TeamWhat We DoCase StudiesToolsBook a consultation →
CASE STUDY. KNOX SYSTEMS · FEDRAMP AUTOMATION. B2B · SEO + GEO

Searched.Answered.

A FedRAMP compliance-automation platform, indexed for the old web and structured for the new one. This is how iExcel made Knox Systems visible in both search and AI answers.

Engagement
SEO + GEO · Multi-phase
Scope
Audit · Keywords · On-page · Schema
Buyer
SaaS pursuing federal authorization
Signal
AI Overviews · Traditional SERP
Secure data center — cloud compliance infrastructure
// SECURE INFRASTRUCTURE · AUTHORIZATION-READY
// The Squint Test · 2 receipts
86%
Net-New Ranking Footprint
86% of the tracked keyword positions were entirely new against the pre-engagement baseline — unranked, then ranked.
20%
Page-One Visibility · Share of Tracked Keywords
One in five tracked keywords reached the first page of results against the program's own baseline.
// The Stack

Mainstream tools. Rigorous execution.

Every workstream on the Knox engagement ran through platforms the team already trusted. No proprietary black boxes — the stack Knox saw is the stack Knox can run against next quarter.

SEMrushTechnical audit
Keyword universe
Search ConsoleCrawl coverage
Query verification
Google Analytics 4Traffic + engagement
Baseline reporting
Business SchemaEntity markup
Site-wide
FAQ SchemaQ&A markup
Blog-wide
Breadcrumb SchemaHierarchy markup
Blog-wide
Article SchemaLong-form markup
Citation-ready
HubSpotCMS + workflow
Publishing plumbing
AI OverviewAnswer-engine surface
Coverage tracked
Chapter 01 · The Setup

Federal market. Consumer search problem.

FedRAMP is the door. A SaaS company that wants to sell into the U.S. federal government — Department of Defense, civilian agencies, the entire GovCloud footprint — needs FedRAMP authorization to get past procurement. The process is technical, expensive, and long. It is also the single fastest-growing search category in cloud compliance.

Knox Systems builds into that gap. A compliance-automation platform that shortens the authorization runway for SaaS companies pursuing FedRAMP status. The product is real. The buyer is real. The traffic engine that connects them wasn't.

When iExcel picked up the engagement, the pattern was clean. Strong technical audience. Real product depth. And a search surface that was leaking on all three fronts — site health errors that kept the crawler working harder than it should, metadata that didn't map to how buyers actually query, and zero schema instrumentation for the AI-answer layer that had quietly become the second front of federal-buyer research.

U.S. Capitol columns — federal buyer context
Chapter 01 · The federal channel
Chapter 02 · The Audit

Site health. Backlink gap. Baseline.

The engagement opened with a full SEMrush technical audit — crawl coverage, sitemap integrity, canonical logic, indexability, page-speed signals — cross-referenced against Google Search Console for what the crawler was actually reporting back. Then a backlink-and-competitor-keyword pass: which FedRAMP-adjacent domains were consolidating referring domains, which query clusters competitors were ranking on that Knox wasn't showing up for, and where the defensible click-share sat inside the existing footprint.

The output wasn't a slide deck. It was a ranked, dated, ownership-assigned backlog: what to fix, what to write, what to schema, what to leave alone. Every item pinned to a workstream — dev checklist, metadata, on-page copy, blog reoptimization, schema deployment — and paced against a launch window the team could hit.

The audit set the baseline. Everything after Chapter 02 is what happened when the backlog started clearing.

Analytics dashboard — audit and reporting context
Chapter 02 · The baseline read
Chapter 03 · The Keyword Universe

141 rows. One theme.

The keyword research collapsed the FedRAMP category into a single working universe: 141 rows, roughly 51,000 monthly searches in aggregate, 55 of them already triggering an AI Overview in the SERP. That last number is the tell. More than a third of the theme was being answered by a generative summary before the reader ever clicked a blue link — which meant the on-page work had to do two jobs at once. Rank the page. Feed the answer.

The top of the list looked like the category itself. The head term. The marketplace query. The definitional queries a buyer types on the way in. The certification and compliance queries a buyer types once they know they need it.

fedramp
18,100 / mo
// AI OVERVIEW
fedramp marketplace
8,100 / mo
// AI OVERVIEW
what is fedramp
2,400 / mo
// AI OVERVIEW
fedramp certification
1,900 / mo
// AI OVERVIEW
fedramp compliance
1,000 / mo
// AI OVERVIEW
141 total rows · ~51,120 aggregate monthly volumeAI Overview triggered on 55 keywords
55 / 141
Working universe · Triggering AI Overviews
Roughly 39% of the theme was already served by a generative summary before the reader ever clicked a blue link. The on-page work had to earn the click and the citation in the same pass.
Cloud infrastructure — data center scale
// THE NEW SEARCH ERA
"The buyer types the query into two search boxes now. The page has to answer both."
— Engagement Recap · Knox Systems × iExcel
Chapter 04 · The Coverage Board

Five lanes. One launch window.

The work ran on five parallel lanes. Every lane had an owner, a checklist, and a status pill that only moved to LIVE when the change was pushed to production and verified in Search Console. This is the board the team ran against — pinned in the shared workspace, updated on every merge, closed out on 11/13/2025.

