# Recommendations: Gavel Exec

Based on the representative run (overall 80.47). Gains are counterfactual estimates from spec 2.2 arithmetic, not measurements.
Page-fixable ceiling: 85.4 — pageFixableCeiling rescores the representative run with every page-level unit met (copy, layout, on-page markup). It excludes site-wide files and scans (llms.txt, robots, markdown, URL, ora.ai, Lighthouse) and product flows (sign-up, checkout, scheduler, chat). Estimate, not a measurement.

## R-01 · Fix critical issue: Capterra rating aggregates all Gavel products rather than Gavel Exec

Type: page-fixable · effort S · predicted +0 overall (Content +0)

**Why**

- The page displays a 4.9-star Capterra rating with 51 reviews to support Gavel Exec, but explicitly notes in the fine print that this rating applies to 'Gavel, all products' (including its historical document automation platform Documate), not Gavel Exec specifically.
- Evidence: Capterra
- ★★★★★
- 4.9
- 51 reviews · Gavel, all products

**Change**

- Remove or correct the defect at 5 Customer Results & Case Studies \[PROOF\] so the page, its markup and its agent-facing files agree.

## R-02 · Fix major issue: Embedded demo iframe blocked by cookie consent gate

Type: page-fixable · effort S · predicted +0 overall (Content +0)

**Why**

- Embedded iframe content (video or interactive element) is blocked until the visitor accepts tracking/marketing cookies.
- Evidence: Iframe is blocked. Accept cookies to load it.PreferencesAccept AllPreferencesManage Consent Preferences EssentialsAlways

**Change**

- Remove or correct the defect at 2 Hero Section \[PO,DEMO,TRX\] so the page, its markup and its agent-facing files agree.

## R-03 · Performance: lighthouse

Type: site-level · effort S · predicted +1.3 overall (Performance +13.5)

**Why**

- performance:lighthouse scores 0.9 — Lighthouse-based Performance 87; target: 100 (maximum gain).

**Change**

- lighthouse: Lighthouse-based Performance 87; target: 100 (maximum gain).

## R-04 · AI readiness (ora.ai): domain

Type: site-level · effort S · predicted +1.2 overall (Agentic best practices +8.1)

**Why**

- AG6:domain scores 0.7 — ora.ai AI readiness 71 for the domain; target: 100 (maximum gain). (now: "ora.ai grade None for gavel.io, scanned 2026-10-24T17:36:08.857Z")

**Change**

- domain: ora.ai AI readiness 71 for the domain; target: 100 (maximum gain).

## R-05 · Buyer order & first screen: order

Type: page-fixable · effort S · predicted +0.9 overall (Layout & navigation +5.8)

**Why**

- LN2:order scores 0.6 — 15 section order inversion(s) against the buyer order PO → SUM → PAF → PROOF → DEMO → FEAT\_B → FEAT\_U → BQ → UQ; target: reorder.

**Change**

- order: 15 section order inversion(s) against the buyer order PO → SUM → PAF → PROOF → DEMO → FEAT\_B → FEAT\_U → BQ → UQ; target: reorder.

## R-06 · Buyer questions: Self-contained answers

Type: page-fixable · effort S · predicted +0.7 overall (Content +2.4)

**Why**

- BQ-D scores 0 — \[evidence not found on the page; scored 0\] Between 50% and 79% of answers restate the subject (Gavel Exec/Exec). (now: "Does Gavel Exec work inside Microsoft Word?
- A: Yes. Gavel Exec runs inside Microsoft Word as an add-in and online at exec.gavel.io...")

**Change**

- Self-contained answers: At least 80% of answers restate the subject (name the product or feature) = 1; 50–79% = 0.5.

**Competitors that meet it**

- GC AI (BQ-D): "Can GC AI Make a First Pass on Incoming Commercial Agreements? Yes. Run a Playbook against the agreement... Can GC AI Review Both the Business Discussion and the Documents? Yes. With Agent Connectors, GC AI can bring relevant email and Slac"

## R-07 · Format variety: formats

Type: page-fixable · effort S · predicted +0.6 overall (Multimedia +6)

**Why**

- MM3:formats scores 0.6 — Fewer than 5 distinct visual formats; target: 5 (e.g. real screenshot, annotated visual, diagram, table, video).

**Change**

- formats: Fewer than 5 distinct visual formats; target: 5 (e.g. real screenshot, annotated visual, diagram, table, video).

