The lore itself.

Most companies have a narrative document their team rarely opens. We build the engine that puts your voice, positioning, and proof into every AI session your team runs — every Claude, every Cursor, every ChatGPT — loaded on every call. Run as your subscription.

read on
§ 01 / on the rewrite tax

Every founder is paying a rewrite tax they can’t see. The homepage gets quietly updated when it stops sounding right. The board update copy contradicts last quarter’s deck. A new hire pitches the company differently because the wiki page they read is half-true. Marketing rewrites because nothing they inherit is on-brand. Recruiters use phrases that miss the ICP, and candidates self-select wrong.

Most positioning is built once and walks away. Brand agencies look outside — competitors, analysts, market data — and produce a deck. Fractional CMOs work inside — your team, your meetings — and produce a strategy. Both are partial. Both go stale within months. The narrative that holds up is the one that combines all of it — the homepage, the JDs, the founder’s podcasts, the team’s lived experience, the customer’s words — and keeps combining as the company moves.

We build that system, and run it as your subscription. The audit examines every external surface and every internal voice. Katie synthesizes the canonical narrative from all of it. The engine deploys it into every AI tool your team already uses — every Claude, Cursor, and ChatGPT session loaded with your voice on every call, your positioning a slash command away, your proof points one query deep. The loop captures the team’s continuous signal as they work and routes it back through Katie’s review. The narrative stays alive because the system around it is.

The document is the snapshot. The engine and the loop are the product. The loop is the moat — every other firm sells the snapshot.

Not a doc that gets refreshed annually. Infrastructure that runs.

— a working definition

"Lorem ipsum is what happens when you don't have your real words. Lore ipsum is what we build so you do. edition 01 / on the canonical narrative

§ 02
The artifact

One artifact. Ten sections. Built to be read by tools.

The Narrative Architecture is the canonical truth your team's LLMs read every time they generate copy — wired into Slack, Cursor, Claude, Cowork, anywhere they already work.

Ten sections, A through J. Six are universal. Three flex by engagement scope. The tenth — Internal Working Trail — is firm-internal and never reaches your team; it carries the alternatives Katie considered and the evidence behind every call she made. The other nine become the canonical truth.

A · Foundation
Mission, why-we-exist, what-we-do, what-we-believe, origin story.
B · Identity
Category, ICP, wedge, anti-positioning, company descriptions.
C · Market context
Current positioning, industry shifts, two futures, why-now, why-us.
D · How we play
Strategic vision, unfair advantages.
E · Pillars
Three to five, by buying job. Each with claim ladders and proof points.
F · Audiences
Stakeholder-level translations, not segment-level. Typically three.
G · Voice & language
Voice descriptors, vocabulary, banned phrases with reasons, story bank.
H · Customer success
Externally-approved stories with attribution rules.
I · External-ready
Press boilerplate, speaker bios, FAQs, talking points.
J · Working trail
Firm-internal. Reality grid, gap classification, alternatives, stress test results. Yours to audit; never shipped to the team.

Every section carries frontmatter that defines its lifecycle, distribution, and engagement scope. The artifact is what your tools read. The system that keeps it accurate is what you subscribe to — see the loop and the engine, below.

§ 03
The loop

The team's signal, captured continuously

The narrative is wired into your team's existing communication channels — Slack, email, or whatever they already use. Signal at the source. Routed back. Fed into approved updates.

Sales call
A customer asks a question the rep can't cleanly answer. The gap surfaces in the journal, gets reviewed, and either becomes a new objection response, a new banned phrase, or a new customer success story.
Podcast appearance
A team member changes how they pitch a feature. Drift gets caught early — in the team's signal, not in the press six months later when it's already out the door.
Board meeting
The founder makes a positioning shift on the fly. It propagates through the engine within the next sprint, with founder approval on every material change.
New hire
Their first week, they get a personalized kit drawn from the canonical narrative — not a wiki page that's been quietly out of date for two quarters.

Material updates require active founder approval. Refinements pass through a 24-hour window. Additions flow continuously and show up in the changelog. The narrative stays alive because the loop is the architecture, not a feature.

§ 04
Operating principle

Engine surfaces. Katie decides.

The contrarian bet underneath the product.

Most AI-powered narrative tools assume authorship. They ship generic output and call it personalized. Our stance is the opposite. The narrative is human-authored work; the engine is the deployment surface that makes it usable across every output a team generates.

AI never produces canonical narrative output that ships to clients on its own. It surfaces options with evidence — Katie reads the surfacing, Katie makes every call. Her decisions become the canonical narrative.

tension options evidence Katie's call canonical narrative

The differentiator isn't AI sophistication. It's the human judgment underneath, made deployable.

