Buying the technology is the easy part
Enterprise technology doesn't arrive in a vacuum. And for all the promises of automation and efficiency, the surrounding organization doesn't magically change when a new tool gets switched on.
Jasmine Libert
Senior VP Growth and GTM
"I like to think about this stuff in my free time. "
AI adoption is accelerating at a crazy pace, but turning that investment into organizational capability takes considerably more work. A little while ago, I wrote about the durable investments organizations can make as they figure out where AI belongs in their business. My argument was fairly simple: we don't know exactly where AI is going, but investments in good data, measurement, governance and organizational understanding will remain useful across a lot of possible futures.
However, there's another part of AI readiness that I've been thinking about since. What happens after you buy the technology? Because, increasingly, access isn't the problem.
Recent research from Bain found that 95% of marketing organizations are using AI. Only 6% said it was having a significant impact on performance. McKinsey found something similar: nearly 60% of marketers use AI multiple times a week, but fewer than 10% have begun capturing value across end-to-end workflows.
That's a pretty substantial gap between having the technology and using it well. And having spent a lot of time looking under the hood of enterprise MarTech stacks, that gap doesn't surprise me.
We’ve seen this movie before
The technology may be new, but the challenges of getting organizations to use technology effectively are very familiar.
Enterprise technology doesn't arrive in a vacuum. It lands in an environment with existing platforms, integrations, data, processes, budgets, approval structures, and people who already have quite a lot on their plates (can you tell that I’m speaking from experience here?!).
And for all the promises of automation and efficiency, the surrounding organization doesn't magically change when a new tool gets switched on. Things actually get harder before they get easier, at least in the short-term.
There's some interesting research from Vertice that puts the broader problem into perspective. In Q2 2026, 65% of SaaS licences were either unused or underutilized. For enterprises with more than 10,000 employees, Vertice estimates the annual cost of that unused and underused software at $18.9 million (brb, I’m going to start my new lucrative career in SaaS sales…).
That's a ton of money to spend on technology you're not getting full value from.
Now, to be fair, underutilized doesn't always mean wasted. Organizations buy capacity in anticipation of growth, and software can deliver value even when every licence isn't being used every day, but that’s not really what we’re talking about here. I suspect anyone who's spent time working with enterprise technology has encountered at least a few platforms that were purchased with great enthusiasm and subsequently used for a fraction of what they were supposed to do.
It's a remarkably common situation. A business identifies a problem, buys a platform to solve it, implements the most immediate requirements and moves on (rinse and repeat). Meanwhile, the platform continues to evolve, the people who originally championed it move into different roles, and a few years later somebody proposes buying something new to solve a problem the existing technology could probably handle.
I've seen versions of this more times than I can count.
AI isn’t going to magically fix that
This is where I think some conversations about AI ROI become too focused on the technology itself. Let's say an AI tool can reduce a task that previously took four hours to twenty minutes. Fantastic. That's a meaningful productivity improvement, and there are plenty of reasons to be excited about it.
But what happens to those twenty minutes of work once they're finished?
Does the output still need to go through three rounds of approvals? Does someone need to manually move it into another system? Is the underlying data reliable? Does anyone know enough about the subject to recognize when the AI has produced something completely wrong but very confidently?
These things matter.
And they matter even more as we move beyond using AI to help individuals complete tasks and start expecting it to support analysis, make recommendations, personalize customer experiences and eventually take actions across multiple systems.
That's a much bigger ambition than helping someone write an email faster.
McKinsey found that only 28% of marketers surveyed said their organizations were fundamentally redesigning teams and workflows around AI. Most were adding AI to processes that already existed.
Which, to be clear, can be a perfectly reasonable place to start. I'm a big believer in experimentation, and there's no reason to redesign an entire organization every time a software vendor announces a new feature (thankfully!).
But if we're expecting AI to fundamentally improve business performance, we need to look at how the rest of the organization operates. The tech can only take us so far.
Look at the tools you already have before you buy anything new
I genuinely wish that more organizations would do this.
Enterprise platforms have become incredibly sophisticated and they evolve quickly. Features get added, integrations improve, capabilities expand. Meanwhile, internal teams change, documentation gets outdated, and institutional knowledge can walk out the door when people leave.
It's entirely possible for an organization to own technology that can solve a problem and have nobody internally who knows it can do that. There's also a difference between knowing a feature exists and having the skills, data and processes to use it effectively.
So when a business identifies a new requirement, I think there's enormous value in looking at the existing environment first. What capabilities are already available? Are they configured properly? Are people trained to use them? Does the data support what they're trying to accomplish? Is there someone responsible for making sure the implementation continues to meet the needs of the business?
