Most AI conversations are capacity conversations. The opportunity is knowing what can be solved today, and where AI can take us next.
By Ron Rosansky, CEO, Akkadian Labs
Nearly every executive conversation I have eventually gets to AI. The use cases vary. The objective rarely does.
Organizations want more capacity. They want to serve customers better, move faster, reduce repetitive work, and make better use of the people and systems they already have. They want to grow without adding cost and complexity at the same rate.
That is what makes AI compelling. It is also why the conversation should start with the outcome, not the label.
I’ll start with our own experience. AI has changed how we build, even as our platform has long delivered the capacity organizations are now trying to create.
AI changed how we build
AI has already changed how people work in nearly every role at Akkadian. The clearest example is software development.
For years, our senior developers—many with 10 to 15 years of experience in workflow automation, integrations, and security—wrote the code behind the capabilities our customers rely on. Today, those same developers have built AI agents that they now curate within the architecture, standards, and security practices they established.
For certain types of development, that is allowing us to move roughly six times faster.
The number is not the point. What it does for customers is.
We can bring requested capabilities into the platform faster, shorten the time between a customer need and a working solution, and increase the value of the platform more quickly than before.
Our developers’ experience matters as much as it ever did. They set the architecture, establish the guardrails, review the work, test the output, and protect the security and quality of the platform. AI is not replacing their domain knowledge. It is extending its reach.
That is the promise of AI at work: experienced people applying judgment and expertise at a much greater scale.
It is also why I am deliberate about where we apply it next.
Complexity has moved, not disappeared
Collaboration technology has genuinely improved. Cloud platforms are more capable, native interfaces are easier to use, and APIs have opened the door to real integration.
Individual tasks are simpler. The environment around them is not.
Most large organizations operate across multiple vendors, cloud and on-premises systems, identity platforms, service-management tools, licensing models, security policies, locations, and business applications.
A single role change can affect calling, devices, numbers, licenses, permissions, identity, and connected services. A migration can involve thousands of users and years of accumulated exceptions.
The hard part is no longer completing a task in one console. It is coordinating the full process across the organization.
That is where domain intelligence matters. A platform built for collaboration lifecycle automation carries years of operational knowledge: which systems are involved, the sequence changes must follow, the dependencies that can quietly break a process, and the policies that must hold.
Akkadian turns that knowledge into governed execution across Cisco, Microsoft, Zoom, and hybrid environments. We are not simply automating clicks. We are applying a proven operating model to work that would otherwise depend on manual effort, scarce expertise, and coordination across teams.
Capacity only works when it is governed
Every executive exploring AI is also working through how it will be used responsibly. The work has to be authorized, policy-compliant, predictable, traceable, and, where appropriate, reversible.
Governance is not a luxury. The challenge is moving faster without losing it.
A process is not more efficient if it creates a new source of risk. It is not scalable if the organization cannot trust the outcome. And it is not finished if no one can explain what happened, who initiated it, and which rules were followed.
Akkadian was built around that reality. Role-based access determines who can perform which work. Standardized workflows apply policy consistently. Detailed audit records preserve attribution. The organization can extend the ability to complete routine work without giving up control of the underlying environment.
One large nonprofit health system shows what that looks like in practice. It supports tens of thousands of team members across hundreds of care locations in several states, with a complex unified communications ecosystem serving clinical and administrative teams across the organization.
With Akkadian, the organization established a standardized, governed model for onboarding, role changes, and offboarding. Its help desk now completes more than 90% of routine joiner, mover, and leaver activity, while the UC team retains control over policies, permissions, workflows, and exceptions.
That is real capacity.
Routine work no longer waits for a senior UC engineer. The help desk can respond quickly and consistently. The UC team can spend more time on architecture, resilience, complex migrations, security, and the exceptions that genuinely require its expertise.
The standards hold regardless of location or who performs the work. The experts define the operating model, the platform enforces it, and the help desk works safely within it.
The goal is not to remove people from work. It is to remove work that never needed a person.
What AI changes, and what it does not
I do not expect today’s boundary between AI, automation, and human work to stay where it is.
General-purpose agents are getting better at reading unfamiliar environments, writing one-off scripts, and handling narrow tasks that once required custom development or purpose-built software. For low-risk, low-consequence work, that will often be enough. Some tasks that platforms handle today will be absorbed this way. That is progress.
What is harder to assemble on demand is everything around the task: who was authorized, which policy applied, the sequence the systems had to follow, what the record needs to show a year later during an audit, and how to reverse the change when something goes wrong.
That is much of the durable value large, regulated organizations are buying. It also becomes more important as the number of platforms, teams, policies, and dependencies grows.
So the useful question is not whether AI will change collaboration operations. It will.
The question is which parts of the operation are narrow tasks and which are governed processes. They should be evaluated differently, and confusing them is expensive in both directions.
That distinction is also guiding how we approach AI in the Akkadian Platform. We will use it where it can simplify the experience, surface better insight, accelerate an outcome, or help customers manage complexity more effectively. Where governed automation already solves the problem reliably, customers should benefit from what is proven today.
Before funding an AI initiative for collaboration operations, it is worth knowing what is already solved.
In this domain, much of the promised outcome is available now: more capacity, less manual effort, better use of expertise, and governance that holds as the enterprise scales.
The technology may be new. The outcome is not.
Find out what is already solved.
A short working session to map which parts of your collaboration lifecycle are narrow tasks, which are governed processes, and where capacity is available today.
