Microsoft scores your referrals. Here's what moves the score.
Partner Center attaches a Referral Confidence Score to what partners share — Microsoft calls it a quality snapshot that shows sellers which referrals stand out. Outbound opportunities are rated High, Medium, or Low by a machine-learning model, and every inbound co-sell referral is checked for completeness and quality at submission. The score is decided by work done before the referral exists.
What Microsoft checks
Real-time guidance during creation
While a referral is being created, Partner Center flags missing detail and explains what would make the referral stronger — before it's submitted, not after it's ignored.
Passive data-quality flags
Partner Center watches for patterns that undermine confidence: a customer email domain that matches the partner's own, deal values that are unusually high or low, a close date inside seven days.
ML scoring on outbound opportunities
A machine-learning model rates outbound opportunities High, Medium, or Low — a consistent read across every referral, shown to sellers with an explanation of what makes strong ones strong.
Automatic checks at submission
Since March 2026, all inbound co-sell referrals are checked automatically for completeness and quality the moment they're submitted. There is no unscored path in.
Source: Microsoft Partner Center documentation and the March 2026 Partner Center announcements. Microsoft controls the score and its naming — this page is reviewed quarterly.
The part nobody mentions
The score isn't in the API. You can only influence it upstream.
Microsoft does not expose the Referral Confidence Score in the Partner Center Referrals API. No CRM integration, Wyra's included, can pull it onto your records — a claim to the contrary is a claim about something Microsoft doesn't offer. That makes the Microsoft score different from AWS's Opportunity Quality Score, which does come back through the Partner Central API.
It also makes the conclusion sharper, not weaker: if you can't read the score back, the only way to manage it is to deserve it — complete referrals, real customer need, plausible values and dates, qualification that stands up when a Microsoft seller opens it with 14 days to decide.
Every cloud now grades partner pipeline
The pattern is bigger than Microsoft. AWS scores every shared opportunity and routes it to a co-sell motion. Microsoft scores referral confidence and checks quality at submission. Google Cloud rebuilt its competency framework around contribution to validated closed/won work. Different vocabularies, one direction: the provider measures what partners bring, and the measurement is set upstream.
AWS
Opportunity Quality Score — 0–100 with a trend, exposed through the Partner Central API and returned onto the record.
Microsoft
Referral Confidence Score — High / Medium / Low from an ML model, visible in Partner Center, not exposed in the Referrals API.
Google Cloud
No published referral score — capability is measured instead by pre-sales and post-sales contribution to validated closed/won opportunities.
Where Wyra fits
Wyra works on the side of the line the score is decided on. Every prospect is qualified in writing — the fit, a why-now, and the evidence — or disqualified with the reason written out, before anything sends. Deals your team converts land in your CRM on Microsoft's sales stages and sync two-way with Partner Center when you choose to share them. What arrives in the Referrals workspace is a referral with its homework attached — which is what every layer of Microsoft's checking is looking for.
Referral Confidence Score: common questions
What is the Referral Confidence Score?
It's the quality signal Microsoft Partner Center attaches to referrals. Microsoft describes it as a quality snapshot that helps sellers see which referrals stand out, with an explanation of what makes them strong. Outbound opportunities are scored High, Medium, or Low by a machine-learning model, and partners get real-time guidance on missing detail while a referral is being created.
What does Microsoft check on a referral?
Three documented layers. Real-time guidance flags missing detail while you create the referral. Passive data-quality checks flag suspicious patterns — a customer email domain matching your own, unusually high or low deal values, a close date inside seven days. And since March 2026, every inbound co-sell referral is automatically checked for completeness and quality at submission.
Can I read the score back through the Referrals API?
No. Microsoft does not expose the Referral Confidence Score in the Partner Center Referrals API, so no CRM integration or co-sell tool can pull it onto your records — any tool claiming otherwise is claiming something Microsoft doesn't offer. The score is visible in Partner Center, and the way to move it is upstream, in the quality of what you share.
How does it compare to AWS's Opportunity Quality Score?
They're the same idea from two clouds. AWS scores every shared opportunity 0–100 with a trend and exposes it through the Partner Central API; Microsoft scores referral confidence High/Medium/Low and keeps it inside Partner Center. Both mean the era of sharing thin pipeline quietly is over: the provider now grades what partners bring, consistently and at scale.
How do I improve my Referral Confidence Score?
By fixing what Microsoft measures before the referral exists: complete detail, a real customer need, plausible deal values and dates, and qualification that stands up. That work happens in your pipeline, not in the Partner Center form. Wyra produces it by default — written insights and a why-now on every prospect before outreach sends — so the referrals your team chooses to share arrive with their evidence attached.
The score is set before the referral exists.
Book a demo and see the qualification your referrals would carry — written, evidenced, and ready for a seller with 14 days to decide.
Book a Demoarrow_forward