# Marcus Delgado, Director of Revenue Operations at Latterly (B2B SaaS, ~160 employees, Series B) — read of Sales Stage Predictor AI, May 30, 2026

> Nine years in RevOps, three of them trying to get reps to update Salesforce honestly. Currently running HubSpot + Clari + Gong. Two kids, coaches U10 soccer on Saturdays, bikes to the office when it's not raining in Portland.

## How I got here

I was Googling "deal stage accuracy HubSpot pipeline" because our Q2 forecast call was a mess and I wanted to see what other people were doing about it. Clicked through a few listicles, found a comparison article that mentioned "AI stage prediction tools," and this was one of three links I opened in tabs. I did not come here from an ad. I came here from frustration.

## What I clicked first

The subheadline grabbed me: "Your reps guess. Your deals sit in 'negotiation' for 6 months. Your forecast misses." That's my life. I didn't even finish reading it, I scrolled straight to the feature list because I've seen this promise before and I want to know HOW, not what.

The comparison table caught my eye too, specifically the row "Reps hide stalled deals" under Manual Pipeline Review. That's accurate and kind of funny that they just say it plainly.

## Where I paused

The testimonial from Sarah Chen: "We moved 6 deals from 'hopefully closing' to 'actually closing' just by moving resources to the ones that had real momentum." I stopped here because that's exactly the behavior I want to change. If that's real, I want to talk to Sarah Chen. But the attribution is "VP Sales, FinTech SaaS" with no company name. That's a yellow flag.

The second one from Michael Rodriguez is even more specific: "$180K ARR saved." That number is precise enough to feel real. Or it's precise enough to feel constructed. Hard to tell.

## What I distrusted

The bottom of the page broke my brain. There's a section that says "How honest is this idea, really?" and then lists: "60/100 Adoptability. $-21,000 Year-1 take-home (Fermi). 1 in 7 Meaningful-success odds." And then: "Honest disclosure: we don't have live customers on this idea yet. We shipped the strategy package; you ship the customer conversations."

I had to read that three times. So... this is not a product. It's a product idea someone is selling as a kit? The pricing page above this shows $299/month plans with a free trial, but the disclosure says there are no live customers. So what am I trialing? A wireframe? A pitch deck? The pricing section and the disclosure section are describing two completely different things and they're on the same page.

If you scroll to the bottom and read carefully, you realize "Adopt the build" for $99-$199 means you're buying the code starter and outreach pack to build this yourself. That's not a SaaS product. That's a franchise idea with a Figma file attached.

## What would convince me

If this were a real, live product: two reference customers I can call, in companies of my size, with their actual company names attached. Not "FinTech SaaS." A company name. One 15-minute call with a revenue ops person who's had it running for 60+ days and can tell me their false-positive rate on the confidence scores.

Specifically on the confidence scores: "95% confident this deal moved to negotiation? We tell you. 60% confident? We flag it for your review." I want to know what the baseline accuracy is when the model says 95%. Does that mean 95 out of 100 of those actually close as negotiation stage? Or is it a softmax score that doesn't correspond to real-world accuracy? That distinction matters a lot and the page doesn't address it.

## What I'd ask in an email reply

1. The page shows a pricing table with $299/month tiers and a free trial, but the footer says there are no live customers yet. Can you help me understand what I'd actually be signing up for?

2. The confidence scores, like "95% confident this deal moved to negotiation": what does that number actually represent? Is it a model output that's been calibrated against real outcomes, or is it a raw probability that hasn't been validated?

3. You mention 40+ engagement signals including "competitor research." How do you detect that? Are you crawling browser activity, reading email content, pulling from Gong transcripts? That question isn't about privacy, it's about whether the signal is actually available in my stack.

## Verdict: dismissive

The actual product problem is real and the framing is sharp. But the page is advertising a live SaaS while disclosing at the bottom that it's an unbuilt idea for sale as a kit. Those two things cannot coexist on the same page without destroying trust. I closed the tab.

---
*Memo by skeptic persona, generated 2026-05-30. Studio breaks own self-grading loop.*