AI Doomsday? Or Does Waiting Become the Risk?

Executive Summary

Risk of Delay: For end-user AV and IT collaboration technology managers, the primary strategic risk of artificial intelligence (AI) is inaction rather than early adoption. Delaying practical AI deployment incurs compounding opportunity costs in organizational learning and workflow transformation.

Focus on Practical Friction, Not AI Hype: Technology managers must separate speculative existential AI debates from immediate workplace challenges, applying available AI capabilities to eliminate friction in collaboration, room scheduling, and user support.

Redesign Workflows for True Transformation: Drawing on the electricity analogy, using AI merely to speed up legacy processes yields incremental efficiency, whereas redesigning workflows around AI capabilities drives fundamental organizational transformation.

Manage the Hidden Risk of Inaction: While the risks of taking action (e.g., product selection error, user rejection) are visible, the risk of waiting is invisible and compounding—costing organizations unaccumulated learning and uncaptured adaptability.

Adopt a Software-Era Foundation Strategy: Organizations must move past static hardware-era purchasing mindsets and choose adaptable cloud and software foundations today that allow continuous evolution alongside maturing AI roadmaps.

It is almost impossible in recent weeks in particular, to have a targeted conversation about a specific AI application without the baggage of much bigger conversations coming into play. Jobs. Human relevance. Safety. Societal disruption. And most recently, even existential risk to humanity itself. That’s before we even begin considering the potentially extraordinary positive capabilities on the horizon.

All those conversations, particularly the difficult ones, matter. We should all have a thoughtful voice in them.

But positive or negative, they largely concern futures that most of us cannot reliably define, let alone individually control or deliver. My worry in this moment is that we cannot allow the scale and emotion of those debates to bleed into the much smaller, but far more practical decisions about what we individually CAN actually pursue and influence with the AI capabilities already in our hands today.

Applied to an end-user AV technology manager, I think that distinction matters enormously.

Deliver Organizational Impact and Outcomes

Your immediate challenge is likely not to resolve the future of artificial intelligence. It is to improve collaboration, remove friction, support users, make better use of space and help people decide, learn and innovate more effectively. Ultimately, it is to deliver the organizational impact and outcomes that collaboration technology has always foundationally existed to serve.

So perhaps, for a moment, we should stop talking about “AI” at all.

Rather, let’s try to just discuss what work we are trying to improve. What frictions are we trying to remove? What valuable capabilities have we historically wanted, but could never justify because some traditional form of constraint made it unviable or impractical?

Only then might we ask whether any of that equation might have changed in an AI age.

Many of us have heard the analogy of AI as the new electricity, and I can’t help but keep coming back to it.

People did not transform their lives because they became enthusiastic adopters of electricity as an abstract and somewhat universally applicable technology. They adopted lighting, refrigeration, appliances, machinery and countless other applications that made useful things newly possible. In industry, simply replacing an old power source with electricity created some value. But the larger transformation came when organizations redesigned factories and workflows around what distributed electric power now enabled.

I think there is a useful lesson in that for us.

Transforming the Workflow with AI

Using AI to make yesterday's workflow faster can absolutely create value. Redesigning the workflow around what AI makes newly possible is where transformation begins.

Consider something as everyday as organizing a meeting.

Particularly in the post-COVID hybrid workplace, we have placed even more logistical work onto a meeting organizer simply trying to bring peers together in person. Having consciously and thoughtfully considered who needs to attend and why, there is huge friction in working through where people will physically be, what room is available, whether it offers the right capability, etc.—let alone what happens when things change in the interim.

Executives typically have EAs to absorb much of that effort and friction. Most of the rest of us do not. So I’m excited as platforms such as Microsoft Places are beginning to address pieces of that problem through more intelligent workforce, workplace, and collaborative workflow coordination. At first glance, the business case is obvious: save somebody ten, twenty, perhaps thirty minutes arranging and managing the logistics of a meeting.

Useful? Absolutely. And that’s primarily the focus of implementation ROI analysis at present.

But I think there is a more important opportunity that sits underneath that efficiency premise.

Every new workflow carries an adoption threshold. The value people believe they will receive has to outweigh not only the practical change effort required, but the cognitive and emotional effort of doing something differently. It’s the “Crossing the Chasm” proposition we all know. If the friction is too high, even objectively better experiences struggle to gain adoption.

Our industry has spent years responding to that reality through simplification and standardization. We reduced choices. Simplified interfaces. Removed capabilities users struggled to understand. We lowered the friction floor largely by reducing the complexity we exposed, often designing toward a lowest common denominator. A significant compromise—even if justifiable—to supporting the diversity premise that gives collaboration so much of its value.

But in this moment, I believe AI potentially gives us a different mechanism.

Intelligence Absorbs Complexity

Rather than removing capability to make an experience manageable, what if intelligence can increasingly absorb complexity on the user's behalf?

Now the meeting example becomes much more interesting.

