How many prospects can I run through this workflow at once?

Cap a single batch at a size your pitch calendar can actually use. Most tools that support bulk prospecting cap a single import around 20 domains, since each one triggers a full report and a huge batch just slows every report down. Run a second batch next week for anything you didn't get to.

Does running a report on a prospect's site require their permission?

No. A visibility report reads public information, the prospect's own website and how AI assistants already talk about them. It's no different from running an SEO audit on a competitor's public site. Treat it like any other prospecting research, for a real pitch, not for anything else.

What happens to the property if the prospect becomes a client?

Nothing has to change. A prospect is already a full property under the hood, just flagged separately so it doesn't clutter your client list. If they sign, you switch the flag and keep going, no new setup, no lost report history.

Should I run one report at a time or always in batches?

Batch it whenever you're comparing more than one prospect for the same pitch cycle, since running them the same day on the same scan keeps every comparison apples-to-apples. A single hot lead that comes in outside a batch is fine to run on its own.

How to Prospect New Clients with AI Visibility Reports

Q: What's the single most important thing to include in the pitch?

One specific, competitor-named finding, not the full report. "AI recommends your competitor for this exact search and never mentions you" is a sentence a prospect remembers. A full breakdown of every metric is something they skim and forget.

To prospect new clients with AI visibility reports, build a list of target domains, run them through a full AI visibility scan as one batch, then read each report for the single finding that makes the best opening line. Package the strongest findings into a white-labeled report and send it before the call, not after. That's the whole workflow: one batch, one read-through per prospect, one branded deliverable that starts the meeting for you.

Most agency pitches lead with a capabilities deck. Slides about the team, past clients, a services list the prospect has already seen from two other agencies this month. It doesn't land, because it's about you, not them. A report that shows a prospect exactly how AI assistants describe their brand right now, and which competitor keeps showing up instead of them, lands every time. It's specific, it's about their business, and it's something they can't put together themselves in an afternoon.

This guide works with any AI visibility workflow; the callouts show how each step runs in Spyglasses.

Step	What you're doing	What tells you it's working
1. Build your prospect list	Compiling target domains into one batch	A validated list, ready to run
2. Run reports at scale	Scanning the whole batch across AI platforms	Every prospect has a completed report
3. Turn reports into a pitch	Finding the one number that opens the call	A specific, competitor-named finding per prospect
4. Share the report	Packaging and sending the branded export	A white-labeled PDF or link in the prospect's inbox

Build your prospect list

Start with a list of domains, not one property at a time. If you're prospecting a vertical, pull ten or twenty target accounts from wherever your business development process already lives, a CRM export, a conference attendee list, a referral pipeline, and get all of them into one batch. Comparing several prospects side by side on the same day, scanned the same way, gives you an apples-to-apples pitch instead of one-off snapshots built weeks apart.

Keep the batch a manageable size. A batch that's too large slows down how fast any single report finishes, and you only need enough prospects to fill your pitch calendar for the next couple of weeks, not your entire target account list at once.

Run AI Visibility Reports at scale

Once the list is validated, run it. A full AI visibility report tests the brand across the major AI platforms in one pass: how consistently AI assistants describe the business, how often it shows up in discovery queries next to competitors, and which sources AI cites when it answers questions about that category. All of that comes from a domain name, no login or access to the prospect's own accounts required.

A batch of reports takes real time to finish, sometimes a few hours for a full list, since each one runs its own set of platform queries. Start the batch in the morning and it's usually done well before your next round of pitch calls.

Pull quote: Nobody remembers a capabilities deck. Everybody remembers the meeting where you showed them exactly where they were losing ground. — “Nobody remembers a capabilities deck. Everybody remembers the meeting where you showed them exactly where they were losing ground.” — Spyglasses

Turn each report into a pitch

A full report has a lot in it: consistency scores, share of voice, discovery queries, citation sources, competitor mentions. Don't hand a prospect all of it and hope something lands. Read each report looking for one finding, the single number or comparison that makes the meeting worth taking.

The strongest opening finding is usually a named competitor. "ChatGPT recommends [competitor] for this exact search and never mentions you" is a sentence a prospect remembers. A low consistency score works too, especially paired with a specific wrong detail: an outdated price, a discontinued product, a description that reads like it's about a different company. Save the rest of the report, the discovery query breakdown, the full citation list, for after they've agreed to the meeting.

Once you know which finding leads the pitch, package it. Download the report as a white-labeled PDF or PowerPoint with your agency's branding, not Spyglasses', and send it ahead of the call, or hand it over live as a leave-behind. A prospect looking at a polished, branded deliverable with their own brand's data on it reads it as evidence you already understand their problem, before you've said a word.

Send it before the call if you can. A prospect who's already opened a report showing a competitor beating them shows up to the meeting with a question, not a blank stare, and that question is the one you wanted them to ask.

A pitch built on their data closes faster than a pitch built on yours

The workflow doesn't change much once you've run it a few times: build the batch, run the scan, read for the one finding that stings, send the branded report before you ask for the meeting. The step teams skip most often is the read-through, sending the full raw report instead of leading with a single sharp finding, and it's the step that makes the difference between a prospect who skims and a prospect who calls you back.

If the prospect signs, the property doesn't start over. It's already a real property with its own report history, so the next step is baselining it properly and correcting the brand profile everything else builds on. From there, tracking visibility over time turns the one-time report you pitched with into the ongoing number you report against.