple site, a very clean dataLayer and a very clear customer journey… and here are where the problems start to arise. In my (admittedly, somewhat limited) implementation experience, I have yet to come across a site that ticks all these boxes. If a company doesn’t know what their checkout page looks like, how will an AI know? So far so good on the job front.

However, we can’t deny that when Launch It is integrated with a tool like Adobe Launch, an API-powered tag management system, it becomes a very powerful asset. Using APIs, we can take the SDR created by Launch It, and automatically deploy code through Adobe Launch. It can add extensions, map data elements and create rules within minutes, and all with a simple click of the ‘Create Implementation’ button. So, whilst a bit more work is needed before we’re able to use AI to automatically create an SDR, the future of automated tagging is perhaps not so far away.

Launch It is currently a concept technology only, and whilst I think there’s still a long way to go before we start relying solely on AI-generated analytics, it’s an interesting step that leverages new capabilities like AI and automation. Whilst the idea of cutting down the time-consuming strategy and planning process of an implementation is a nice one, human intervention will remain very much needed. Hopefully this means my job is safe…for now.