## R-08 · Priority order: Setup stated early in the text agents read; Access stated early in the text agents read; Security & data stated early in the text agents read (+4 more)

Type: page-fixable · effort M · predicted +1 overall (Agentic best practices +6.4)

**Why**

- F7 scores 0.3 — first mention at 85% of the agent text (now: "Adding Gavel Exec to Word takes about 30 seconds.")
- F4 scores 0.4 — first mention at 79% of the agent text (now: "Start with 25 free queries per user and collaborate with your team during the free trial.")
- F6 scores 0.4 — first mention at 74% of the agent text (now: "Zero Data Retention agreements with the AI model providers Gavel uses, including OpenAI.")
- F5 scores 0.8 — first mention at 32% of the agent text (now: "15 → about 5 hours First-cut review time per contract, after Martin built an in-licensing playbook of about 80 rules")
- F3 scores 0.9 — first mention at 21% of the agent text (now: "Start free with 25 queries per user, then pay $160 per user per month.")
- F1 scores 1 — first mention at 20% of the agent text (now: "Built for transactional attorneys and legal teams, Gavel Exec is AI contract review and drafting software that runs inside Microsoft Word as an add-in and on th")

**Change**

- Setup stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that states what is needed to start or how long setup takes.
- Access stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that states how to start: free trial, free plan, or 'request a demo' as the only route.
- Security & data stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that names a certification or states how customer data is handled.
- Proof stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that names a customer with a result, or gives a quantified outcome or independent benchmark (customer counts and logos are not F5).
- Price stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that states a paid price amount, or explicitly states the paid pricing model (e.g. 'custom pricing per seat'). Free plans/trials are F4 only..
- What it is & output stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that names the product together with what it produces or does (one sentence).
- Audience stated early in the text agents read: In the first 15% of the text an agent reads, a body-text sentence that names who it is for.

**Competitors that meet it**

- GC AI (F1): "GC AI takes on the first pass of commercial contract review and prepares targeted redlines against your approved positions, so you can focus on the trade-offs that get agreements signed."
- GC AI (F2): "Help Sales close deals on terms that work for the business."
- GC AI (F4): "Try for Free"

## R-09 · Coverage per content type: Fitting visual for SUM; Fitting visual for FEAT\_B; Fitting visual for UQ

Type: page-fixable · effort M · predicted +0.7 overall (Multimedia +6.7)

**Why**

- MM1:SUM scores 0.5 — Section 4 contains structured feature breakdowns, loosely illustrated.
- MM1:FEAT\_B scores 0.5 — Enterprise security section 9 highlights SOC 2 Type II and safeguards textually.
- MM1:UQ scores 0.5 — FAQ section answers implementation and setup questions textually.

**Change**

- Fitting visual for SUM: A meaningful visual that fits this content type in a section tagged with it.
- Fitting visual for FEAT\_B: A meaningful visual that fits this content type in a section tagged with it.
- Fitting visual for UQ: A meaningful visual that fits this content type in a section tagged with it.

## R-10 · Problem & output: Problem- or outcome-led headline

Type: page-fixable · effort S · predicted +0.3 overall (Content +0.8)

**Why**

- PO-1 scores 0.5 — Catches off-market terms before negotiation expresses a clear outcome/problem, but lacks a quantified/measurable result, qualifying for 0.5. (now: "AI contract review in Microsoft Word that catches off-market terms before you negotiate")

**Change**

- Problem- or outcome-led headline: The H1 names a problem the buyer has or a measurable result they get. Qualifiers score 0 (where it runs, 'trusted', 'AI-powered', 'smart'); a soft outcome without a measurable result scores 0.5; category plus qualifiers only = 0.

## R-11 · Features · buyers: Integrations

Type: page-fixable · effort S · predicted +0.3 overall (Content +0.8)

**Why**

- FB-2 scores 0.5 — Mentions Google sign up and notes iManage/NetDocuments require contacting them; primary environment is Word/Web. (now: "Gavel Exec works inside Microsoft Word (desktop, Word on the web and Microsoft 365) and on the web, and you can sign up with Google. Using a document management")

**Change**

- Integrations: Lists integrations beyond the primary environment (CRM, storage, identity, e-signature…). Primary environment only = 0.5.

**Competitors that meet it**

- GC AI (FB-2): "when Docusign CLM and Intake agent are enabled for your GC AI organization. An administrator can configure the intake workflow, routing, and playbook review for a designated CLM folder."

