Proof, with the gaps left visible

Here is what we can prove today.

We are early. This library contains working demos and modeled business cases, not a hidden roster of customer wins. Every card separates three different questions: did the system run, whose data did it run on, and was the financial outcome verified? A demo can prove that a workflow runs. It cannot prove that a client saved money.

Current evidence inventory

These counts are generated from the published library. They update when an item’s evidence status changes.

Demo-tested systems
4
Modeled outcomes
12
Client-verified outcomes
0
The evidence ladder

Three levels. No promotion without evidence.

A source can support an assumption. A test run can support a system claim. Only production measurement against an agreed baseline can support a client outcome.

01 / model

Modeled scenario

Shows the arithmetic for a plausible process at a stated size. Inputs, residual human work, running costs, and a range should be visible.

Does not prove: that a system ran or that a client achieved the result.

02 / observed

Observed system evidence

Reports what a working build did under named test conditions, including data type, sample size, exceptions, and review workload.

Does not prove: production adoption, attribution, or financial savings.

03 / verified

Client-verified outcome

Requires production use, a written baseline, a matched reporting period, source records, agreed adjustments, and client approval of the published result.

Can support: a measured client outcome, within the stated scope and period.

Savings verification protocol

The number is agreed before the build and re-measured after it.

The signed scope is the measurement plan. It defines what counts, where the source data comes from, and which changes require an adjustment.

  1. 01
    Draw the boundary.

    Name the process, cost lines, source systems, and period being measured. Costs outside that boundary do not enter the result.

  2. 02
    Record the baseline.

    Use exports, invoices, time records, or another agreed source. Show frequency, volume, loaded labor cost, error cost, tools, and existing review time.

  3. 03
    Make the projection inspectable.

    State every assumption, deduct runtime and residual review, separate one-time cash release from recurring savings, and show a conservative range.

  4. 04
    Instrument the build.

    Log runs, exceptions, human interventions, failures, and relevant cost changes. A green dashboard is not enough.

  5. 05
    Compare like with like.

    Re-measure the same line items, then adjust for material changes such as order volume, staffing, wages, prices, or tool changes.

  6. 06
    Settle from the verified number.

    Do not count the same benefit twice. If verified savings are lower than projected, the fee is corrected under the signed agreement.

Method references, not borrowed credibility

Our structure is informed by the Efficiency Valuation Organization’s IPMVP framework for measurement boundaries, before-and-after periods, and adjustments; the HM Treasury Green Book for separating appraisal from evaluation, exposing optimism bias, and testing assumptions; and the U.S. GAO Cost Estimating and Assessment Guide for documented inputs, source data, sensitivity analysis, and updating estimates with actual costs.

Those references support the measurement discipline, not the dollar figures in our scenarios. See.ke does not claim IPMVP adherence, certification, or government endorsement.

Open the assumptions

Scenarios and demo builds

Demo-tested systems appear first. Open a card to inspect the inputs, limits, and available test evidence. A published benchmark supports only the assumption it actually measures, never the outcome by itself.

For founding clients

Help create the first verified row.

Early clients get priority build slots and direct access to the builder. In return, we ask for disciplined measurement and, only with written permission, an anonymized or named case study. Commercial terms do not depend on giving us a testimonial.

Inspect a sample Savings Audit
The Savings Audit

Bring us one process. We’ll show you how it would be measured.

20-minute fit check · no obligation · measured audit for qualified processes.