● Backbone
Internal · 2026
Backbone
Velocity is the
new product-market fit.
Document v1
Manifesto
+ Grand Execution Plan
90 days
Part one
An internal call to arms.
01
Section

The
Manifesto.

Six convictions about how Backbone builds, ships, and wins from this day forward.
● Backbone — Manifesto
01 / The world changed

Product-
market fit
is dead.

Or rather —

It's no longer enough. Finding the fit used to be the moat. Today, anyone with a laptop and an AI agent can find a fit by Friday.

What separates winners from corpses is not what you build — it's how fast you ship, learn, and ship again.

BACKBONE
03 / 17
● Backbone — Manifesto
A new fit
“

Velocity-fit
is the new
product-market fit.

Speed of iteration · the only durable advantage
BACKBONE
04 / 17
● Backbone — Manifesto
02 / The bet

AI is the
operating system
of how we work.

Not a tool. Not an experiment. Not a Q3 initiative.
Every team. Every function. Every day. This is the company now.

Two choices.
01 Ride the wave.
02 Get drowned by it.
We're choosing the wave.
BACKBONE
05 / 17
● Backbone — Manifesto
03 / The new shape of work

Everyone
is a builder.

The handover is dead — in both directions. If you work at Backbone, you sit at the front of the value chain. You understand the customer pain, you shape the solution, you own the outcome.

Sales
Spot friction. Prototype the fix in Claude Code. Ship value this week — not in six.
Ops
Tired of a manual workflow? Build the agent yourself. Don't beg for bandwidth.
Engineering
The ticket-taking era is over. You're not a task executant — you're a problem owner.
BACKBONE
06 / 17
● Backbone — Manifesto
04 / Org & authority

Drop the
chain of
command.

Inside the product, there is no hierarchy.
There is the customer, the problem,
and whoever can solve it fastest.

Titles still exist — for clarity, comp, accountability. They are not permission slips. They are not gates.

✕Before · the relay

Twenty handovers,
twenty losses of signal.

Director approves
PM scopes
Designer mocks
Eng builds
QA tests
Sales sells
Each step adds latency. Each handover loses signal.
●Now · horizontal

One builder.
One customer.
One loop.

Builder
Customer
Prototype
Ship
Measure
Iterate
The best idea wins — not the most senior. The fastest prototype gets the feedback — not the most political.
BACKBONE
07 / 17
● Backbone — Manifesto
05 / How we decide

The arbiter
is the data.

"Best idea wins" on the whiteboard isn't the same as in production. We are not running a debate club.

Shipping isn't the victory. Customers using what you shipped — that's the victory.

Signal
Verdict
Action
Real usage, growing
it works
Double down →
Real usage, flat
it's stuck
Iterate hard →
No usage, dead
it didn't land
Kill it →
No shame in killing what didn't land. The shame is leaving it limping in production because someone's ego is attached to it.
BACKBONE
08 / 17
● Backbone — Manifesto
06 / Name the elephant

Fear is the
enemy.

Not budget. Not tools.
The reason most companies will fail this transition is fear — and fear never built anything.

The fears we kill, this year
— Fear of breaking something
— Fear of looking stupid
— Fear of not knowing the syntax
— Fear of working differently next year
— Fear of being replaced
— Fear that what you mastered is now table stakes
We don't fight fear with speeches.
We fight it with reps — build something small. Break something. Fix it. Repeat.
BACKBONE
09 / 17
● Backbone — Manifesto
07 / Amplification, not replacement

AI doesn't
replace you.
It amplifies you.

The leverage goes to the human who learns to direct the tool — not the one who waits for someone else to do it.

1→5
Proposals per week.
From one a week to five — same person, same job.
1→4
Features shipped per month.
Monthly cadence becomes weekly. Compounding from day one.
BACKBONE
10 / 17
● Backbone — Manifesto
08 / Why now, not in six months
80%

Of Backbone.
Not ten.

We want to decentralize creation power to 80% of Backbone — not the 10% who happen to write code today.

In six months, our competitors will have done it. The compounding curve on AI productivity is brutal. Every week you wait, the gap widens.

10–20×
Expected throughput lift if we pull this off. Not 2×. Ten. Maybe twenty.
BACKBONE
11 / 17
Part two
A manifesto without execution is a poster.
02
Section

The Grand
Execution
Plan.

The stack, the principles, the 90-day march, and how we'll know it's working.
● Backbone — Execution
01 / How a Backbone builder ships

The Stack.