## R-12 · Transaction reachability: Reachable transaction: buy; Reachable transaction: ask

Type: page-fixable · effort S · predicted +0.3 overall (Layout & navigation +2.1)

**Why**

- LN5:buy scores 0.5
- LN5:ask scores 0.8

**Change**

- Reachable transaction: buy: A visible CTA, link or widget on the page that starts this transaction.
- Reachable transaction: ask: A visible CTA, link or widget on the page that starts this transaction.

## R-13 · Machine-readable signals: Markdown on request; Markdown complete; llms.txt (+1 more)

Type: site-level · effort S · predicted +0.3 overall (Agentic best practices +2.3)

**Why**

- MR-8 scores 0 (now: "Accept: text/markdown → text/html; charset=utf-8")
- MR-9 scores 0 (now: "word ratio n/a")
- MR-10 scores 0 (now: "/llms.txt 404")
- MR-11 scores 0 (now: "not linked")

**Change**

- Markdown on request: Accept: text/markdown returns text/markdown.
- Markdown complete: The markdown keeps ≥ 80% of the HTML's visible words under their headings (0 if MR-8 fails).
- llms.txt: /llms.txt returns 200 with content.
- Page in llms.txt: /llms.txt links this page (or its markdown version).

**Competitors that meet it**

- GC AI (MR-10): "/llms.txt 200"

## R-14 · URL keywords: Descriptive URL, not brand-only; Tier A keyword in URL; Second relevant keyword in URL

Type: site-level · effort M · predicted +0.5 overall (Agentic best practices +3)

**Why**

- UK-1 scores 0 — Last segment is brand-only.
- UK-2 scores 0 — matched none
- UK-3 scores 0 — matched none

**Change**

- Descriptive URL, not brand-only: The path contains a word describing the product's function, object or audience and its last segment is not only brand tokens.
- Tier A keyword in URL: The path contains a Tier A category term (partial = 0.5).
- Second relevant keyword in URL: The path contains a Tier B term or a second distinct Tier A term.

## R-15 · Chunk quality: Self-contained section text (2:CH-1); Self-contained section text (2:CH-2); Self-contained section text (3:CH-1) (+1 more)

Type: page-fixable · effort M · predicted +0.2 overall (Agentic best practices +1.4)

**Why**

- AG4:2:CH-1 scores 0
- AG4:2:CH-2 scores 0
- AG4:3:CH-1 scores 0
- AG4:3:CH-2 scores 0

**Change**

- Self-contained section text (2:CH-1): The section names its subject, doesn't open with a dangling reference, labels example data, and gives units/baselines for numbers.
- Self-contained section text (2:CH-2): The section names its subject, doesn't open with a dangling reference, labels example data, and gives units/baselines for numbers.
- Self-contained section text (3:CH-1): The section names its subject, doesn't open with a dangling reference, labels example data, and gives units/baselines for numbers.
- Self-contained section text (3:CH-2): The section names its subject, doesn't open with a dangling reference, labels example data, and gives units/baselines for numbers.

## R-16 · Illustrativeness: V19

Type: page-fixable · effort S · predicted +0.1 overall (Multimedia +1.3)

**Why**

- MM2:V19 scores 0.5 — Product visual rated 50; target: 75 (clearly shows what the feature does with legible, realistic content). (now: "Three reviewed agreements, including a Series A preferred stock purchase agreement, each marked with a check")

**Change**

- V19: Product visual rated 50; target: 75 (clearly shows what the feature does with legible, realistic content).

