From App to Smart App

 

Mobile apps have become a feature of our daily lives, helping us be more efficient in our work and enabling continuous engagement with friends and colleagues. Apps have won this special place in our busy days through being purposeful, doing what they do extremely well, by delivering the right service, at the right time on the right device. Technologically, this revolution in mobile work and interaction was enabled by powerful, hand-held devices and ubiquitous connectivity. But it wasn’t the technology that drove the adoption, it was the key apps that were designed to make our lives and jobs easier that made them a success.

The next transformative enabling technology is AI, but like phones and networks, it’s not the AI itself that will drive adoption. It’s intelligent use of intelligence, the application of AI to our day-to-day activities, that will make it a universal helper at the scale of the phone. Nobody needs “an AI” any more than they needed “an app” or before that “a website”. What people need are new and better ways of working, using AI and associated technologies as almost invisible force multipliers. We believe that the best way to deliver those is via the smart app.

Quite simply, smart apps extend the existing capabilities of apps through the use of artificial intelligence and other new technologies. A smart app retains all the key aspects that made it purposeful, building on the existing foundations to augment the human user. Smart apps add the right intelligence to the already potent mix of capabilities that are immediately to hand. The smart part of any app is likely to be largely invisible, blended so effectively into the app that it feels completely natural to the user.

This is in stark contrast to the normal approach which puts the AI aspect to the front. For most people, AI is complex to understand and invokes large amounts of data and specialist modelling approaches. While some application areas really do benefit from this approach, for example, medical diagnoses, AI also works extremely well to enhance many everyday situations, making tasks simpler, less repetitive and more reliable. In fact, many people already rely on AI-based systems without ever realising it.

Do I need to change everything to go smart?

The best place to start a journey to an unfamiliar destination is often from a familiar place, so existing apps and services often make ideal starting points for being transformed into smart apps. One of the most widely adopted uses of AI that many are not even aware is AI falls exactly into this category: Gmail’s Smart Replies. When you reply to an email in Gmail you are offered three short responses based on an AI analysis of the original message. For an invite to a meeting your three responses might be “Yes, I’ll be there,” “Sorry, can’t make it” and “I’m not sure yet.” The short, punchy, personalised responses are so effective that they are already used for 13% of all replies and have been shown to save thousands of billable hours for the large consultancies.

While few apps have the usage of Gmail, it is a perfect example of where adding a smart capability to an existing, widely-deployed app returned hours of time to people. Adding smart functions to most existing enterprise apps can provide productivity benefits, and consumes benefit from streamlining, personalisation and the resulting simplicity.

Any new build app, whether for mobile, voice devices such as Amazon Alexa, or even fitness apps for wearables, benefit from adding smart functions. At CAF all of our clients are asking for AI, carefully woven into the fabric of the experience so that it may not even be noticeable. Smart apps can be updates of well-used apps or brand new ones, offering up new efficiencies.

What sort of apps benefit from the Smart App treatment?

Take a look at some of your existing apps. Where can automation help? Any good app should propose an address rather than requiring the user to type it in, and smart apps extend this natural reduction in data entry. There are many functional areas of apps where information can be deduced using basic AI techniques.

An example might be a for a field-service engineer closing a ticket. The app can scan the ticket content, consider the time taken, and any parts consumed, and propose a completion report. This could vary from a selection of single-word close options to a short, personalised text. Humans don’t handle large numbers of choices very well and having a system to pre-select a number of likely cases can dramatically reduce cognitive load and hence improve response accuracy and quality.

Recommendations is another area that can streamline usage of an app. Instead of making the user trawl through a long list of possible cases, pop the most likely one or ones to the top. While the user may not even notice the shortcut, it saves time and reduces the likelihood of error. How can we tell the most likely selection? Often through looking at banal pieces of information such as the time of day, the location or what’s in the user’s diary.

Most home carer visits to a specific patient will have the same closing status, yet those carers have to hunt through a long list of options. Simply check what the most common response for that user is and make it the first one on the list. Make the second one the most common response for that carer across all their patients. You can guarantee that by default the most often selected values are at the foot of the list in traditional enterprise systems. It’s time for that to change.

In a similar vein, maybe you always update your sales figures at 2pm each day; a smart app will recognise and make sure that if your sales information is only a single tap away. No more diving into menus or hierarchies of files. Box can already do this, and the capability can easily enrich any app with a predictable lifecycle. Even when usage has no regular pattern, preferences and actions can be modelled and optimised, resulting in user delight and corporate gains.

How do I get started?

As with most projects, getting started is the most difficult step. Ignorance of the possibilities, fear of failure, and a concern about needing to have in-depth technical knowledge are the usual inhibitors of any activity. While the current hype and fearmongering about AI makes it even worse, there are many parallels to the introduction of consumer and enterprise mobility and we can learn from that. CAF has built over one hundred successful mobile apps, often introducing them into areas where mobility had yet to tread. This provides us with insight into how to address these three areas of worry head-on.

Taking each concern in turn, look to areas where users stumble, trolleys are abandoned, or where there are open text fields. These are all ripe to have automation applied to providing responses, and that’s the first steps into using AI. Fear of failure is a cultural problem, both in society and within many organisations. Any team looking to overcome that must have the support of senior management and a license to experiment. The worst outcome is that the team addresses the third concern: lack of knowledge. A smart apps project that doesn’t offer spectacular gains still provides learning and experience to the company.

At CAF we offer three specific packages designed to help get past these initial phases of worry. We offer The AI Challenge, an executive seminar and advisory session that allows senior managers to tell the corporate story of AI with clarity. We also have the Smart App Challenge, a workshop that explores AI-based innovation, identifies a key concept and builds a working prototype, all within the space of five days. And finally, we have a range of smart accelerators that bring smart apps products to market quickly through the use of predefined, reliable, common development patterns.

So if you want to join the smart apps revolution and get a jump on your competitors, come in for a cup of tea and a chat about how we can make it happen or send us a note. We'll get right back to you.

 
Richard Marshall