§ 05
How we work

AI-native, not AI-assisted

We don't run this like a traditional consultancy. The firm itself is the same system you're buying.

Every prompt, every skill, every template the firm uses lives in a single git repository — our company OS. Each engagement gets its own private repo, structured the same way. Claude is the interface across both. We built the infrastructure to back it.

What you see
Less overhead. Faster turn. Higher fidelity than a consultancy of comparable depth. A proposal in your inbox before the discovery call ends.
What we keep
Every engagement teaches the system. New tensions surface new skills. New audiences extend the templates. The next client gets the compounding benefit of every prior one.
What's unique
The architecture pattern is emerging across AI-native services firms. What's unique is what we put inside it: Katie's twenty-five years of practice, encoded.
§ 06
The depth

Twenty-five inputs, three phases

Most positioning consultants look at three to five inputs. We look at twenty-five — every one cited, every claim sourced, every decision documented with three or more alternatives considered.

External (firm-internal — before we talk to your team)

Internal (with your founder and team)

Synthesis (Katie, solo)

Total: twenty-five distinct inputs and decisions per engagement. Every one with a path back to source material. The Internal Working Trail is the audit log; you can read it any time.

§ 07
The alternatives

What we replace

Five paths a founder considers. What each one does well, and where it leaves the rewrite tax.

Brand consultancy
$75K – $200K
delivers Full positioning, voice, visual identity
misses Not deployed as infrastructure. Goes stale in six months.
Fractional CMO
$8K – $20K / month
delivers Strategic guidance, execution oversight
misses No canonical artifact. Their voice — not yours — leaks into the company.
Brand agency
$50K – $150K
delivers Visual identity, copy templates
misses Not the strategic narrative. Not in the team's tools.
DIY (founder + ChatGPT)
Hours of founder time
delivers Fastest, cheapest start
misses No canonical artifact. Output drifts daily.
Status quo
Invisible
delivers
misses The rewrite tax. Already happening; uncosted until measured.

We replace the strategic + deployment portion of all five. Standard pricing through the first five engagements is $10K up-front and $3K/month for the full stack. Two-year cumulative cost versus the alternatives runs $80K – $400K for partial coverage. We provide complete coverage at lower cost because the infrastructure pattern carries the leverage.

§ 08
The work

Custom, layered, priced to scope

What you subscribe to is L3 + L4 — the engine running in your team's tools, the loop capturing signal continuously. L1 + L2 are the foundation that produces the canonical baseline they run against.

L1
Foundation
The Narrative Architecture itself — Katie's audit, synthesis, and the canonical artifact your company runs on.
once
L2
Implementation
Markdown deployment so any LLM tool your team uses can read the narrative natively. About 70% of L3's tool behaviors are replicable here with templates and agent prompts.
once
L3
Engine
Live MCP server in your team's tools. The irreplaceable parts: the feedback loop, multi-tool consistency, live updates as Katie approves changes, future-proofing as new tools emerge.
subscription
L4
Loop
Continuous capture of the team's signal — through Slack, email, or whatever channel they already use. Lived experience updates the narrative.
subscription

L1 and L2 are the one-time foundation. L3 and L4 are the ongoing product — engine deployed, loop active, narrative evolving with the company. Pricing is judgment-based and built live in the sales call. You leave the call with a proposal in your inbox — no waiting, no follow-up build.

§ 09
Pricing

Three shapes, intro-priced

Through the first five engagements, prices are below where they'll land. Lock current rates by signing now.

Standard
$10K up-front · $3K / month
L1 + L2 + L3 + L4

The full stack. Annual narrative refresh. Quarterly check-in. External monitoring of public surfaces.

Engine-only
$10K up-front · $300 – $500 / month
L1 + L2 + L3

Architecture, implementation, and engine deployed. No annual refresh, no monthly call. For teams that want the engine self-managed.

Document-only
$15K – $20K one-time
L1 + L2 (+ 6-month engine trial)

The artifact and templates without ongoing engine. Six-month L3 trial included so you can decide whether to subscribe after seeing it run.

Pricing is judgment-based and built in the sales call — these tiers are anchors, not menus. Engagements that need extra audience translations, additional pillars, or accelerated timing flex from here.

§ 10
Discovery

Book a 30-minute call

We'll listen, ask a few sharp questions, configure a proposal live, and you'll have it in your inbox before the call ends.

Or send your team's most recent investor update or homepage hero — we'll send back a one-paragraph read on its narrative shape. No pitch, just our eye.

katie@loreipsum.co →