Presentation Tie Out A.I. Skill
Hand Claude your investor deck, IC memo, OM, or lender package before it goes out. It reads every number on every slide, checks them against each other, and checks them against your model when you provide one.
This is a numbers check, not a design review. It will not tell you the layout needs work or the returns look thin. It tells you whether the figure on slide 4 still agrees with the figure on slide 17, and whether either one agrees with the model.
What It Checks
The same metric repeated across slides, so the cover, the executive summary, and the appendix all say the same thing
Arithmetic that should foot: line items against totals, percentages against their stated base
Derived metrics that should reconcile: price per unit, cap rate, LTV, DSCR, equity multiple
Scenario mixing, where the exit cap traces to the upside tab while the IRR traces to base case
Scale and basis errors: $1.2M shown for a model value of $1,200, dollars per SF presented as dollars per unit, trended shown where untrended was meant
Leftover template text: placeholders, XX%, and the prior deal's sponsor name still sitting on slide 2
How It Reports
Findings come back in chat as Critical, Medium, Minor, Ambiguities, Unverifiable, and Coverage. The Unverifiable section is never omitted, because numbers baked into a chart image cannot be read and you need to know that rather than assume silence means clean. It never edits your deck. Providing the model is optional, and without one the skill still checks the deck against itself.
Pairs With
The CRE Model Audit skill checks whether your model's formulas are right. This one checks whether your deck matches the model. Run both before a package leaves the building.
Perfect For
Anyone sending a deck to investors, IC, or a lender. Especially useful on white-label work, where last deal's client name on slide 2 is the kind of error that costs a relationship rather than just a correction email.
Requires: Claude with file uploads and code execution. Format: .skill
Hand Claude your investor deck, IC memo, OM, or lender package before it goes out. It reads every number on every slide, checks them against each other, and checks them against your model when you provide one.
This is a numbers check, not a design review. It will not tell you the layout needs work or the returns look thin. It tells you whether the figure on slide 4 still agrees with the figure on slide 17, and whether either one agrees with the model.
What It Checks
The same metric repeated across slides, so the cover, the executive summary, and the appendix all say the same thing
Arithmetic that should foot: line items against totals, percentages against their stated base
Derived metrics that should reconcile: price per unit, cap rate, LTV, DSCR, equity multiple
Scenario mixing, where the exit cap traces to the upside tab while the IRR traces to base case
Scale and basis errors: $1.2M shown for a model value of $1,200, dollars per SF presented as dollars per unit, trended shown where untrended was meant
Leftover template text: placeholders, XX%, and the prior deal's sponsor name still sitting on slide 2
Hand Claude your investor deck, IC memo, OM, or lender package before it goes out. It reads every number on every slide, checks them against each other, and checks them against your model when you provide one.
This is a numbers check, not a design review. It will not tell you the layout needs work or the returns look thin. It tells you whether the figure on slide 4 still agrees with the figure on slide 17, and whether either one agrees with the model.
What It Checks
The same metric repeated across slides, so the cover, the executive summary, and the appendix all say the same thing
Arithmetic that should foot: line items against totals, percentages against their stated base
Derived metrics that should reconcile: price per unit, cap rate, LTV, DSCR, equity multiple
Scenario mixing, where the exit cap traces to the upside tab while the IRR traces to base case
Scale and basis errors: $1.2M shown for a model value of $1,200, dollars per SF presented as dollars per unit, trended shown where untrended was meant
Leftover template text: placeholders, XX%, and the prior deal's sponsor name still sitting on slide 2
How It Reports
Findings come back in chat as Critical, Medium, Minor, Ambiguities, Unverifiable, and Coverage. The Unverifiable section is never omitted, because numbers baked into a chart image cannot be read and you need to know that rather than assume silence means clean. It never edits your deck. Providing the model is optional, and without one the skill still checks the deck against itself.
Pairs With
The CRE Model Audit skill checks whether your model's formulas are right. This one checks whether your deck matches the model. Run both before a package leaves the building.
Perfect For
Anyone sending a deck to investors, IC, or a lender. Especially useful on white-label work, where last deal's client name on slide 2 is the kind of error that costs a relationship rather than just a correction email.
Requires: Claude with file uploads and code execution. Format: .skill
The Fractional Analyst’s
Presentation Tie Out A.I. Skill
Upgrade your underwriting process with our resources that deliver speed, accuracy, and confidence across all your deals
Hand Claude your investor deck, IC memo, OM, or lender package before it goes out. It reads every number on every slide, checks them against each other, and checks them against your model when you provide one.
This is a numbers check, not a design review. It will not tell you the layout needs work or the returns look thin. It tells you whether the figure on slide 4 still agrees with the figure on slide 17, and whether either one agrees with the model.
What It Checks
The same metric repeated across slides, so the cover, the executive summary, and the appendix all say the same thing
Arithmetic that should foot: line items against totals, percentages against their stated base
Derived metrics that should reconcile: price per unit, cap rate, LTV, DSCR, equity multiple
Scenario mixing, where the exit cap traces to the upside tab while the IRR traces to base case
Scale and basis errors: $1.2M shown for a model value of $1,200, dollars per SF presented as dollars per unit, trended shown where untrended was meant
Leftover template text: placeholders, XX%, and the prior deal's sponsor name still sitting on slide 2
How It Reports
Findings come back in chat as Critical, Medium, Minor, Ambiguities, Unverifiable, and Coverage. The Unverifiable section is never omitted, because numbers baked into a chart image cannot be read and you need to know that rather than assume silence means clean. It never edits your deck. Providing the model is optional, and without one the skill still checks the deck against itself.
Pairs With
The CRE Model Audit skill checks whether your model's formulas are right. This one checks whether your deck matches the model. Run both before a package leaves the building.
Perfect For
Anyone sending a deck to investors, IC, or a lender. Especially useful on white-label work, where last deal's client name on slide 2 is the kind of error that costs a relationship rather than just a correction email.
Requires: Claude with file uploads and code execution. Format: .skill