Sometimes the answer really is to buy something new (just don’t put all your trust in the SaaS sales person!). But sometimes the opportunity is sitting in a platform you already pay for.
Ditto for data
I think there's a similar issue with how organizations approach data.
Storage has become relatively inexpensive, and we've spent years building systems capable of collecting enormous amounts of information. Which is wonderful… in theory. It has also made it very easy to collect data simply because we can.
I've written before about data hoarding (that’s the technical term), and I think it's worth revisiting in the context of AI. Organizations accumulate information across platforms, departments and customer touchpoints, often without a particularly clear idea of how it will be used. The argument is usually that the data might be valuable someday.
Maybe! But keeping it has costs beyond storage.
Someone has to maintain the pipelines, manage access, document definitions, monitor quality, respect privacy and consent requirements, and hopefully figure out what all that information means. And as anyone who's worked with a large data environment knows, those responsibilities have a way of multiplying.
There's an opportunity cost, too. Every dataset, integration and platform requires some amount of attention, expertise and maintenance. Those resources could be going toward improving the information that the organization already relies on to understand customers and make decisions.
I'm certainly not suggesting we stop collecting data. But I do think we should be more deliberate about what we're collecting, why we're collecting it, and whether we have the capacity to make it useful.
AI makes this conversation so much more relevant. The quality, context and accessibility of data are increasingly important as we ask technology to do more of the analysis and interpretation for us, but collecting more information doesn't automatically make an organization better informed. We've had plenty of time to learn that lesson.
Technology has an operating cost beyond the invoice
When organizations evaluate technology investments, the conversation tends to focus on licensing, implementation and infrastructure costs. Those are important, obviously, but they're only part of what it takes to make technology useful. There's also the cost of developing the organizational capability to use it.
People need training. Processes may need to change. Integrations need to be maintained. Data needs to be reliable and accessible. Governance needs to reflect what the technology is being used for, and someone needs to understand the business well enough to recognize whether it's delivering the intended results. And all of that needs to continue long after the initial implementation is finished.
I think this is especially important when evaluating AI investments, because the pressure to adopt new capabilities is enormous. It's easy to get excited about what a product can demonstrate, it's considerably harder to assess what it would take to make that same capability work consistently inside a real organization, with all the constraints that come with it.
A sophisticated platform might be capable of extraordinary things. But if an organization doesn't have the skills, data, processes or resources to operate it effectively, much of that potential may remain just that: potential.
A simpler solution that people understand and use well may deliver far more value. And that brings me back to readiness.
Readiness includes the organization around the technology
I still believe strongly in the value of investing in good data foundations. Reliable, connected and well-governed data gives organizations more options, whether they're investing in AI or trying to improve the way they measure and understand their business, but there's another layer to readiness that deserves just as much attention.
Does the organization have the skills to use the technology? Do people understand the data and the processes behind it? Can teams evaluate the output and make decisions based on it? Is there a plan for maintaining and improving the capability as the business changes?
These aren't particularly glamorous questions, they don't make for very exciting product announcements, either. But they're the questions that determine whether an investment delivers value beyond the initial enthusiasm.
And they're questions we spend a lot of time thinking about at Napkyn.
Our work often starts with understanding what an organization is trying to accomplish and what it already has in place. That might lead to a new technology investment, but it might also mean improving an existing implementation, connecting systems, strengthening measurement, simplifying a data architecture or helping a team develop the knowledge to get more out of its existing tools.
We work across strategy, analytics, engineering, marketing effectiveness and enablement because these things are connected in practice. The best technical solution in the world won't accomplish much if the organization can't use it effectively.
I think there's a real opportunity right now for businesses to take a closer look at the investments they've already made, especially as they consider what comes next with AI.
Technology will keep evolving, and there will always be something new worth exploring. That's part of what makes this industry interesting.
Before adding another platform to the stack, it's worth asking whether we're getting everything we should from the ones we already own. You might be surprised by the answer.
Wondering whether you're getting the most from your data and MarTech investments?
At Napkyn, we help organizations understand what they already have, identify where value is being lost, and make informed decisions about what to improve next. Let's talk.
More Insights
Building Scalable Serverless Data Pipelines on Google Cloud
Cem Bakar
Cloud Architect
Oct 7, 2026
Read More
Understanding the IAB TCF v2.3 Update: New Vendor Disclosure Requirements Explained
Rob English
Director, Data Solutions
Jun 17, 2026
Read More
CM360 Path to Conversion Reporting: Full Customer Journey Attribution
Monika Boldak
Associate Director, Marketing
Jun 3, 2026
Read More
More Insights
Sign Up For Our Newsletter
Napkyn Inc.
