Are you using an outdated model for measuring ROI on AI investments? The risk calculus that technology leaders rely on was built for a different era, one of long implementation cycles, high switching costs and tools that needed to earn their place over years to justify the investment.
Those assumptions don’t make sense anymore. Organizations that haven’t figured that out are leaving real value on the table every day they wait.
The economics of technology implementation have changed so significantly that the old decision-making models are now a liability. Instead of asking yourself if a solution will still be relevant in three years, you can now ask if that same tool will provide real-world value in under six months. Time to value being measured in weeks, not months, changes everything about technology planning for years.
How We Got Here
For most of the history of enterprise technology, caution was the right instinct. Implementing a new system was expensive, slow and disruptive. Migrations were painful. Training took months. Vendor lock-in was real. If you made the wrong call, you lived with it for a long time. Tech debt wasn’t a buzzword—it was an actual organizational weight that accumulated from decisions made under pressure with incomplete information.
In that environment, a rigorous evaluation process made sense. Long procurement cycles, detailed RFPs, phased rollouts and extended ROI windows were rational responses to real risk. The cost of being wrong was high enough that the cost of moving slowly was worth paying.
That calculus worked until the variables changed. And they’ve changed dramatically.
Here’s What’s Different Now
The timeline from idea to working implementation has collapsed. What used to take a dedicated development team six months can now be prototyped in days and deployed in weeks. The cost curve has dropped just as sharply. Solutions that would have required significant capital investment two years ago are now accessible at a fraction of the price with far less infrastructure required to support them.
This isn’t just about AI tools being cheaper or faster to set up. It’s aboaut what that shift means for how you evaluate them. When implementation cost is low and time to value is short, the traditional ROI model breaks down. You’re no longer making a multi-year bet. You’re making a near-term decision about whether a specific capability is worth the cost of having it right now.
The ERP Problem
Here’s an example I use with clients because it makes the shift concrete.
A company is 18 months out from an ERP go-live, and it has a real operational problem today. There’s a meaningful AI application that would address it directly, generate measurable ROI quickly and be deployed in a matter of weeks. The instinct, shaped by years of traditional technology planning, is to wait. Build it into the ERP. Do it right. Don’t create something that will just need to be replaced.
That instinct is costing them 18 months of value.
If the solution can be deployed quickly, delivers real ROI during its window and gets retired cleanly when the ERP goes live, that’s not waste or a failed project. That’s exactly how smart capital allocation should work in a world where implementation cost and cycle time are no longer the constraints they used to be.
We saw this firsthand in our own work at Impact. For our unified support operations team, the initial development of an AI gatekeeper bot took roughly two weeks and generated positive ROI within days of launch. We used that solution while our Professional Services Automation tool was still under development, and it continued to return value throughout the 12 months it took to build the longer-term system.
The fact that something is temporary doesn’t necessarily make it wrong. In many cases, it might be the right fit. The organization solved a real problem, captured real value and moved on. That’s a win by any honest measure.
Rethinking Tech Debt
Tech debt, as most organizations understand it, is the accumulated cost of short-term decisions that create long-term problems. You cut a corner to ship faster and spend the next two years paying for it. You built on a platform that didn’t scale, and now you’re locked in. The debt compounds and eventually comes due.
The concept of tech debt assumed high switching costs and that the wrong decision would follow you. But the math changes when implementation costs are low and the tools themselves are evolving fast enough that yesterday’s deployment may already be approaching its end of life. A solution that lives in your organization for 10 months, solves a real problem and gets retired doesn’t leave debt behind—it leaves results.
The risk, instead, comes from not building something temporary. The risk is waiting for conditions that may never arrive or building something permanent around assumptions that the market has already moved past. We need to challenge organizations on what tech debt really means in this environment before leaders use it as a reason to avoid decisions that should get made.
A Different Planning Model
If we accept that all this is true, let’s explore what an updated ROI model looks like.
First, ROI windows need to be shorter. If a solution can’t demonstrate meaningful value in 90 days, that’s useful information. It doesn’t mean the solution is wrong, but it means the evaluation should reflect that timeline rather than projecting value over three years and hoping it materializes.
Second, organizations need permission to build things that aren’t meant to last. Not every AI implementation should be engineered for permanence. Some should be engineered for the next six months with a clear understanding of when and why they’ll be retired. That’s a legitimate strategy, not a compromise.
Ultimately, leaders need to ask different questions up front. Not just about what the cost will be and what they’ll get in return but also about what the cost is of not having that capability for the next year while they wait to do it properly. In a faster environment, that cost is real, and it belongs in the model.
The Question Worth Asking
The organizations that are winning right now aren’t necessarily the ones with the most sophisticated AI deployments. A lot of them are simply the ones where their leaders gave permission to move. They looked at a problem, identified a solution that could address it quickly, ran the math on a shorter timeline and made the call.
The ones still waiting are often looking for a level of certainty the current environment won’t provide. They’re asking whether a tool will still be relevant in three years when the right question is about what it’s worth will be in the next two quarters.
That’s not recklessness—that’s how the math actually works now.
The challenge I regularly put to any organizational leader is straightforward.
- Look at the problems in your organization that AI could address today
- Ask what it would cost to deploy something in the next 60 days
- Ask what the value of solving that problem would be over the next year, even if the solution doesn’t last beyond that
- Ask whether the old planning model you’re using actually fits those numbers or whether you’re applying a framework built for a different era to a decision that has changed
The ROI is there. The timeline to capture it has shortened. The question is whether your planning process has kept up.