Yes, an intelligent agent might find the time, the room and the appropriate technical capability. But what if – building upon that new more practical and tactical efficiency foundation - it can also remove enough friction from richer ways of collaborating? That truly transformative new workflow opportunities now themselves become more practical? Premises like identifying and proposing:

• Who should actually be in the room?
• What perspective or experience might be missing?
• What collaborative methodology best fits the intended outcome?
• How should the space or technology respond differently if the purpose is ideation rather than decision-making?
• And what needs to happen to turn conversation into immediate action?

We have always been capable of asking many of those questions. The problem is these are complex and difficult to manage and measure. Applying that level of thought, coordination and facilitation to every meeting would historically have required far more human effort and practical friction than the return could justify.

So most meetings never received it.

That is the equation AI can begin to change.

The transformation opportunity is not simply doing the same work with less effort. It is making valuable new work viable that the old effort equation never practically allowed us to do consistently or at scale.

Efficiency still matters. In fact, I think it may be foundational. But perhaps its greatest value is not the minutes we recover, but in the new foundations it unlocks.

Efficiency Removes Constraints. Transformation Exploits Their Absence.

If collaboration itself, rather than simply collaboration technology, is becoming an increasingly critical driver of organizational innovation and change, that distinction matters enormously. Better access to knowledge, more relevant perspectives, stronger conversations and better decisions can lead to faster learning, greater innovation, greater adaptability and ultimately competitive advantage.

Now the return has the potential to move beyond linear efficiency toward something much more compounding.

[ Check out other insightful collaboration and workplace AI articles from Byron Tarry here. ]

And that is precisely why the risk of “waiting” for bigger AI discussions to progress deserves more attention.

Here’s a different take on that concept. Our organizations and careers have largely been built around managing the visible risk of action. What if I choose the wrong product? What if the platform does not deliver? What if users reject it? What if something better arrives next year?

Those are legitimate questions.

The Risk of Doing Nothing is Always Harder to See.

There is rarely an invoice for value never discovered because we waited. No budget variance captures the capability our teams failed to develop. No project review measures learning we never accumulated, or a collaboration improvement that never had a chance to emerge.

That has always made inertia deceptively comfortable.

While we have spent decades optimising how we manage the risk of action, I think we now need to become much better at considering & managing the risk of inaction.

That does not mean becoming reckless. It means recognizing that risk management can no longer be nearly so synonymous with change avoidance.

Certainly amid an abundance of opportunity sometimes waiting will absolutely be the right answer. Some technologies will be immature. Some risks will be unacceptable. Some opportunities are simply misaligned. Some organizations simply will not yet be ready.

But waiting is not neutral.

If the opportunity we are deferring is merely a small efficiency gain, delay may carry a relatively small cost. But if that efficiency is the first unlock in a chain that enables a better workflow, better collaboration, and ultimately better organizational performance, then the opportunity cost compounds very differently.

And I believe that is where we increasingly sit as an industry.

I recently had a conversation with Sam Kennedy from NetSpeek about the strange relationship our industry has historically had with roadmaps.

In a hardware-centric world, showing customers too much of what was coming next could often create a reason not to buy what existed today. If tomorrow's product offered materially different capability, while today's hardware was largely static and might remain installed for five or seven years, waiting could be entirely rational. Roadmap conversations themselves therefore carried risk.

But I wonder whether we are carrying hardware-era buying instincts into a software-era capability model.

It’s clear that increasingly more of the future will arrive through agile software, cloud platforms, intelligence layers and services layered onto physical foundations already deployed.

That must therefore change the question. Rather than asking "What might I miss if I commit today?", perhaps we should increasingly ask, "What can I choose today to create meaningful value now while also maintaining a credible path toward what comes next?"

In This Moment of Uncertainty, You Cannot Buy the Future.

But you can choose foundations today that give you permission and confidence in a path to evolve into it.

That mindset should influence how we interrogate products, architectures, and partners. Clarity of roadmap should now be seen as an opportunity on all sides to increase confidence in today's decisions, as opposed to exposing a dead end.

This is not about predicting perfectly. It is about maintaining progression when perfect prediction is unavailable.

And that brings me back to where we started.

The current AI conversation is full of questions nobody can answer with confidence. But we cannot allow uncertainty about that future to blind us to the opportunities sitting within reach today, to the leadership we can show, the foundations we can build, and the transformational outcomes we can already begin to pursue.

This does not require us to be optimistic about every possible AI future. Nor does it require us to dismiss its risks. I think it just requires something much more practical: the discipline to separate what is speculative from what is actionable, and to focus our energy where we still have agency.

Today we CAN remove unnecessary friction. We CAN rethink workflows designed around old constraints. We CAN make decisions that preserve a path to tomorrow. And we CAN challenge ourselves as to whether fear of an unknowable future is becoming a misguided justification for preserving a status quo we already know is imperfect.

Because in this moment, standing still is also a bet.