skip to Main Content
Back to Insights

With the emergence of AI, sensors, and new forms of connectivity, we’re seeing the first examples of intelligent software agents — systems that provide helpful services with little or no setup, management, or maintenance from users.

It is no exaggeration to say that over the next generation, the creation of vast interconnected networks of intelligent software agents will fundamentally change the relationship between people and technology.

At Scale, we’re calling this picture of the future the “Intelligent Connected World”. A trend-of-trends, the ICW involves a reinvention of applications, devices, and practically everything digital. It carries with it new and better ways for people to interact with their technology, transforming even the most complex software systems into helpful, predictive, and responsive experiences.

And while this transition is already underway, we’re still in the earliest moments of this new era.

Intelligence in Every Thing

A simple phrase goes a long way to describing what we’re talking about here: helpful software.

You might think of most software applications as helpful, but consider how much work we do on behalf of our software. Salesforce is helpful to a sales team only after they learn to reserve time every day to keep their records updated, or hire dedicated staff to administer the system. Adobe Photoshop is helpful to a team of digital artists after years of training and countless hours configuring its interface.



Intelligent software agents empower the Augmented Worker to quickly determine whether or not to take an action , like launching a campaign, based on real-time data the system itself gathered.
Intelligent software agents let Augmented Workers quickly determine whether to take an action, like launching a campaign based on real-time data the system itself gathered.

Helpful software doesn’t feel like software at all. Picture sitting down at your desk and asking your CRM how best to spend your next hour. Or explaining to a photo editor how you want an image to look, then moving on to other tasks while it does the work.

As developers and data scientists get better at hand-crafting machine learning algorithms and leveraging massive datasets, “helpful” features are showing up in applications. The ICW looks beyond these narrow improvements to a time when intelligence is available from practically every tool we use — digital or otherwise.

The Intelligent Connected World’s Enabling Technologies

We’re entering a new era in which software does the work for the user. It’s happening now because three areas of innovation are maturing simultaneously:

  1. Machines are increasingly able to interact with other machines without human intervention
  2. A mesh of interconnected systems leveraging cheap computing power generates machine intelligence that is constantly improving
  3. Ubiquitous “sensory” data capture devices provide endless data streams used to continuously improve intelligent digital systems

Let’s take a look at each of these trends in turn.

The use of machine learning, and deep learning in particular, to augment individual applications is well underway. But entirely new capabilities are possible when machine intelligence is embedded in practically every digital system, and all of those systems have access to and learn from each other. Over time, that transition will be as significant as the shift from client-server to cloud or mainframe to desktop.

The “connected” part of the ICW encompasses the familiar connectivity between devices and networks as well as the coming explosion of interconnection between machines. This latter distinction is important because — as with humans — intelligence in digital systems increases with increased interconnectedness. Deep learning’s neural nets mimic neuronal pathways in the human brain; likewise, the sensors and discrete data-capture devices springing up everywhere mimic a human’s sensory organs, providing a steady stream of raw data that intelligent machines can make use of to improve their capabilities.

These sensor networks call to mind the Internet of Things (IoT). Yet in practice IoT remains a mostly aspirational description of capabilities that can be enhanced by sensor networks. Taking that potential further, however, the ICW envisions those sensor networks become truly ubiquitous: meshes of sensors blanketing everything.

Clearly, the transition from today — with Tech’s many promising yet disparate technologies — to the fully integrated vision of the ICW involves fundamental changes in the interface, application, and infrastructure layers of the software stack.

Which brings us to intelligent software agents.

Intelligent Software Agents and Purpose-Built Intelligence

We call the systems that power helpful software intelligent software agents, which provide purpose-built intelligence. Intelligent software agents put the I in the ICW, leveraging perception (connected sensors) to generate context (understanding of the world). The power and potential of these agents derive from the number of agents and their ability to interact with each other on our behalf.

These agents are built using the many tools in the AI toolkit, but interacting with them — using them to get work done — will be a distinctly surprising experience. Today, developers use algorithms to make software seem smart, though the way we interact with these apps remains the same. The software agents of the ICW, vast interconnected networks of constantly learning systems, have sufficient intelligence to do work on our behalf with little or no input, data entry, or fiddling with settings.

For the foreseeable future and beyond, these agents will remain narrowly focused — purpose-built — rather than generally intelligent. Better-than-human domain experts? Yes. Hal from 2001? No.

When you look far out into the ICW era, you start to see areas of potential technical and commercial disruption. Here are a few such insights — not predictions — into ways the ICW might alter the technology landscape:

  1. Software fades into the background. Individual applications and platforms are always available but largely invisible until needed.
  2. A handful of interfaces connect us to limitless capabilities. Voice- and gesture-recognition interfaces connect users to systems-of-systems that provide services.
  3. Software may become abstracted (de-branded). 1 and 2 above suggest that users won’t open Excel, for example, to perform a complex calculation but command an intelligent agent to get the work done for them. If we no longer click a branded icon on a screen to launch a program, our relationship with software (and software companies) changes fundamentally. The marketing of IBM Watson hints at one way this could manifest, in which Watson is positioned as powering other applications. You don’t access Watson, you benefit from its presence.

Don’t Underestimate the Value of “Helpful”

Though counterintuitive, Scale’s vision of the Intelligent Connected World implies that technology will be more personal than ever. Eventually the manual effort that we consider normal today goes away. The experience of this new world will be to us what it was like for the first bookkeeper who got her hands on VisiCalc: a fundamental shift in work itself.

We are venture investors, so our interest is focused on the creation of new markets and disruptions of existing markets catalyzed by this trend-of-trends. But after thinking deeply on the implications and potential, it’s hard not to feel excitement and optimism for this fast-approaching world of intelligent helpful technology.

Eric Anderson co-wrote this article.

Back To Top