204-78 George Street, Ottawa, Ontario, K1N 5W1, Canada
Napkyn US
6 East 32nd Street, 9th Floor, New York, NY 10016, USA
212-247-0800 | info@napkyn.com
Buying the technology is the easy part
Enterprise technology doesn't arrive in a vacuum. And for all the promises of automation and efficiency, the surrounding organization doesn't magically change when a new tool gets switched on.
Jasmine Libert
Senior VP Growth and GTM
November 10, 2026
"I like to think about this stuff in my free time. "
AI adoption is accelerating at a crazy pace, but turning that investment into organizational capability takes considerably more work. A little while ago, I wrote about the durable investments organizations can make as they figure out where AI belongs in their business. My argument was fairly simple: we don't know exactly where AI is going, but investments in good data, measurement, governance and organizational understanding will remain useful across a lot of possible futures.
However, there's another part of AI readiness that I've been thinking about since. What happens after you buy the technology? Because, increasingly, access isn't the problem.
Recent research from Bain found that 95% of marketing organizations are using AI. Only 6% said it was having a significant impact on performance. McKinsey found something similar: nearly 60% of marketers use AI multiple times a week, but fewer than 10% have begun capturing value across end-to-end workflows.
That's a pretty substantial gap between having the technology and using it well. And having spent a lot of time looking under the hood of enterprise MarTech stacks, that gap doesn't surprise me.
We’ve seen this movie before
The technology may be new, but the challenges of getting organizations to use technology effectively are very familiar.
Enterprise technology doesn't arrive in a vacuum. It lands in an environment with existing platforms, integrations, data, processes, budgets, approval structures, and people who already have quite a lot on their plates (can you tell that I’m speaking from experience here?!).
And for all the promises of automation and efficiency, the surrounding organization doesn't magically change when a new tool gets switched on. Things actually get harder before they get easier, at least in the short-term.
There's some interesting research from Vertice that puts the broader problem into perspective. In Q2 2026, 65% of SaaS licences were either unused or underutilized. For enterprises with more than 10,000 employees, Vertice estimates the annual cost of that unused and underused software at $18.9 million (brb, I’m going to start my new lucrative career in SaaS sales…).
That's a ton of money to spend on technology you're not getting full value from.
Now, to be fair, underutilized doesn't always mean wasted. Organizations buy capacity in anticipation of growth, and software can deliver value even when every licence isn't being used every day, but that’s not really what we’re talking about here. I suspect anyone who's spent time working with enterprise technology has encountered at least a few platforms that were purchased with great enthusiasm and subsequently used for a fraction of what they were supposed to do.
It's a remarkably common situation. A business identifies a problem, buys a platform to solve it, implements the most immediate requirements and moves on (rinse and repeat). Meanwhile, the platform continues to evolve, the people who originally championed it move into different roles, and a few years later somebody proposes buying something new to solve a problem the existing technology could probably handle.
I've seen versions of this more times than I can count.
AI isn’t going to magically fix that
This is where I think some conversations about AI ROI become too focused on the technology itself. Let's say an AI tool can reduce a task that previously took four hours to twenty minutes. Fantastic. That's a meaningful productivity improvement, and there are plenty of reasons to be excited about it.
But what happens to those twenty minutes of work once they're finished?
Does the output still need to go through three rounds of approvals? Does someone need to manually move it into another system? Is the underlying data reliable? Does anyone know enough about the subject to recognize when the AI has produced something completely wrong but very confidently?
These things matter.
And they matter even more as we move beyond using AI to help individuals complete tasks and start expecting it to support analysis, make recommendations, personalize customer experiences and eventually take actions across multiple systems.
That's a much bigger ambition than helping someone write an email faster.
McKinsey found that only 28% of marketers surveyed said their organizations were fundamentally redesigning teams and workflows around AI. Most were adding AI to processes that already existed.
Which, to be clear, can be a perfectly reasonable place to start. I'm a big believer in experimentation, and there's no reason to redesign an entire organization every time a software vendor announces a new feature (thankfully!).
But if we're expecting AI to fundamentally improve business performance, we need to look at how the rest of the organization operates. The tech can only take us so far.
Look at the tools you already have before you buy anything new
I genuinely wish that more organizations would do this.
Enterprise platforms have become incredibly sophisticated and they evolve quickly. Features get added, integrations improve, capabilities expand. Meanwhile, internal teams change, documentation gets outdated, and institutional knowledge can walk out the door when people leave.
It's entirely possible for an organization to own technology that can solve a problem and have nobody internally who knows it can do that. There's also a difference between knowing a feature exists and having the skills, data and processes to use it effectively.
So when a business identifies a new requirement, I think there's enormous value in looking at the existing environment first. What capabilities are already available? Are they configured properly? Are people trained to use them? Does the data support what they're trying to accomplish? Is there someone responsible for making sure the implementation continues to meet the needs of the business?
