Stop Getting in Your Own Way: AI and the Basics of Organizational Adaptability [Part 2 of 3]

Topic 2: Failing to inspect the results of your work.

How many times has your organization spent bookoos* of money on a project or initiative that promised to generate savings, increase revenue or produce some other kind of value only to never go back and check if the investment paid off? 
 
I got my career going early with a start-up while I was in college. Starting then, I began interacting with large organizations – universities, national non-profits, Fortune 500 organizations, and multi-national conglomerates. I had made up a story that “real companies” have it figured out. Reality check – they don’t. I’m constantly astonished by how successful companies are in spite of themselves. Relating this to the topic at hand, only a handful of the organizations I’ve encountered have had the discipline to truly follow-through and check the return on investment or value generated from their initiatives. It’s disappointing on so many levels. 
 
A hallmark of being an adaptable organization is the intentionality of adjustments. I advocate for a hypothesis driven approach to this. Meaning, We believe if we -do some thing- that -this other thing- will happen. And if we measure -these things- then we will know we were right if -these results- occur. But let’s get real, it can be easy to develop the discipline of stating the hypothesis upfront. It is an entirely different thing to follow-through and measure to see if you were right. And it doesn’t actually matter if you “were right” – what matters is measuring it and learning from what happened so you can have more evidence to use in making your next hypothesis / adjustment. 
 
Why do organizations rarely follow through on the measurements to check return on investment? My experience indicates it is because there is simply too much other ‘more important’ work going on. I’d argue there is no work more important than evaluating the results of your work so you can intentionally make the next set of changes. Without this discipline you don’t have organizational adaptability, you have organizational adhocedness** (yes, that should be a word).

Having the right systems and processes in place for evaluating value / ROI definitely helps, but now that AI is readily available – it’s a game changer.

For some perspective on how it can work, let’s turn to guidance from my colleague Bryce Arii:

Start with the value, not the technology. The point of agentic AI is never the agent. The point is a business capability that runs better and a workflow that costs less to operate. So, state the hypothesis in those terms. We believe improving this capability lowers this cost or frees up this capacity, and we'll know by measuring these numbers before and after. If you can't name the capability and the cost it touches, you don't have an AI initiative, you have a science experiment. 

Then do the unglamorous work before the agent ever runs. This is where most of the value is won or lost, and it has nothing to do with the model. Build a pilot around a real business case, one capability with a number attached. Get the process documented and stable enough to automate, because you can't improve a workflow you haven't defined. Get the data in order so the agent has something trustworthy to act on and you have a clean baseline to measure against. Get the systems connected so the work and the measurement live in the same place. Skip these and you're not running a pilot, you're running a demo. 

Then measure the value, not the activity. With the baseline captured and the pilot scoped to a single business case, the after-action gets simple. Did the capability improve. Did the cost move. Did the capacity get freed and then actually redeployed. This is where AI earns its keep on the inspection side too, pulling the cycle time, the labor hours, the error rate, the dollars touched, and drafting the before-and-after against the hypothesis you stated up front. The point isn't proving the agent ran. It's proving the business case was real. 

And keep the value framework honest. Cost savings only count if they show up somewhere a CFO can find them. Reduced hours that never leave the payroll, capacity that never gets redeployed, a faster cycle nobody monetizes. Those are notional savings, not real ones. A pilot that improves a number on a slide but never moves the P&L didn't fail at AI. It failed at the business case, which was the whole point. 

Mastering organizational adaptability requires dozens of organizational capabilities to be running at peak performance. The basics of achieving high levels of adaptability means you need to have the discipline to inspect and adapt. That means continually checking in on the results of your work. 

In the final article from this series you’ll learn about minimizing non-value add repetitive or manual tasks to create capacity for high-value knowledge work. 

This series highlights only a few of the many things that can unlock a greater level of adaptability in your organization, and they happen to be 3 that can be directly addressed by using AI. 

This article is brought to you by me, Leslie Porter, and my collaborator Bryce Arii. Bryce has been a long-time team member and collaborator with Adaptivity and I can truly say this series wouldn’t have happened without him. 

If you’re curious how Adaptivity can support your journey to leverage AI to maximize organization adaptability, we’d love to have a chat. Start a conversation today. 

* Bookoos: A slang term derived from the French word beaucoup, meaning a lot, much, or great in quantity. 

** Adhocedness: Having the state, condition or quality of being ad hoc.