SEO + GEO Coverage Board
SNAPSHOT · 11/13/2025 · ALL LANES CLEARED
// LANE 01 · TITLES
42FIXED
Duplicate title tags and H1/title mismatches across the working page set — rewritten to unique, buyer-query aligned patterns.
Cleared
// LANE 02 · METAS
57WRITTEN
Missing meta descriptions authored against the FedRAMP query universe. TOFU, MOFU, and BOFU intent split explicitly.
Cleared
// LANE 03 · SCHEMA
4TYPES
Business, FAQ, Breadcrumb, and Article schema deployed site-wide. GEO-readiness prerequisite for the generative-answer layer.
Live
// LANE 04 · BLOGS
AllGEO-READY
Key Highlights sections, FAQ blocks, and structured intros re-cut so every long-form asset is legible to both readers and answer engines.
Live
// LANE 05 · AI OVERVIEWS
55COVERED
Every keyword in the working universe that already triggers an AI Overview now has an on-page target built for citation.
Live
+32
Metadata updates approved · Then implemented
Of 56 pages reviewed, 34 were optimization-ready. 32 approved rewrites shipped on 11/13/2025 — 16 TOFU, 10 MOFU, 6 BOFU.
Chapter 05 · Dev Checklist

Every error had a count.

The dev-SEO checklist was the least glamorous part of the engagement and the most consequential. Site-health signals compound. A duplicate title on a hub page doesn't just hurt the hub — it dilutes internal linking, splits ranking equity, and confuses the crawl budget on every URL that pointed into it. The audit surfaced five error classes with counts attached, ordered by blast radius.

Every one of them was cleared by 11/13/2025. The checklist below is the before-and-after Knox's dev team was working against — the same report we handed over on kickoff, marked closed on the same run.

Error class
Before
After
Missing meta descriptionsDiscovered by crawl audit
57
0
Duplicate H1 / title tagsRanking-equity drain
42
0
Title tags too longSERP truncation risk
20
0
Title tags too shortMissed keyword opportunity
14
0
Pages with multiple H1sSemantic hierarchy break
4
0
0
Outstanding checklist errors · Post 11/13/2025
Five error classes, all cleared in the same release window. Scope claim, not ranking claim — the crawler now sees a clean site.
Engineer workstation — multi-monitor dev environment
Chapter 05 · Errors cleared
// Ranking Footprint · Top-Three Share
9%

Nine keywords in a hundred. Top three.

Of the tracked ranking positions in the position-change export, 9% landed in the top three results against the program's own pre-engagement baseline — the same technical, metadata, and schema work that cleared the crawler also moved the needle where buyers actually click.

Chapter 06 · Blog + Schema Sweep

Content the reader keeps. Structure the machine reads.

The blog reoptimization ran on a simple rule: the reader still comes first, and the answer engine reads exactly what the reader reads. That meant Key Highlights sections at the top of every long-form post — the same summary a reader skims and the same summary an AI Overview cites. FAQ blocks at the bottom, written against the queries the keyword universe surfaced, matched to FAQ schema so the answer engine could ingest them without guessing.

Site-wide, four schema types went in — the GEO-readiness prerequisites. Business schema so the brand entity is machine-legible. FAQ schema on every post that carried a Q&A. Breadcrumb schema across the blog so the crawler and the answer engine both understand the hierarchy. Article schema on long-form so citations round-trip correctly.

None of it changed how the pages read to a human. All of it changed how the pages read to everything else.

// Site-wide

Business Schema

Organization identity, contact points, and service graph — deployed on every page so the brand entity is unambiguous to the answer engine.

// Blogs

FAQ Schema

Every blog FAQ block matched to structured markup. The reader sees the Q&A. The answer engine sees the JSON-LD underneath it.

// Blogs

Breadcrumb Schema

Hierarchy exposed for every long-form URL — cleaner crawl paths, cleaner SERP breadcrumbs, cleaner citation surface for AI Overviews.

// Long-form

Article Schema

Author, publish date, and body structure marked up so the piece can be cited as an article rather than as a generic web page.

Code and encryption abstract — schema layer
Chapter 06 · The schema layer
Cybersecurity abstract — encryption context
// THE COMPOUND
"Traditional search and AI answers run on the same underlying craft. Do the craft, and both surfaces reward you."
— Engagement Recap · Knox Systems × iExcel
Chapter 07 · The Read

SEO. And GEO.

If you sell into a regulated market — FedRAMP, HIPAA, SOC 2, ISO 27001, PCI, StateRAMP — and your buyer's research motion starts with a query typed into a Google box and a generative answer engine, the Knox engagement is the shape of what to do about it.

Audit the site health that the crawler is quietly downgrading you for. Rewrite the metadata so the pages match how the buyer actually asks. Deploy the four schema types that make the brand, the FAQs, the hierarchy, and the long-form legible to the answer engine. Instrument the AI-Overview surface with the same discipline you apply to the blue-link surface.

The position-change export told the same story from a different angle. Against the pre-engagement baseline, 86% of the tracked keyword positions were entirely new — pages that didn't rank at all, now indexed and ranked. 20% of that footprint reached page one; roughly 9% broke into the top three. None of it happens without the site-health, metadata, and schema work landing first.

Most agencies still sell one search era at a time. The buyer is already living in both.

86%
Ranking Footprint · Net-New vs. Baseline
Of the tracked keyword positions in the position-change export, 86% had no prior ranking at all — the technical and structural work turned invisible pages into indexed ones.
Security lock — federal buyer context
Chapter 07 · The read

Two search boxes. One craft.

If your buyer starts research in Google and finishes it in an AI answer, your on-page needs to survive both surfaces. Book a 15-minute sanity check on your setup — we'll tell you where the traditional-search work is already covering the GEO work, and where it isn't.

Book a 15-min sanity check →
Book →