Terenzo

TerenzoROI methodology

How we measure the ROI of AI.

One formula, applied the same way every time: the hours a workflow returns, priced at what those hours actually cost you, set against the full cost of building and maintaining the change. Every input traces to a named interview in your company. Nothing is estimated from a vendor benchmark, and nothing is ever fabricated.

Fig. 01What we count

Three things, all of them measurable.

01

Hours returned

Time a specific person stops spending on a specific task, in hours per week, stated by the person who does the work and confirmed by their manager.

02

Rework avoided

The cost of catching errors late: reconciliations, corrections, credit memos, re-sent invoices. Counted only where the current error rate is known.

03

Cycle time

Days removed from a process that has a dollar attached to waiting: the close, quote-to-cash, collections. Counted at the carrying cost of the delay.

Fig. 02What we refuse to count

If it cannot survive a CFO's questions, it is not on the slide.

  1. Vendor-claimed multipliers

    "Teams report 40% productivity gains" is marketing, not measurement. We measure your workflow or we do not claim the number.

  2. Revenue attribution without a causal path

    If we cannot draw the line from the change to the dollar, the dollar stays off the math. Upside that might happen is commentary, not ROI.

  3. Time too scattered to redeploy

    Saving forty people two minutes a day is a rounding error wearing a big number. We count hours only where they consolidate into capacity someone can actually use.

  4. Adjectives

    Faster, smarter, more efficient: none of these are numbers. If a benefit is real but unquantifiable, we say so in plain words and leave it out of the total.

Fig. 03The math

One formula, shown with its arithmetic.

Annual savings = hours saved per week × people × loaded hourly rate × 52
Year-one ROI = (annual savings − implementation − 12 months of maintenance) ÷ (implementation + 12 months of maintenance)

Worked example, with illustrative numbers, not a client result: three analysts each spend seven hours a week assembling the monthly reporting pack. 3 × 7 × $52 × 52 is $56,784 a year, call it $57,000. If the automation costs $30,000 to build and $1,500 a month to maintain, year-one cost is $48,000 and year-one ROI is 19%. Year two, the same savings against maintenance alone, is over 200%. The audit runs this math on every recommendation, with your figures.

Fig. 04How the numbers stay honest

Four rules, applied to every figure.

  1. Traced to a source

    Every hour count names the interview it came from. The person who does the work stated it; their manager saw it before we did the multiplication.

  2. Validated before the readout

    Your stakeholders walk the findings before we present them. A number nobody in your company recognizes never reaches the deck.

  3. Assumptions on the slide

    The rate, the headcount, and the hours sit next to the total, so the math can be checked in the room, not defended after it.

  4. Remeasured after go-live

    The maintenance retainer includes measuring actuals against the audit's estimate. A projection nobody checks is just a hope with a decimal point.

Fig. 05See it applied

This math has a deck.

The audit ends in a short executive readout: opportunity matrix, phased roadmap, and this ROI math applied to your numbers. A sample, built for a fictional company with the arithmetic shown, is public.

View the sample readout Book a call