Examples

Example 1: Turning a Client-Call Summary Into Proof Points

A SaaS client wrapped a quarterly review call, and the account manager wrote up a short summary: the client's inbound leads tripled over ninety days, their head of growth said the new content engine "finally made our pipeline predictable," and a benchmark from the call put their cost per lead at roughly half the category average.

Step 1: Feed the source. On the Proof Points library page, the account manager gives Proof Points the call summary to extract from.

Step 2: Candidates appear. Within moments, the page shows three new candidates in Needs Review, each tied back to the summary:

• Outcome: "Inbound leads increased 3x over 90 days." Category: lead generation. Confidence: high.
• Quote: "Finally made our pipeline predictable," attributed to the client's head of growth. Category: content. Confidence: high.
• Stat: "Cost per lead roughly half the category average." Category: lead generation. Confidence: medium.

Proof Points kept every number exactly as the summary stated it. It did not round the 3x up, and it did not turn "roughly half" into a precise figure. A line in the summary that was just the account manager's own opinion about the client's enthusiasm was left out: there was no source behind it, so it is not a proof point.

Step 3: Note the client name. Because the first two candidates name a specific client, Proof Points has prepared a public-safe version of each, restating the same outcome as "a B2B SaaS client" with the identity removed.

At this point nothing is usable yet. These are candidates sitting in Needs Review.

Example 2: Running Scrutiny, Then Approving

A content lead has accumulated a dozen candidates over a few weeks from several client calls and a case study. Before drafting season, they want to clean the set up and decide what the team is allowed to cite.

Step 1: Run Scrutiny. The content lead clicks Run Scrutiny. Proof Points spots that two candidates describe the same retention result captured from two different calls, merges them into one point with both sources, and scores the rest. The counts update: ten points now show as Scrutiny Passed.

Step 2: Read the result. The content lead opens the Scrutiny Passed view. The strongest evidence has risen to the top, ranked by score. One candidate scored low: a vague claim about "great engagement" with no number attached. It survived as a candidate but ranks at the bottom.

Step 3: Understand the gate. Passing Scrutiny ranked these points; it did not release them. None of them can be cited yet. The content lead opens the Needs Review queue to make that call.

Step 4: Approve and reject. Working through the queue:

• The retention outcome with two sources: Approve. It joins the citable library.
• The "great engagement" claim with no number: Reject, with a short reason ("no metric, not citable"). Proof Points uses that reason to extract better next time.
• A strong client quote that names the client: Approve. The library keeps its anonymized public version for any external draft.

The library count for usable, approved points goes up. Only the approved points are now available to the drafting agents.

Example 3: A Draft Cites an Approved Point

A week later, Marketing Content is drafting a blog post for the same SaaS client's industry, and Marketing SEO is preparing a supporting piece.

Step 1: The draft is grounded. When the blog draft appears in the app, it carries a highlight at the top:

This Draft Is Grounded In 2 Proof Points From Your Client Calls

Each proof point shows as a chip. The draft cites the tripled-leads outcome and the "predictable pipeline" quote, both in their anonymized form, because the source named a client.

Step 2: Check the source. Before approving the draft, the content lead selects a chip. It opens the Proof Points page, scrolled straight to that exact piece of evidence, so the original outcome and its source are right there. The claim in the draft matches the evidence behind it.

Step 3: Confidence to approve. Because the draft only used points the content lead had already approved, there is nothing fabricated to catch. The evidence is real, the public-safe version protects the client, and the source is one click away. The content lead approves the draft.

This is the payoff of the two-gate process: the drafting agents never reach for an unapproved or unverified claim, because the only evidence they can touch is what you released into the library.