Agile delivery, the public cloud, fancy interactive micro-frontends…there have been many transformative advances in how applications are built over the last ten years.
But what hasn’t really budged a micron is the contents of these apps and how intelligent they are internally.
Everyone in the enterprise today is building apps more quickly, on better foundations and wrapped up in nicer packaging—and this is great! But the business logic of the apps is the same as it was ten years ago.
Having stood static for so long, this intelligence gap at the application layer represents an astonishing opportunity to leap ahead of competitors, provide stand-out service to customers and adapt rapidly to market conditions.
It’s one of the biggest opportunities to make a leap forward in capability that exists in the enterprise.
And enterprises have the capacity to make that leap: they are investing heavily in artificial intelligence (AI) and data science capabilities. But they are building them in isolation from the rest of their business and are struggling to take these capabilities to production (and there are good reasons why).
So we find ourselves in a situation where enterprises have an intelligence gap they need to fill at the application layer, as well as growing AI capabilities. Yet, the two have not yet been brought together.
AI-enabled applications do just that: bringing the power of AI to the application in ways that can provide real competitive advantage.
In this blog, we’ll look at the three main ways you can enable your apps with AI (along with real-life examples) before examining the business value that results.
Applications can be powerfully enabled by AI in a number of different ways:
1) Intelligent data sources
2) Intelligent business logic
3) Intelligent data connectedness
Let’s take a closer look at each in turn.
There has been very little change in how applications read from data sources over the last ten years.
The scope of databases that apps read from has broadened slightly beyond your traditional relational databases to include a few other types of database but that’s about it.
One straight-forward way to increase the intelligence of applications, then, is to connect them to a more intelligent data source: AI capabilities or ML models.
An example use case here is personalised recommendation engines. Amazon or Netflix’s website will be connected to an intelligent ML model that is crunching the numbers in real time on what any particular user might want to be offered next.
And if you doubt the business value of that, just consider what a different quality of service either Amazon or Netflix would provide without personalised recommendations!
Taking advantage of intelligent data sources requires that ML models can be easily deployed to production and the app in question can connect to it. This is commonly known as MLOps: a framework for deploying ML models into production reliably and efficiently.
It also requires that models are updated continuously with up-to-date data. But many enterprises, instead, have static and siloed datasets that they use to underpin their models. Over time, the ‘drift’ between the data that underpins the model and the real-life situation grows larger and larger. They are using yesterday’s data to predict tomorrow, which is just not good enough for the most valuable use cases.
That’s why enterprises must sort out their back end and put in place a seamless, automated, end-to-end pipeline so front-end apps can connect easily to intelligent data sources in real time, interact with these models and read, interpret and present them.
If you have a ‘dumb’ application and a ‘smart’ data source, that’s an improvement. But for some more advanced use cases you will need to weave greater intelligence into the business logic of the application itself.
The basic principle is to distribute intelligence away from a central location and to embed it closer to the sources of data.
There are a few ways of doing this.
For example, by adding computational intelligence closer to where data is produced.
One example here would be edge computing, where an ‘edge’ device such as an IoT sensor or smartphone, has local computation power. Another is TinyML, a technique for integrating reduced and optimised ML apps into hardware. An example is Amazon’s Alexa unit that has an ML model embedded at the local level that does natural language processing at that level, but also has access to wider ML capabilities in the cloud.
By embedding modern data processing paradigms such as real-time processing, streaming and analytics capabilities into the application.
In this way, apps can become part of the network topology of streaming analytics/processing and can then provide real-time feedback to the wider system, which can then respond in turn—creating a highly-responsive real-time user experience.
A great example of this in action is Uber.
The Uber app, as well as the back-end systems that support it, are incredibly intelligent. The user is part of a web of real-time processes that allow them to see their location in real time, the location of their driver and receive up-to-date pricing and availability.
This would not be possible if both the data source and the business logic of the app were not intelligent, allowing it to analyse, process and present that information in a personalised way to the user in real time.
We have seen how to make applications more intelligent by using AI/ML-based data sources, embedding smarter business logic and now we will look at the final approach: connecting data intelligently.
Making meaningful connections between separate data points (or sets) is what turns that data into knowledge.
When businesses say they want data, what they really mean is that they want knowledge.
The knowledge is what allows them to deeply understand their customer, their industry and their own business.
The knowledge graph is the best tool that exists for intelligently connecting separate data sources and representing the knowledge that emerges from them.
This technology can identify data sources and link them together, intelligently determining which field in data set A corresponds to which field in data set B and C etc. and showing the deep interconnection between them.
This enables insanely powerful real-time services.
Take Google Maps, for example.
It is capable of determining and modelling the connections between complex data sets (transport networks, distance to destination, train timetables and cancellations, traffic and road closures, personal transport preferences, weather and so on) and analysing these in real time to deliver an optimal route for the user in a few seconds.
What is going on behind the scenes so that you know you have to get the 14:53 to Farringdon in order to get to your dentist appointment on time is intelligent on all three levels: intelligent data sources, combined with intelligent business logic and intelligent data connectivity.
It’s nuts! And very, very useful for users.
And this intelligence has the potential to create massive business value when used correctly.
There are four main avenues of business value that AI-enabled applications open up.
Intelligence is about providing the right thing at the right time.
The more intelligence you have baked into your application, the better equipped it will be to provide specific and personalised services to each consumer at the right time.
As consumer expectations increase (why would people want to use something these days that isn’t real time?!) that user satisfaction leads to more consumption.
Also consider how providing relevant, personalised up- and cross-sells or the right discounts and offers can really help to drive additional revenues, without even having to create new products or services.
More intelligence, happier users, greater revenue.
It doesn’t matter that users like your product if they like your competitors’ even more.
As I stated at the beginning, there is a big intelligence gap at the application layer, which represents a huge opportunity to jump ahead of your competitors in functionality and usability.
So many companies are stuck with very traditional user experiences built on a foundation of very traditional tech.
This means that there is a massive opportunity to leap ahead with even small increases in capability.
Intelligent applications are well-connected to intelligent data sources and therefore to the conditions of the market and their customer base.
This means they can detect and respond to important shifts in those market conditions and consumer behaviour in real time.
Your apps become much more adaptable and responsive to the world.
If you want to stand out from your competitors you need to be seen to be on the cutting edge.
Take the banking industry, for example. By this point, even the most traditional banks have managed to get a decent-looking app together. That’s now the new normal.
The question is: what’s the next step? That’s where the challenger banks like Monzo and Revolut have stepped up their game by embedding AI and ML into their services.
If you provide outdated (or even just ‘not cutting edge’) apps you risk being labelled a ‘dinosaur’ and having your brand image suffer accordingly.
Many enterprises have the data capability to get started with AI-enabled applications, they just need some help automating and scaling it.
Mesh-AI is a global consultancy that uses data, machine learning and artificial intelligence to deliver transformative outcomes for enterprise organisations.
Our mission is to make data your competitive advantage.
Get in touch to see if we can help you create amazing AI enabled Apps!
If you want to be competitive, you need to sort your data constraints, and that's where Mesh-AI can help. Identify the areas in your organisation that require the most attention and solve your most crucial data bottlenecks. Get in touch with us at hello@mesh-ai.com for a Data Maturity Assessment.
Interested in seeing our latest blogs as soon as they get released? Sign up for our newsletter using the form below, and also follow us on LinkedIn.
.png)
.png)
