4 Things to Consider Before Investing In Artificial Intelligence

Press enter to search
Close search
Open Menu

4 Things to Consider Before Investing In Artificial Intelligence

By Salah Kamel - 07/18/2018

Many businesses have invested in Artificial Intelligence (AI), hoping to automate tasks and produce analytical insights. In retail and e-commerce, AI is often considered the path to a dynamic, personalized, and frictionless customer experience on the web and in stores.

But without feeding AI high-quality raw information, businesses risk failing to replicate successes that increase revenue while building customer loyalty. Before implementing AI and potentially unlocking value, they should ask themselves four critical questions.

1. Is the data clean enough?

Regardless of its type, AI is only as good its data. Bad data yields bad analytics, which can lead to costly business decisions. Good data comes from clean data—which starts with location.

When companies scale, they often collect data in multiple systems: CRM, ERP, even spreadsheets. This tends to contaminate data with conflicts, redundancies, trivialities, and duplications. Stored in a “data hub,” data will be crisp and clean and ready to serve AI.

2. Is there enough data?

Big data now has a dozen definitions, but fundamentally it remains what Gartner’s Doug Laney identified in 2001 as the “3 V’s” of data management: data volume, velocity, and variety.

Today’s digital systems produce so much data that failing to capture it all can hinder capabilities. When AI relies on limited data, even its accuracies will be unhelpful.

Businesses shouldn’t mine customers for every bit of personal info. That’s counterproductive and, in a GDPR world, possibly crippling. But optimizing AI with predictive analytics requires more than a tiny slice of information.

For instance, knowing customer purchase history and product details necessitates a robust loyalty program or troves of third-party data. Multiple customer touch-points? Without synchronous collection, the most important data risks being excluded.

3. Will AI yield a better outcome than current systems?

Today’s most sophisticated AI can drive a car and recognize an object. But AI’s future effectiveness is still a dark unknown.

Powerful AI is not by default a superior solution. While AI can do many things more quickly, accurately, and safely than a human—search the web, create a scientific theory, survive deadly environments—it can’t even read for comprehension.

Trusting a chatbot to handle routine customer queries is rational: one European telco found that chatbots could resolve 82% of calls autonomously. Trusting it to speak with a dissatisfied customer is riskier.

4. What’s the business goal with AI?

Technology helps businesses succeed—not the other way around. Even with the rise of tools like software-as-a-service (SaaS), a company remains driven by business goals underpinned by business values.

When implementing AI, these goals should be articulated in plain language—read legal documents, recognize skin cancer—so that AI can complement humans, especially if employees are taking on technical and intellectual debt to power a platform that could threaten their job.

Summary

In the age of big data, businesses have asked themselves what to do with all the new information. Now they must contemplate how AI alters their mission. But the best AI on the planet won’t optimize a supply chain, deliver better customer service, and reach new markets if not designed to meet those needs. Only humans, equipped with good data and smart business goals, can do that.

-Salah Kamel, CEO and founder of Semarchy