For a truly customer-centric model, we need to solve the ‘last mile’ of AI

For a truly customer-centric model, we need to solve the ‘last mile’ of AI
Tim Burke is CEO of Affinio, the marketing strategy platform that leverages the interest graph to understand today’s consumers. With over 12 years of experience in building companies and developing technology that solves some of the world’s hardest problems, Tim co-founded Affinio to mine the billions of relational network connections that exist within any given social audience to shed data-driven light on who each audience segment is and what they truly care most about. For more information, visit:, or follow on Twitter @Affinio.

I’ve yet to work with an enterprise that didn’t have customer-centricity as its top goal. Customer centricity has been the driving force behind countless digital transformation initiatives, and the reason why so many enterprises invest in building a data science team and culture designed to help them better understand their customers.

Just committing to a customer-centric approach is worthy of some accolades, in my opinion. Accepting that one doesn’t know everything about one’s customers is the first step in serving them better, ergo the massive investments in data lakes, analytics, AI and a host of other tools to provide deep insights. So why do organisations still struggle to provide the level of customer centricity they strive for? The answer: it’s nearly impossible to enable customer centricity when tens of thousands of employees are more or less in the dark about their customers.

This gets to the heart of the challenge. Customer-centricity requires enterprises to get critical assets – first-party data in the data lake, AI-driven enquiries, concise answers – into the hands of their workforce in a timely manner.

The high cost of not-quite customer centricity

The benefits of customer centricity are well known and have been discussed at length over the past five years. Less talked about are the costs of not quite achieving true customer centricity. Those costs are calculated in the opportunities that are lost because a sales team couldn’t answer the precise questions a prospect asked. They’re the sales pitches that went nowhere because the data science team is overloaded, and couldn’t provide the insights that made those pitches more relevant to the prospect. They’re the marketing campaigns that delivered underwhelming results because the messaging didn’t quite resonate with the target market. And they’re all the net-new opportunities a brand lost because a sales team didn’t have access to affinity data (i.e. customers that like X also like Z) that would have unlocked potential revenue for them.

Of course, AI tools are complex, models need to be optimised and the data that feeds them must be carefully curated so that it can answer the questions field workers have. But you know what? Google managed to democratise the global body of knowledge. At any point, and on any device, one can get real-time insight into a remote country’s GDP, or find out if the prescription you just filled interacts with an OTC product you take.

What’s more, the interface is so simple that anyone can find what they need by just typing in their questions into the search bar (there’s a good chance that Google will even complete the query for you!). The challenge of simplifying complex AI queries of vast pools of data has been solved before.

Is it unreasonable to imagine a similar experience with an enterprise’s first-party data? I think not. What’s more, I think that this level of functionality would radically transform how we define customer centricity.

I’m going to go a little John Lennon on you and ask you to imagine a world where all employees can answer complex and highly creative questions while meeting with a customer or advertising agency. Imagine the impact on an enterprise when everybody is able to access the vast amount of customer data that in a data lake, but more importantly, leverage it to bring their immediate tasks at hand to a whole new level.

The last mile of AI

Back in the late 1990s, when telcos were rolling out their fiber networks, there was a lot of talk of “the last mile,” meaning the not-insignificant challenge of bringing all that sophisticated networking technology to the buildings and homes of the actual people who wanted to use it.

In terms of AI, the last mile is putting the deep and rich understandings into the hands of everybody in the organisation so they can glean AI-driven insights on a need-to-know basis, wherever they are. 

What does the last mile look like? We can take our cue from Google Search here. The AI last mile needs a simple UI that supports simple queries. It needs real-time responses presented in easy to understand ways. And it needs to be mobile first, so that wherever the user is — in a conference room with colleagues or across the desk from a prospect — he or she can offer useful insights to any question that’s asked.

Until we solve the challenge of the last mile, whole sections of the enterprise won’t get access to the customer insights they need, and that in turn, try as they might, enterprises themselves will never achieve true customer-centricity. 

Interested in hearing industry leaders discuss subjects like this and sharing their use-cases? Attend the co-located IoT Tech Expo, Blockchain Expo, AI & Big Data ExpoCyber Security & Cloud Expo and 5G Expo World Series with upcoming events in Silicon Valley, London and Amsterdam and explore the future of enterprise technology.

View Comments
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *