Your scores and recent movement. The Coach can compare the latest run with prior context and point to what likely changed.
How the Coach remembers
A coach without memory is a search engine. This is what memory looks like.
Four sessions. Nineteen days. The Coach knows what Marcus fixed in session 1 when he asks a question in session 4, and uses that history to give him a different answer than a stranger would.
What you see
The actual Coach conversation across four sessions.
Coach Session · Spans 3 Weeks
Example: pre-seed B2B SaaS founder, four sessions over 19 days. Anonymized.
PV
PitchVault Coach
Working with Marcus · DevTools SaaS · Pre-seed
Session 1 · March 14, 9:42 AM · v1 score: 54
My VaultMoat is 28. Investors keep saying my moat is weak. What do I actually do about it?
Your deck names AI-powered as the moat. That is not a moat. It is a category. The 28 reflects that the Coach cannot find a switching cost, a network effect, or a data flywheel anywhere in the deck. Three things would move this score: name what specifically gets harder for a competitor each month you operate, name what compounds with usage, or name what a $50M-funded entrant would have to build that they cannot shortcut. Pick one to start. The slide rewrite is about 2 hours.
Session 2 · March 18, 7:14 PM · v2 score: 61 (+7)
Score went up but VaultMoat only moved 4 points. Why?
You rewrote slide 9 to describe integration depth. That is a real switching cost mechanism, which is why VaultMoat moved at all. But it scored 32, not 60+, because the rewrite asserts the cost without quantifying it. The Coach is looking for how many integration points an average customer has at month 6 versus month 1. What does ripping you out and replacing you cost in engineering hours? You wrote the story. The score wants the data.
Session 3 · March 26, 10:08 AM · v3 score: 68 (+14)
Got an investor pass yesterday. They said interesting, but the moat feels thin. I just rewrote the moat slide. What gives?
The moat slide moved your VaultMoat from 32 to 56. That is real. But the investor is not reading slide 9 in isolation. They are cross-referencing it with the traction slide. Your traction slide still says growing fast without numbers. A moat without traction proof reads as a thesis, not a moat. The pattern is: investors believe the moat after they believe the traction. Fix slide 5 next. About 45 minutes if your numbers are clean.
Session 4 · April 2, 6:30 PM · v4 score: 76 (+22)
I am investor visible. Should I send to my warm list now or wait?
Send. The deck moved 22 points across four versions in 19 days. That trajectory is part of what investors see. Two notes: VaultRisk is still 58, so the GTM slide will draw questions about CAC payback. Have a one-line answer ready. And the same investor who passed in March will respond differently to a 76-point version. Do not skip them. The pass was specific feedback. This is the response.
About this session
An anonymized reconstruction of live Coach behaviour.
Founder name, company, timestamps, slide numbers, and company specifics are changed. The cadence, score movement, and reasoning pattern show what persistent memory is meant to do.
This is what 19 days of memory does that 19 fresh prompts cannot.
How it works
Most AI tools see each prompt as a fresh request. The Coach sees a relationship.
In session 1, Marcus has a moat problem. The Coach diagnoses it specifically: he named a category, not a mechanism. In session 2, the score moved but VaultMoat barely moved, so the Coach explains the missing data instead of asking Marcus to re-explain the deck.
In session 3, Marcus logs an investor pass that sounds like a moat objection, but the Coach connects it to the traction gap and the moat history. In session 4, the deck has moved 22 points and Marcus asks whether to send. The answer uses trajectory, remaining risk, and the feedback Marcus recorded.
What the Coach remembers01 02 03 04
Four kinds of context the Coach can pull from.
Your Action Plan history. Which steps were marked done, which were attempted, and which still need proof.
Your founder-entered raise context. Investor feedback, passes, blockers, and notes are useful when you log them.
Your context. Stage, sector, geography, raise target, and founder constraints you have shared.
Honesty contract
What the Coach won't do.
Won’t help you fake traction. If a fix needs evidence you do not have, the Coach says so.
Won’t tell you what you want to hear. The default is honest, not flattering.
Won’t replace your judgment. The Coach gives context-aware reasoning. You decide what to do.
Why this matters
A score is a moment. A coach is a relationship.
Fundraising is the wrong place to forget what worked last week. The deck moves through five to ten revisions over six to twelve weeks. Every revision is supposed to learn from the last one. A coach without memory makes every conversation start over.
The four scores tell you what is true today. The Coach is what makes the next revision smarter than the last.
Connected pieces
Where the Coach fits with the rest of the loop.
My Vault →
The Coach drives the headline and current blocker on the founder home base.
Action Plan →
The Coach helps keep the ordered fixes connected to what you completed, corrected, and verified on reruns.
Raise Plan →
The Coach can update the plan when you log outcomes, passes, check-ins, or blockers.
PitchVault Coach is included with Investor-Grade Analysis.
One purchase. $199. The Coach carries your context across every revision.
Unlock the full analysis →