From its inception, IT has always been about efficiency. The very first use case for computers was to speed the work of human computers doing long calculations by hand. From that simple, focused beginning, applying technology within the enterprise has been an exercise in how you could leverage it to improve efficiency in narrow bands of the operational landscape.
And it’s been wildly successful.
Since the dawn of the modern computing era, enterprise leaders have methodically moved through the organization using automation to dramatically increase efficiency within each department. And the results have been astounding. Organizational productivity, on an aggregate basis, has grown exponentially for most of the last several decades.
The challenge is that we’re now reaching a tipping point.
All of that functionally-focused efficiency is now negatively impacting employee productivity at exactly the wrong time. The good news is that AI has arrived at exactly the right time to help fix it.
How Improving Efficiency Has Made Employees Unproductive
There’s a pretty straightforward formula for using technology to deliver function-based efficiency: define the handful of core business processes that the business unit leverages to perform its function, understand the data and workflows necessary to execute those business processes, and then deploy a piece of software to capture the necessary information and execute the workflows. More structure creates more efficiency. Rinse and repeat.
While this formula has had its challenges, on balance, it’s been tremendously successful. While some may complain, suggest to them that they simply throw out the software and go back to a manual process and you’ll have a rebellion on your hands. Most organizations would simply be unable to operate without these functionally-oriented systems.
The problem is that these systems have been designed to improve the efficiency of the function. But they have generally paid little attention to the impact on employees who needed support from those functions or otherwise had to use these applications only occasionally to get something done. While functional systems work great (most of the time) for those that use them all day, every day — they are pits of frustration for everyone else.
When you don’t use a system on a regular basis, every time is like the first time. You have to try and remember how to do everything — and its an endless source of frustration and exasperation.
The challenge, of course, is that it isn’t just one system. It’s dozens of them. Every function has its own system and employees have to figure out how to navigate all them at one point or another. Moreover, this collection of applications creates the equivalent of an employee gauntlet in which they have to first figure out which application to use and then remember how to use it.
The obvious result — one that each and every one of us has experienced firsthand — is endless frustration, a huge loss of productivity as you run the gauntlet, and a generally poor experience. Unsurprisingly, most of us just fall back on calling or emailing support teams, which in turn makes those teams unproductive and drives up costs all around.
But it is the very expansiveness of this problem that makes it hard to solve. Addressing it requires that you not only improve access to one system, it demands that you transform how employees interact with all systems.
The good news is that AI has opened the door to solving this problem. The key to doing so: abstraction.
Striking the Balance Between Efficiency and Productivity
The fundamental issue is that the drive to functional efficiency has resulted in a disjointed maze of applications that employees have to access for any given support process. The very thing that delivers value for the function takes it away from the employee.
The question is how to retain the benefits of all that functional efficiency while simultaneously making it easier for the employee to navigate the maze of applications delivering the functional value?
The answer is that you must create a layer that sits between the employee and these functional systems and provides a simple, unified way to enable the employee to get the support they need without regard to the underlying system delivering the support or services they require.
This is called an abstraction layer and it acts as a translator between the employee and these complicated functional systems. It also acts as an advocate, focused on the employee’s perspective and experience rather than the needs of the functional team.
This abstraction layer is the key to simultaneously enabling functional teams to continue to reap the benefits of their focused systems, workflows, and automations, while at the same time creating a mechanism that interacts with the employee on their terms and from their perspective. At the abstraction layer, the focus isn’t on functional efficiency, but rather the productivity and experience of the employee.
There’s just one problem with this approach: the expansiveness of the employee support use case.
The number of administrative processes an employee must interact with are staggering. They must be able to get support from IT when technology doesn’t work. They need to interact with HR for a whole host of issues. They need to file expense reports, check budgets, request purchase orders, and handle a number of other financial issues. They need to request guest access to a building, reserve a desk, move offices, schedule meetings, book a conference room, and on and on and on.
Almost every one of these actions relies on a separate and unique business process — and many of them a dedicated application.
This expanse is why most companies are not attempting to leverage AI to solve this challenge and create an employee support abstraction layer — or aren’t doing it well.
But it is possible to create an abstraction layer that delivers the support employees truly want and need. But it requires a secret sauce made up of information, action, and reasoning.
The Secret Sauce
There are really two issues with the expansiveness of the broad employee support use case:
- The sheer number of applications and business processes that employees must access to deal with administrative issues
- The vast diversity of the types of actions and support they require
In some cases, employees just need an answer to a simple question. In others, they need a very nuanced answer to a very contextual question (something very specific to their unique situation). Sometimes, they need to take a simple action (e.g., submit a request for a new device, etc.), and at others they need to execute a complex, multi-step process and monitor its progress over an extended period of time.
And each of these permutations can exist across each of the countless apps necessary for employee support.
Therefore, to create an effective abstraction layer you must create a mechanism that can act on information requests, take action, and leverage a reasoning engine to stitch them together.
Information Processing Engine: The information processing engine must focus on creating a knowledge graph that can pinpoint the source of any information needed in the context of the specific needs, rights, limitations, and entitlements of each specific employee. Its job is to understand context, locate, and serve information relevant to the employee.
Action Processing Engine: Beyond simply serving up information, an employee support abstraction layer must also enable the employee to take various forms of action to meet their needs. This action must be taken in the dual context of the employee’s situation and in the context of the target system, business process, and operating parameters. It must be able to sustain that action across complex, multi-step worklows. And those actions must be secure and auditable.
Orchestration and Reasoning Engine: Critically, the employee will probably not understand the nuances of the various underlying systems or business processes. Likewise, they will be focused on their perspective and not discriminating between needing information or action, or from which system they require information or action. As a result, the last critical part of the secret sauce is a reasoning engine that can translate their request into the various components, hand-off elements, and orchestrate the end-to-end process.
By bringing these three elements together, you can create an abstraction layer that interacts with an employee on their terms and abstracts away all of the complexity of navigating through the myriad of functional systems — all without disrupting the efficiency those systems deliver to the organization.
How SymphonyAI’s AIW Delivers
What I just walked you through in this blog post is essentially the process that we followed when we were ideating on and developing Agentic AI for Work (AIW).
As we surveyed the enterprise landscape, we realized how impactful the productivity loss was on employees — not to mention the degree of frustration it created.
And, we saw the depth of the challenge — and what it was going to take to address it.
We developed AIW from the ground up to solve these issues. It simplifies often complex, yet critical business processes for employees. It reduces time wasted on handling these mundane, often infrequent tasks, dramatically improving productivity.
And it does it all from an employee-first perspective. This isn’t about putting AI in front of any one given application. It’s about leveraging AI to solve employee productivity issues and removing the administrative burden of interacting with these countless systems and business processes.
And it couldn’t have happened at a better time. Enterprises are under immense pressure to perform, to transform, and to adopt AI. We believe AIW helps them accomplish all three.
Alleviating the productivity drain improves organizational performance. But AIW also becomes a transformational platform laying the foundation of AI success, paving the way for future transformative AI initiatives.