Sometimes the answer really is to buy something new (just don’t put all your trust in the SaaS sales person!). But sometimes the opportunity is sitting in a platform you already pay for.
Ditto for data
I think there's a similar issue with how organizations approach data.
Storage has become relatively inexpensive, and we've spent years building systems capable of collecting enormous amounts of information. Which is wonderful… in theory. It has also made it very easy to collect data simply because we can.
I've written before about data hoarding (that’s the technical term), and I think it's worth revisiting in the context of AI. Organizations accumulate information across platforms, departments and customer touchpoints, often without a particularly clear idea of how it will be used. The argument is usually that the data might be valuable someday.
Maybe! But keeping it has costs beyond storage.
Someone has to maintain the pipelines, manage access, document definitions, monitor quality, respect privacy and consent requirements, and hopefully figure out what all that information means. And as anyone who's worked with a large data environment knows, those responsibilities have a way of multiplying.
There's an opportunity cost, too. Every dataset, integration and platform requires some amount of attention, expertise and maintenance. Those resources could be going toward improving the information that the organization already relies on to understand customers and make decisions.
I'm certainly not suggesting we stop collecting data. But I do think we should be more deliberate about what we're collecting, why we're collecting it, and whether we have the capacity to make it useful.
AI makes this conversation so much more relevant. The quality, context and accessibility of data are increasingly important as we ask technology to do more of the analysis and interpretation for us, but collecting more information doesn't automatically make an organization better informed. We've had plenty of time to learn that lesson.
Technology has an operating cost beyond the invoice
When organizations evaluate technology investments, the conversation tends to focus on licensing, implementation and infrastructure costs. Those are important, obviously, but they're only part of what it takes to make technology useful. There's also the cost of developing the organizational capability to use it.
People need training. Processes may need to change. Integrations need to be maintained. Data needs to be reliable and accessible. Governance needs to reflect what the technology is being used for, and someone needs to understand the business well enough to recognize whether it's delivering the intended results. And all of that needs to continue long after the initial implementation is finished.
I think this is especially important when evaluating AI investments, because the pressure to adopt new capabilities is enormous. It's easy to get excited about what a product can demonstrate, it's considerably harder to assess what it would take to make that same capability work consistently inside a real organization, with all the constraints that come with it.
A sophisticated platform might be capable of extraordinary things. But if an organization doesn't have the skills, data, processes or resources to operate it effectively, much of that potential may remain just that: potential.
A simpler solution that people understand and use well may deliver far more value. And that brings me back to readiness.
Readiness includes the organization around the technology
I still believe strongly in the value of investing in good data foundations. Reliable, connected and well-governed data gives organizations more options, whether they're investing in AI or trying to improve the way they measure and understand their business, but there's another layer to readiness that deserves just as much attention.
Does the organization have the skills to use the technology? Do people understand the data and the processes behind it? Can teams evaluate the output and make decisions based on it? Is there a plan for maintaining and improving the capability as the business changes?
These aren't particularly glamorous questions, they don't make for very exciting product announcements, either. But they're the questions that determine whether an investment delivers value beyond the initial enthusiasm.
And they're questions we spend a lot of time thinking about at Napkyn.
Our work often starts with understanding what an organization is trying to accomplish and what it already has in place. That might lead to a new technology investment, but it might also mean improving an existing implementation, connecting systems, strengthening measurement, simplifying a data architecture or helping a team develop the knowledge to get more out of its existing tools.
We work across strategy, analytics, engineering, marketing effectiveness and enablement because these things are connected in practice. The best technical solution in the world won't accomplish much if the organization can't use it effectively.
I think there's a real opportunity right now for businesses to take a closer look at the investments they've already made, especially as they consider what comes next with AI.
Technology will keep evolving, and there will always be something new worth exploring. That's part of what makes this industry interesting.
Before adding another platform to the stack, it's worth asking whether we're getting everything we should from the ones we already own. You might be surprised by the answer.
Wondering whether you're getting the most from your data and MarTech investments?
At Napkyn, we help organizations understand what they already have, identify where value is being lost, and make informed decisions about what to improve next. Let's talk.
More Insights
Buying the technology is the easy part
Jasmine Libert
Senior VP Growth and GTM
Nov 10, 2026
Read More
What’s New in Google Analytics: More Room for Curiosity
Chris Wright
Associate Director, Training and Development
Oct 28, 2026
Read More
Building Scalable Serverless Data Pipelines on Google Cloud
Cem Bakar
Cloud Architect
Oct 7, 2026
Read More
More Insights
Sign Up For Our Newsletter