4
The Enterprise Guide to
Tracking
AI ROI
This is the fourth and final installment of our AI Deployment Playbook series.
You can start from the beginning here.
TRACK PREP
Track adoption and Build your AI leadership team,
measure success measure baseline, and find
your AI champions
THE AI
PROFICIENCY
CYCLE
APPLY TRAIN
Audit existing workflows to Train the team on safe AI
find AI use cases capabilities and company-
specific policies
The Enterprise Deployment Series #4 The Enterprise Guide to Tracking AI ROI | 1
A letter from
Section’s CEO
Source: Deloitte’s State of
Generative AI in the Enterprise
Q4 Report, January 2025.
Companies are reporting higher and higher AI spend, and citing both qualitative and
quantitative wins from their AI initiatives. And yet 41% of organizations also report
struggling to define and measure the impacts of their generative AI efforts. So which
is it – is there value or not?
The reality is that some of this is PR, as companies are vying to look like leaders in the
new age of AI. And some of it’s real – at Section, we get a 10% productivity lift just from
having our team use LLMs in their daily work. We can see it and we can measure it.
The key to tracking the ROI of your own AI initiatives is to tune out the hype and zero
in on the metrics that matter to your company. Let’s cover the two pronged approach
for doing that.
GREG SHOVE
CEO of Section
The Enterprise Deployment Series #4 The Enterprise Guide to Tracking AI ROI | 2
The two-pronged approach to
measuring AI ROI
In the previous installment, we broke out potential AI projects into two distinct types – personal
wins & workflow wins. Both of these come with different investments and different levels of
impact, so you also need to think about their success differently.
PERSONAL WINS WORKFLOW WINS
Your personal wins are employee-level Your workflow wins are company-level use
use cases that typically cover individual cases and are used by a lot of people – or
or small team workflows and are solved by only a few, but with a high cost savings
via off-the-shelf LLMs or AI tools.
You probably spent additional budget on
You likely didn’t spend any additional AI solutions to enhance the efficiency of
money on these beyond your company- these workflows, so you’ll need to justify
wide LLM, so these wins will help justify this spend.
your LLM investment.
The goal of these use cases is to
The goal of these use cases is typically an remove friction or streamline costs on a
increase in efficiency or quality of output highly valuable use case to the company
for a small group of AI power users. as a whole.
Measuring the impact of workflow wins
For higher-value, higher-investment projects, measure ROI – your time/cost savings relative to
the budget you spent.
Here are 3 steps to doing this well:
1. Establish a baseline. We often don’t do this when we update internal processes, but you
need to know how long a workflow takes when done manually and how much it costs to
accomplish without AI so you can compare.
2. Run a 90 day sprint to verify that the AI application works. Determine that AI can indeed
produce an output equal to or better than a human.
3. Compare the results. Using your established baseline, measure the ROI of the AI solution
based on how much faster or cheaper it can achieve the workflow.
AI ROI = (Baseline Cost/Time – Cost of AI/Time) ÷ AI Investment
To make a bulletproof case, measure your initial results after this 90 day sprint and again in 6 months.
The Enterprise Deployment Series #4 The Enterprise Guide to Tracking AI ROI | 3
Measuring the impact of personal wins
For lower-impact, lower-investment AI projects, you need a more qualitative approach to
measure success. These AI efforts are owned and used by individuals across the org, so trying
to unify them around a single metric won’t work.
You should encourage your team to use the same system as you use for workflow wins, but at
the individual level:
1. Establish a baseline of time/effort to accomplish a task without AI
2. Run a 90-day sprint to verify the AI application works
3. Measure the time/effort and therefore ROI of the AI solution
Note: You should plan for a margin of error or inflation knowing that employees are self-reporting
their wins.
Here’s what your plan should be for measuring the impact of your personal AI wins over the first
30 days, 12 months, and 24 months:
Eventually, the impacts of these personal wins become more obvious as you’re able to map
them to longer-term trends – such as being able to increase output on a team without having
to increase headcount.
The Enterprise Deployment Series #4 The Enterprise Guide to Tracking AI ROI | 4
The long-term ROI
While you measure the impact of your individual AI initiatives, don’t lose sight of the long-term
measure of success: Building an AI-native team.
All of this effort should ladder up to creating a team that can consistently identify big and small
opportunities for AI going forward. And to do this, they need to be knowledgeable about how AI
works and skilled at putting it into practice.
FROM TO
Unknown AI adoption 80% of your organization is an AI
and proficiency Expert, Practitioner or Experimenter
The measure of success here is 80% of your organization in one of the 3 highest AI proficiency
levels within 12-36 months.
Lean into the frameworks you learned in this 4-part series – focus on building AI-readiness into
your deployment strategy, providing high-quality on-going training, and making AI applicable to
your employees via core workflows – and you can build this high-output, high-performance, AI-
native team.
The Enterprise Deployment Series #4 The Enterprise Guide to Tracking AI ROI | 5
What’s next?
You’ve reached the end of the AI Deployment Playbook series! Next is implementation, testing,
and continual iteration.
But this isn’t the end of Section’s support. Here’s where to go next if need you additional
resources:
Learn more about how Section works with companies on their AI goals. Whether
you need help benchmarking your team’s AI ability, redesigning core workflows
1 with AI, or training and upskilling your team on brand-safe AI, we have proven
solutions. Reach out to learn more.
Check out the rest of our playbooks & tools. From our flagship AI Proficiency
2 Report, to a custom GPT creation workbook, to an AI policy creation guide, we
have free resources to supplement every stage of your AI deployment.
Join a free live AI:ROI conversation. We host free webinars with trailblazing AI
3 leaders every month for impactful, no-BS conversations on how to put AI to work
and get real returns from it.
We’re happy to chat about your AI
deployment any time
BOOK A MEETING
The Enterprise Deployment Series #4 The Enterprise Guide to Tracking AI ROI | 6