The perils of natural language and search

Natural language is about to revolutionise search, but if you’re relying on the big tech goliaths to help grow your business, then you need to be aware of a few things — not least that the only business they intend to grow is theirs.

Natural language: artificial intelligence?

Driven by uses such as personal assistant apps on smartphones, consumers expect technology not just to hear them, but to understand them and react intelligently too. True artificial intelligence, the type people associate with Arthur C. Clarke, is still a little way off. But intelligent features such as machine learning and implicit personalisation are already integrated into some natural language applications, enabling them to understanding more about an individual, their likes and preferences, through interaction only.

As the use of natural language within search advances, these capabilities will provide the cornerstone to personalising the results. In the meantime, wearables and the IoT are changing the technology landscape. Devices are getting smaller and if consumers are looking at their smartphones with half an eye to the TV watching Netflix, then the interaction needs to be effortless and fast. It’s one of the reasons why Google, Apple, Microsoft, Amazon, Facebook etc are investing considerably sums of money into a speech-enabled user experience.

But a better user experience isn’t the only reason. Data is the key goal. People reveal an extraordinary amount of information when speaking naturally. Data that can be used immediately to understand more about what customers are searching for and why. Information that can be used to close a sale or even upsell, in addition to providing long term insight, from products trends to website improvements that will increase sales.

Overcoming the obstacles

With intelligent analysis and the ability to ask questions back on the fly, it’s even possible to work out not just a specific request, but the sentiment and reason behind it, such as going on holiday or starting a new job. All additional pieces of information that can be used at a future date. For businesses who are looking to grasp the new wave of natural language in search, this presents two problems.

The first is screen real-estate; without a visual aid, being first in the search results will become even more critical. Consumers might get Cortana or Amazon Echo to read out the first couple of results, but the chances are users will just pick the top of the list and buy the best or the most popular product for many relatively low cost items. Unless, of course, they prefer a particular brand and ask for that.

The second problem is owning the data. It’s feasible in the future that some of the tech goliaths will open up their APIs and allow businesses to develop applications that use pre-installed personal assistants, but those APIs are not available today. However, even if they were, caution has to be exercised. Just consider how some of these companies have used customer data garnered from smartphone users to sideline the telcos, and how they are currently muscling in on the banks with their own payment systems. It doesn’t take much to see that their ultimate goal is to use the data your customers provide them with and steal your business. Or at least the profitable part.

As if this wasn’t enough of an issue, businesses are just waking up to the fact that as consumers increasingly move their lives online, the once direct relationship they had with them has floundered. With this disconnect, enterprises are losing the data required to grow their business in the future. A nice looking website or user friendly mobile app doesn’t build brand loyalty, which will be key if you’re not in the top two search results and no longer like the price of sponsored ads.

But like most things in life, with each problem comes an opportunity. Many enterprises already have the information they need to jump start a re-connection with their customers. It’s wrapped up in chat logs and other conversational data, but instead of treating it as natural language, it’s been flattened into tag clouds of individual words with no meaning or context. If you’re an airline you might see 'book' as your most frequently asked question and presume it’s referring to a flight. It’s not. Most topics around 'book' relate to a seat (ask a certain airline that now charges its customers £5 for the privilege).

A business could start developing its own natural language project today and benefit from it in just three or four months time. One that delivers intelligent interactions between you and your customers, providing them with a superior user experience they aren’t receiving from your competitors and giving you an unparalleled amount of data to improve search results, increase sale conversions, and understand future trends. Or you could just ask Siri, but I think she’s keeping that information to herself.

Andy Peart, CMO, Artificial Solutions

Image Credit: Shutterstock/nenetus