Zycus Horizon US Edition 2026 · September 21-23, 2026 Register Now

Get Started

Merlin Autonomous Sourcing

How much spend are you leaving unsourced?

Not what procurement controls. What has actually been to market.

Size it in four minutes
What it gives youOne figure, split into the part inside your managed number and the part outside it.
What it comparesYour coverage against Ardent Partners research across 326 procurement executives.
What you leave withThe cost of closing it in people or in weeks, and a written brief you can send on.

The Framework

One reported number. Four different realities.

A category joins spend under management the day a contract is signed, and stays there for the life of that contract, tested or not.

Band 01

Competitively tested

Went to market inside your window. The only spend you hold evidence on.

Band 02

Managed but untested

Under contract, inside your reported figure, last contested longer ago than any defensible interval.

The gap
Band 03

Legitimately excluded

Sole source, regulated, locked, or below threshold. You define this band yourself.

Band 04

Unmanaged

Outside procurement entirely. Unsourced too, but already visible and already on somebody's plan.

Also unsourced

Bands two and four are both unsourced. Only one of them is visible.

Assessment

Size your own gap

Eight answers. No savings rate is applied to anything.

Start with the size of the pot

Two numbers most procurement leaders can give from memory.

Indirect and services spend your team could source. Exclude direct materials, taxes and intercompany.

million

Enter figures in the currency you selected. Nothing is converted, and the published research referenced later is USD based.

Under contract or procurement control. Your reported figure, as a share of the spend above.

75%of addressable spend

Now take out what should never go back to market

You draw this line, so nothing you exclude counts against you later.

Sole source, regulated, contractually locked, strategic partnership, or below threshold.

20%of managed spend

This decides how much room the assessment leaves around your next answer. A busy two years flatters your long run rate, a quiet two years understates it.

A real competitive event. Renewals, rollovers and single source extensions do not count.
Quick way to check: open your sourcing system, count events that reached award in 24 months, and divide by the categories that should have run in that time.

45%of what remains

How much sourcing capacity do you have

This is what turns the gap into a number of people or a number of weeks.

People who actually run events, not the whole procurement function.

people

Request to award. APQC benchmarks the range at 52 to 74 days.5

weeks

Roughly. Used only to suggest where a first wave should start.

million

Last two

These shape the recommendation rather than the arithmetic.

Defaults to the four year cap on EU public framework agreements, used as a reference point.1

Closing It

Coverage has only ever had two levers

More people

Scale headcount by the multiplier. An argument most sourcing leaders have already made once, and already know the answer to.

Shorter cycles

Divide the cycle by the multiplier and the same team meets the interval. Hackett puts Digital World Class procurement at 24% shorter sourcing cycles.4

What changed

Most of a cycle is assembly rather than judgment. Framing strategy, finding suppliers, chasing responses, normalizing bids.

Where Merlin Autonomous Sourcing Fits

Your S2P system records the event. Merlin Autonomous Sourcing strategizes and executes it.

An execution layer rather than a replacement, so Ariba, Coupa, Ivalua, Jaggaer, Oracle or SAP stays exactly where it is. Gartner's caution about agent washing is worth taking seriously, and the honest test is whether a commercial decision ever leaves human hands. Here they do not. What changes is how much preparation your team does before reaching them.

Merlin Autonomous Sourcing

Agents perform tasks. Merlin Autonomous Sourcing delivers sourcing outcomes.

Bring your result to a working session and we will scope the first wave against your own categories.

Explore Merlin Autonomous Sourcing Request a demo
32+
Patents in procurement AI
$1Tr+
Spend processed globally
Leader
Gartner® Magic Quadrant™ for Source-to-Pay Suites, 2026
Sources and method
  1. EU Directive 2014/24/EU, Article 33(1). A public framework agreement shall not exceed four years except in duly justified cases. Used as a defensible reference interval, not a rule governing private indirect spend.
  2. Common Paper, Contract Benchmark Report 2024 Q1. Around 90% of B2B SaaS cloud service agreements renew automatically, 29% with an automatic fee increase. Sell side SaaS data, not all indirect categories.
  3. Ardent Partners, CPO Rising 2025. Average spend under management 70.8%, best in class 91.7%, all others 61.1%. Research with 326 CPOs and procurement executives. Study co-sponsored by Zycus.
  4. The Hackett Group, 2025. Digital World Class procurement organizations run 24% shorter sourcing cycles than peers.
  5. APQC Open Standards Benchmarking. Sourcing event cycle time, need identified to contract signed, 52 days at the fastest organizations and 74 at the slowest.
  6. World Commerce and Contracting with Deloitte, The ROI of Contracting Excellence, 2023. Average contract value erosion 8.6% across 1,236 organizations. Shown on this page as context only. It is an average across all contracts, not a measured rate for untested spend, and it is never applied to your gap figure.

Method note. Findings are stated across the range of your tested share estimate, at plus or minus five points, rather than at a single value. Where that range straddles your chosen interval, the assessment reports the direction as unresolved instead of picking a side. Two simplifications apply to the closing figures: required cycle time assumes throughput scales in inverse proportion to elapsed cycle time, and required headcount assumes constant returns to scale in a sourcing team. Neither holds exactly, and both are stated here so the figures can be read with that in mind.

This assessment sizes exposure and capacity. It does not calculate savings, and no savings rate is applied to any figure it produces. Every number in your result is arithmetic on the eight answers you entered. Results are directional and intended to frame a conversation, not to replace a value assessment.

Gartner is a registered trademark and service mark and Magic Quadrant is a registered trademark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and are used herein with permission. All rights reserved. Gartner does not endorse any vendor, product or service depicted in its research publications and does not advise technology users to select only those vendors with the highest ratings or other designation.

New_Zycus - Header 2025
Linkedin-in Youtube Instagram X-twitter Facebook-f

Subscribe to our newsletters

This field is for validation purposes and should be left unchanged.
This field is hidden when viewing the form
Consent

Solutions

Products

Industry

Process

Merlin Agentic Platform

Services

Zycus Community

Company

Resources

Discover

News & Events

Experience

Our Presence

Copyright 2026 Zycus Inc. All Rights Reserved.

CSR Policy   Terms of Use  Privacy Policy  GDPR  Cookie Policy

This field is for validation purposes and should be left unchanged.
Consent

Start typing and press enter to search