## Limitations (a page change can't fix these)

- **Demo — needs a product or sales-flow change** (product): No interactive sandbox or demo on page or 1-click away without signing up (Try free links to account creation). No interactive sandbox or demo on page or 1-click away without signing up (Try free links to account creation). Maximum gain if fully fixed: +0.4 overall (estimate).
- **Demo media — needs a product or sales-flow change** (product): No interactive demo usable without login; target: one on the page (50 points). Needs a product-side demo or sandbox. No interactive demo usable without login; target: one on the page (50 points). Needs a product-side demo or sandbox. Maximum gain if fully fixed: +0.8 overall (estimate).
- **Use (try / sign up) — needs a product or sales-flow change** (product): Use (try / sign up) is scored from the sign-up flow: environments, clicks, verification, installs. A page can only improve it by adding a no-login sample. Use (try / sign up) 47; target: the advertised feature is usable on the page without verification (U = 100). Current reading: Try Gavel Exec free → email\_or\_verification\_asked; 2 environments (+0 known), 1 clicks (+0 known) Maximum gain if fully fixed: +3.2 overall (estimate).
- **Buy — needs a product or sales-flow change** (product): Buy is scored from the checkout path: a plan bought on the page scores 100, a reachable checkout 70 minus steps, price shown without checkout 35, no public price 15. Buy 25; target: a plan can be bought on the page (B = 100). Current reading: price on page true, checkout reachable false Maximum gain if fully fixed: +2.3 overall (estimate).
- **See a demo — needs a product or sales-flow change** (product): See a demo is scored from the booking flow: time slots visible before personal data, fields before slots, leaving the page. Without a visible scheduler it is capped at 35. See a demo 35; target: an on-page scheduler with slots before personal data, earliest slot today/tomorrow (D = 100). Current reading: Request a demo → form\_personal\_data\_asked; scheduler none Maximum gain if fully fixed: +2 overall (estimate).
- **Ask a question — needs a product or sales-flow change** (product): Ask a question is scored by the best contact option: live chat with an online indicator 100 … email only 30. Ask a question 40; target: live human chat with an online indicator (Q = 100). Current reading: phone: (415) 941-6866 shown on the page and clickable tel link. Maximum gain if fully fixed: +1.2 overall (estimate).
- **Talk to a human — needs a product or sales-flow change** (product): Talk to a human is scored by the best option: same-day call or live human chat 100, clickable phone with hours 90 … email only 30. Talk to a human 65; target: a same-day call bookable on the page or live human chat (H = 100). Current reading: phone\_no\_hours: Phone number (415) 941-6866 is provided with 'Call us, 7 days a week' (spec notes days are not hours, so phone\_no\_hours). Maximum gain if fully fixed: +0.7 overall (estimate).

## Coverage ledger

| Component | Score | Status | Refs | Note |
|---|---|---|---|---|
| content · PO Problem & output | 91.7 | recommendation | R-10 |  |
| content · SUM Summary | 100.0 | met |  |  |
| content · DEMO Demo | 85.7 | limitation | L-DEMO-5 |  |
| content · PROOF Trust signals & proof | 100.0 | met |  |  |
| content · FEAT\_U Features · users | 100.0 | met |  |  |
| content · FEAT\_B Features · buyers | 91.7 | recommendation | R-11 |  |
| content · PAF Pricing, access & fit | 100.0 | met |  |  |
| content · BQ Buyer questions | 80.0 | recommendation | R-06 |  |
| content · UQ User questions | 100.0 | met |  |  |
| conversion · U Use (try / sign up) | 47.0 | limitation | L-U-flow |  |
| conversion · B Buy | 25.0 | limitation | L-B-flow |  |
| conversion · D See a demo | 35.0 | limitation | L-D-flow |  |
| conversion · Q Ask a question | 40.0 | limitation | L-Q-flow |  |
| conversion · H Talk to a human | 65.0 | limitation | L-H-flow |  |
| conversion · C CTA placement | 100.0 | met |  |  |
| layout\_navigation · LN1 Coverage | 100.0 | met |  |  |
| layout\_navigation · LN2 Buyer order & first screen | 70.8 | recommendation | R-05 |  |
| layout\_navigation · LN3 Learn-more paths | 100.0 | met |  |  |
| layout\_navigation · LN4 Feature discovery | 100.0 | met |  |  |
| layout\_navigation · LN5 Transaction reachability | 86.0 | recommendation | R-12 |  |
| layout\_navigation · LN6 In-page navigation | 100.0 | met |  |  |
| agentic · AG1 Priority order | 69.7 | recommendation | R-08 |  |
| agentic · AG2 Buyer questions answered in text | 100.0 | met |  |  |
| agentic · AG3 Text-parseable visuals | 100.0 | met |  |  |
| agentic · AG4 Chunk quality | 80.0 | recommendation | R-15 |  |
| agentic · AG5 Machine-readable signals | 71.4 | recommendation | R-13 |  |
| agentic · AG6 AI readiness (ora.ai) | 71.0 | recommendation | R-04 |  |
| agentic · AG7 URL keywords | 25.0 | recommendation | R-14 |  |
| multimedia · MM1 Coverage per content type | 83.3 | recommendation | R-09 |  |
| multimedia · MM2 Illustrativeness | 70.8 | recommendation | R-16 |  |
| multimedia · MM3 Format variety | 60.0 | recommendation | R-07 |  |
| multimedia · MM4 Demo media | 50.0 | limitation | L-MM4-interactive |  |
| performance · performance Performance | 86.5 | recommendation | R-03 |  |