5 ways you can secure your conversational future

Customer expectations are changing, and they’re changing fast. Seduced by the offerings from the big five tech giants, Apple, Microsoft, Google, Amazon and Facebook, people now anticipate intelligent conversational interaction with applications and services across all aspects of their lives. Delivering on those expectations is enterprises’ next big challenge and choosing the right development technology is the first step.

While it may be tempting to team up with the tech goliaths, there are several reasons as to why businesses should consider independent alternatives including data ownership, reach and resources. The following five critical capabilities consider these, as well as other challenges in developing intelligent natural language projects.

A platform that is intelligent on the inside

It’s all very well delivering a sophisticated conversational interface for your customers, but if it takes several months to build, an army of developers to code and is near impossible to maintain, then clearly it isn’t up to enterprise demands.

A natural language platform needs to deliver intelligent automation during development, not just in the finished artefact. It must be adaptable to circumstances, such as not relying on masses of training data, which frequently isn’t available, but able to auto-generate the code.

In addition, it needs to enable developers and business specialists to collaborate during the build. A graphical interface can facilitate this and help keep the development in line with business goals, by allowing the team to quickly pinpoint improvements as the project progresses and to identify new functionality previously not considered.

Before making a larger commitment, consider running pilots or workshops. See for yourself what can be created in a matter of days out of nothing more than data the company already holds.

By having truly intelligent software that reduces the need for highly skilled computational technicians by delivering just as much AI to developers as it does to the user, organisations can build their natural language projects in weeks, not months.

The ability to develop independent of OS or device

No one can predict the future and definitely not when consumers are involved. Don’t waste time and resources building a natural language application that isn’t going to be easily ported to the next big thing whether that’s bots, personal assistants or some other IoT device not yet invented.

For instance, given the number of users Facebook has it’s unlikely the company will be motivated to make its bot deployment infrastructure channel-agnostic. By using a platform that will build an application independent of device, service or operating system it makes it easier to deploy across multiple channels, even propriety ones.

In addition, it will make your application scalable and more usable. Customers prefer choice and rarely commit to just one manufacturer, service or device. Being able to talk to your customers using their preferred method of communication is far more important to your business than adopting a technology that will limit its availability in the future.

Deliver a human-like conversation

As Microsoft recently discovered with its ‘self-learning’ Tay conversational bot, there are dangers with machine learning if the right safeguards aren’t in place. However, don’t underestimate the power of human-like conversation when compared with command driven applications. It is one of the most important aspects in terms of user acceptance and the data it provides.

To be successful an interface needs to be able to remember pertinent facts, recognise and clarify ambiguity, continue conversations as the user switches devices, use other factors such as geolocation as part of the interaction and understand the sentiment of the user.

Try talking to an application to discover how sophisticated a technology’s conversational skills are – but don’t believe the demo. Look for artefacts already developed in the technology. Does it understand complex questions? When a follow on question is asked, does it remember the context? If it carries out an action such as setting a reminder, can it cancel it?

You don’t have to believe that you’re talking to a human, in fact in some cases it’s better that people realise it’s a machine. Consider it a success if you spoke to an interface like another person and received the response you expected.

Keep control of your data

There’s a very good reason as to why the tech giants are already embroiled in a battle for ownership of AI and the speech-enabled interface – data.

People reveal vast amounts of information in conversations and it’s this they want to own and control. They want to use it to increase their bottom line with your customers’ money, eventually cutting you out of the loop. For this reason alone it’s imperative that enterprises generate and keep possession of their own customer data.

Sophisticated analytics are an essential part of any development platform enabling information gleaned from conversations to be fed back into the interaction to personalise it further and provide actionable insights in real-time. The data also helps businesses understand current trends, create data-driven applications and train the application or service further.

Furthermore, enterprises need to be able to seamlessly integrate third party data sources. Make sure the platform you chose goes beyond standard API. While modern data brought in will probably be easily integrated, legacy data, which often provides the backbone to an application, is rarely so conveniently packaged.

Deliver a multilingual experience

Global corporations don’t operate in a single language. Enterprises must consider how easy it is to take a pilot project and transfer it to work across the entire company. The simplest way to do this is to separate the language aspect from the functionality.

An intelligent natural language development platform should enable for approximately 80% of the initial development to be reused when adding a new language. This not only maximises resources, but ensures that the main corporation remains in control while allowing some flexibility for local customs or regulations.

Until recently, building any type of AI application that uses natural language was not easy. In the last five years that has changed dramatically. But like all emerging markets there are huge differences in how each technology achieves this. Choosing which one is right for your business will be one of the biggest decisions you make this year.

Andy Peart, CMO of Artificial Solutions