Every builder — sales, ops, marketing, support, product, engineering — operates on the same pipeline. Every layer is a built-in safeguard so you can keep your attention on the only thing that matters: the customer problem you're solving.

01
Builder
Anyone, any function. The person closest to the customer leads the iteration.
02
Claude Code · Nexus
Your AI pair-builder. Describe what you want, iterate fast, review every output.
03
Claude Design
Design system baked in. Ship interfaces that look like Backbone — no designer required.
04
GitHub repo
Version, review, gate. Code lives in source control, not in someone's Notes app.
05
Vercel + GCP
Vercel for every dev preview — push, get a URL. Production ships on Vercel and GCP.
06
Embedded metrics
Telemetry from day 1. No feature reaches production without instrumentation.
AI Governance
— wraps the entire pipeline
Security
Quality gates
Observability
No shadow IT
BACKBONE
13 / 17
● Backbone — Execution
02 / How we operate

Six execution
principles.

01

Default to building.

If you find yourself writing a ticket, ask first: could I prototype this myself in 30 minutes with Claude Code? Tickets are the last resort, not the first move.

02

Ship small,
ship often.

A working prototype in production beats a polished plan in a doc. Iterate after launch, not before.

03

Pair with AI like
a senior engineer.

Describe context. Push back. Ask for alternatives. Review every output. AI is not a magic box — it's a force multiplier on the human who drives it well.

04

Governance is
non-negotiable.

AI-generated work flows through the governance layer. Speed and safety aren't enemies — bad process is.

05

Show your work.

Every builder demos weekly. Publicly. Inside the company. We are not going to whisper this transformation — we are going to broadcast it.

06

Measure honestly.
Kill without ceremony.

Two weeks after launch — if usage is growing, double down. If it's dead, kill it. Amazon does this ruthlessly. So will we.

BACKBONE
14 / 17
● Backbone — Execution
03 / The 90-day march

The 90-day plan.

Three phases. One direction. Tooling rolls out, then scales, then compounds.

Days 0–30

Ignite.

  • Tooling rollout. Claude Code access for every employee. No approval queue. No gatekeeping.
  • Governance v1. Repos, security baselines, code review gates, data rules — one page, not fifty.
  • Builder onboarding. One-hour kickoff per team. By the end, every attendee has shipped something.
  • Champions. 1–2 AI champions per department. Their job: unblock the team.
  • Weekly Build Demos. Friday, 30 minutes. Whoever shipped this week, shows it.
Days 30–60

Scale.

  • Repo per builder. Every active builder/team gets a dedicated GitHub repo wired to Vercel.
  • Pattern library. Reusable agents, prompts, code patterns. Steal-and-adapt is the default.
  • Cross-department shipping. Sales builds for Ops. Ops builds for Support. The flywheel spins.
  • First metrics. Number of builders, agents in production, idea→ship cycle time. Published weekly.
Days 60–90

Compound.

  • 80% check-in. Measure what share of staff has shipped AI-built work. Fix where it lags.
  • Customer-facing wins. ≥ 5 improvements shipped end-to-end by non-engineers. Tell the story externally.
  • Governance v2. Tighten where it leaked. Loosen where it slowed us.
  • Role redesign. Every job description rewritten around "builder" expectations. Publicly.
BACKBONE
15 / 17
● Backbone — Execution
04 / How we'll know it's working

How we'll know
it's working.

Metric
Today
Day 90 target
Staff who have shipped AI-built work
< 10%
≥ 60% → on the way to 80%
Agents & tools in production
a handful
dozens, owned across departments
Cycle time: idea → production
weeks
days, often hours
Shadow-IT incidents
unknown
zero — everything inside governance
Customer NPS trajectory
current baseline
trending up, driven by shipped value
Shipped features with usage metrics
inconsistent
100% — telemetry by default
Features killed after launch (a healthy number)
≈ 0 — we hold on too long
a regular, public, unsentimental practice
BACKBONE
16 / 17
● Backbone — Closing
— Backbone

The time
is now.

01Velocity-fit is the new PMF.
02Everyone is a builder.
03The org is flat inside the product.
04Fear is the enemy.
05Ownership is the answer.
06AI is the lever.
07The time is now.
Backbone
This week. Open Claude Code and build the smallest useful thing you can imagine in your job. Share it. Be impatient with the work, patient with yourself. Push back — this plan is v1, not the bible.
↑ ↓ to navigate · 17 slides