DoLesson 2 of 6The argument

Modeling the ROI with your own numbers

Build a defensible model using your organization's actual figures, and know which inputs the result is most sensitive to.

40 minIntermediate

In this lesson — 3 parts

Build it yourself
Model capacity, and say where it goes
Test what the answer depends on
Do this in the product
Key questions
What to take away

Part 1 of 3

Build it yourself

The ROI page gives you the model and the structure. What makes it usable in your organization is that the inputs are yours, which is not a formality, because the first question you will be asked is about an input rather than the conclusion. "Where does the four hours a week come from?" is answerable if you measured it and fatal if you inherited it from a vendor deck.

So source what you can. Time spent on the work actually in scope, taken from a sample of real weeks rather than an impression. Your own fully-loaded cost per role. Volumes pulled from your systems. Where you genuinely have to estimate, write down that it is an estimate and what it is based on, a model with one honestly-labeled assumption is far more persuasive than one that presents every figure with equal confidence.

Part 2 of 3

Model capacity, and say where it goes

Following the module's framing, model recovered capacity rather than removed cost. That is harder in a good way: recovered hours are only worth something if you can name what they get spent on, and being forced to name it is what turns the model from an arithmetic exercise into a plan.

Be concrete. More accounts covered per rep, more time in front of customers, a segment that currently gets no coverage at all, renewals started a quarter earlier. Each of those connects to a number the business already tracks, which means the ROI case can be checked against reality later, and a case that can be checked is one people will believe now.

Recovered hours, sourced from measured time rather than assumed
Where those hours go, named specifically
The existing business metric each destination should move

Agent run accounting in RevTech, by workflow and team. — Credit usage broken down by workflow, team and individual run, giving agentic work the same cost visibility as any other system. Screenshot of the RevTech application; sample data.

Part 3 of 3

Test what the answer depends on

Before presenting anything, move each input by twenty-five per cent in both directions and watch what happens to the result. Two things come out of this. You learn which one or two inputs the conclusion actually rests on (and it is usually not the ones that generated the most discussion) and you get the range rather than the point estimate.

Present the range. A single number invites the room to argue about that number, and the argument is unwinnable because it is precise about something inherently uncertain. A range with the sensitivity stated invites a conversation about the assumption that matters, which is the conversation worth having and the one where you have already done the work.

Now open the tool

The RevTech ROI model

This lesson is the orientation. The material itself lives on its own page, kept in one place so it stays current everywhere it is referenced.

Do this in the product

Build and stress the model

Open the ROI model and list every input it needs.
For each, find a real figure from your own systems or a measured sample. Mark the ones that are genuinely estimates.
Model recovered capacity rather than removed cost.
For each block of recovered capacity, name the work it gets spent on and the existing metric that should move.
Move each input up and down by twenty-five per cent and record the effect on the result.
Identify the one or two inputs the conclusion actually depends on.
Write the case as a range with those sensitivities stated, not as a single figure.

Key questions

You should be able to answer each of these from memory before opening it. Recalling the answer is what makes it stick; recognizing it when you read it does not.

Why does it matter that the inputs are yours rather than a vendor's?

The first question in the room is always about an input, not the conclusion. "Where does the four hours a week come from?" is answerable if you measured it and fatal if you inherited it from a deck.

Covered in full: The RevTech ROI model →

Why present a range rather than a single number?

A single number invites an unwinnable argument about precision on something inherently uncertain. A range with its sensitivities stated moves the conversation to the assumption that actually matters.

What to take away

Model it on your own numbers. A vendor's benchmark is not an argument you can defend.
Say where the recovered capacity went, or the saving is theoretical.

Teach this lesson

The argument in 4 slides, for presenting it to your team

Slide 1 of 4

A model you did not build is one you cannot defend

The first question in the room is always about an input. Vendor numbers cannot answer it; yours can.

Slide 2 of 4

Use figures you can source

Time your team actually spends on the work in scope
Your own fully-loaded cost, not an industry average
Volumes from your systems, not estimates

Slide 3 of 4

Model capacity, not cost removal

Recovered hours are only worth something if you can say where they go. Name the work they get spent on.

Slide 4 of 4

Test the sensitivity

Move each input by a quarter and see what breaks. Present the range you believe, not the single best figure.

The argument you will have to defend

Why week one is noisier than week four

All lessons Start working with agents

Demo

Try the demo.

See agents carry the repeatable work of GTM across sales, marketing, customer success, and RevOps. Every action prepared, reviewed, and recorded. Fictional data, real product.

Explore the demo