Skip to main content
Case studies

Case Studies: What We Have Measured, and How to Check It Yourself

Reference results from Cresolv One installations, with the client masked, the method stated and the date given. Where we only have a worked example, we label it illustrative.

Reference case: figures checkedReference case, October 2026
Exact counts where measured
Illustrative examples labelled
  • Client names masked by default
  • Method and date stated
  • Illustrative examples labelled
676 / 676
Annual tracker figures matched (India reference case)
125 / 129
Quarterly competitor-deck figures matched
3,800 / 3,800
Deck figures traced to document, page and printed line
About 25 min
To read 52 annual reports for 4 companies

Case studies

How we report results

A result is only useful if you can tell what it was measured against. So every number on this page comes with its context, and we hold ourselves to a short list of rules.

Reference case: competitor financials checked against a client's own tracker

We tested CompeteIQ on our India reference installation, a listed manufacturer, in October 2026. The test used that company's own hand-built material as the answer key: an annual source tracker covering FY14 to FY25 and the quarterly competitor decks the team had produced by hand.

CheckResult
Annual source tracker, standalone, FY14 to FY25676 of 676 figures matched
Quarterly competitor decks125 of 129 figures matched
Deck figures traced to document, page and printed line3,800 of 3,800
Reading 52 annual reports for 4 companies, FY14 to FY26About 25 minutes
Errors found in the client's consolidated tracker7

The useful part is not the match rate on its own. Every figure in the output carries a link back to the printed line and page it was read from, so when a number differs from the client's tracker, the difference can be settled by opening the filing. That is how the 7 errors in the consolidated tracker were identified.

Reference-case results from one production installation. Your own results come from a reconciliation during a pilot on your decks. Read the full product description on the CompeteIQ page.

Want the same reconciliation on your own tracker?Share a past benchmarking deck and we will reproduce it.
Request a Reconciliation

Worked example: turning plan demand into a buy list (illustrative)

Not every product has a published client case yet. For the AI Procurement Planner, here is the netting logic on demo numbers, so you can see how a buy-list line is built and where you would check it.

StepQuantity
Highest plan demand across sources1,200
Add stock awaiting a component+ 150
Subtract stock on hand− 400
Subtract quantity on open purchase orders− 300
Subtract quantity in transit− 200
Net need450
Vendor MOQ is 250, so the order rounds up to500

Illustrative example with demo data. Your own figures come from your ERP exports. See the AI Procurement Planner page for what works today and what is planned.

Where to look next

If you are weighing a specific product, each page says what it does today, what is planned, and what a pilot would use as its starting data.

Is this the evidence you need?

A good fit if

  • You want evidence and a method before you commit to anything
  • You can share a small sample of your own data from one entity, team or category
  • You are comfortable with masked client names and stated methods

Probably not yet if

  • You need named public client references today
  • You want a result quoted without a baseline from your own process
Start Small

Check It on Your Own Data

A pilot is the way to turn a reference case into your own number. Pick one entity, team, gate or category and compare the output with your current process.

1
Share a sample

A small batch of your own documents or exports from one entity, team or category.

2
We run it on your data

Configured to your policy and run next to your current process, with exceptions listed and explained.

3
Review together

Walk the results line by line and compare them with how your team does the work today.

See How a Pilot Works

How a result gets onto this page

Every number follows the same rules, so you can tell what it was measured against.

BaselineSay what it was measured against and when
CountExact counts where measured, "up to" where estimated
MaskDescribe the client type, name it only with agreement
LabelDemo-data examples are marked illustrative
PublishOne installation is presented as one installation

Frequently Asked Questions

How we report results, what a reference case means, and how to check a result on your own data.

Why are the clients not named?

We mask client names by default and name a client only when they have agreed to it. A reference case tells you the type of company, the month it was measured and the method, so you can judge whether it looks like your situation.

What does "India reference case" mean?

It means one production installation at a listed Indian manufacturer, measured in October 2026 against that company's own hand-built material. It is one installation. It shows what the product did there, not what it will do for you; your own results come from a pilot on your data.

How do you phrase results?

Measured counts are given exactly, with the context they were measured in. Estimates are given as ranges or with a qualifier, for example "up to" or "around". We state the baseline where there is one, and we do not quote a time saving until we know how long the work takes you today.

Is there a case study for every product?

No. Today there are two measured cases: the CompeteIQ reconciliation described on this page, and the AP Automation case at an e-commerce company, which shows counts and screens but no before-and-after figure. For the other products, the product pages say what works today and what is planned, and the pilot is how you check it on your own data. We add a case here only when the client has agreed and the figures are checked.

What is an "illustrative" example?

A worked example built on demo data to show the logic, such as the buy-list netting example on this page. It is not a client result, and each one is labelled as illustrative.

How is a pilot different from a demo?

A demo uses our data. A pilot uses yours: you pick one entity, team, gate or category, share a sample, and we run it next to your current process so you can compare the two. See the pilot page for the steps.

What do you need from us to start?

A small sample of your own documents or exports from one entity, team or category, and a person who can review the output with us. We do not ask you to rebuild anything, and each product page lists what it needs.

How do I get in touch?

Use the form on this page, WhatsApp us on +91 7738829118, or email sales@cresolv.com. Tell us which product you are looking at and what you would want to test.

Talk to us

Tell us what you would want to test

Share a few details and we will set up a call. If you already know which product you are looking at, say so and we will use your kind of data in the walkthrough.

What happens next

  1. We reply to schedule a call.
  2. We walk through the nearest reference case or worked example.
  3. We agree one entity, team or category for a pilot, if it makes sense.

Prefer WhatsApp? +91 7738829118 or sales@cresolv.com.

We use your details only to respond to this enquiry.

See It on Your Own Data

Pick one entity, team, gate or category, share a sample and compare the output with your current process.