Backed by programs from

Our announcements

A self-improving AI front office

The front office that improves itself.

Our RSI agent workflow runs your front office — marketing, sales, and customer service — for a roofing crew or a software team. Each loop learns from its own results, and what one learns, the others use. You approve what it keeps.

RSI Agentcycle 07
Sample cycle
  1. RunGTMLinkedIn series + 3 GEO fixes shipped
  2. MeasureCustomerCallers keep asking about weekend pricing
  3. ReviewSalesSame question stalls 2 open deals
  4. CommitAll loopsOne pricing answer → site, front desk, follow-up
Cost per accepted asset · illustrative↘ compounding

§ 1Thesis

Why Florus exists, and the number that would prove us wrong.

Recursive self‑improvement will define the AI era. We point it at go‑to‑market, sales, and service.

Why these three loops

  1. Recurring work. Campaigns, calls, and follow-ups repeat, so lessons get reused.

  2. Fast feedback. Replies, meetings, and bookings arrive in days, not quarters.

  3. Reviewable output. Every asset, call, and booking can be accepted or rejected — auditable.

  4. Shared customers. The same people move through all three loops, so a lesson travels.

The claim we’ll be held to

“Same budget, more accepted work — every cycle.”
Accepted assets per dollar
The headline number.
Repeat-defect rate
Does a fixed mistake stay fixed?
Asset reuse
Are playbooks actually being used?
Cross-loop reuse
Does a lesson from one loop get used in another?
Human override rate
How often you have to step in.
Governance alpha hypothesisIllustrative only. Cost per accepted asset stays flat or rises for ungoverned agents and falls across cycles for a governed RSI loop.Cost per accepted asset, by cycle12345678
  • Ungoverned agents
  • Governed RSI loop
  • Governance alpha — the gap between them
Fig. 1Cost per accepted asset across eight cycles. A hypothesis we are instrumenting, not a reported result.

§ 2The RSI loop

Every cycle ends smarter.

GTM, sales, and service each run the same cycle — and keep what worked once you approve.

  1. 01RunShip campaigns, make calls, answer customers.
  2. 02MeasurePull back replies, meetings, bookings, and reviews.
  3. 03ReviewScore results. Find the root cause.
  4. 04CommitTurn lessons into assets — with your approval.
  5. 05CompoundStart the next cycle — in every loop — from what was learned.
Explore the RSI Agent
The RSI loop: run, measure, review, commit, compound — then the next cycle starts with what was learned.cycle n → n+1starts smarter01 Run02 Measure03 Review04 Commit05 Compound

§ 3Connected loops

Three loops. One system that learns from all of them.

Each loop gets better on its own. Connected, a lesson learned in one becomes the starting point for the other two.

Three RSI loops — GTM, sales, and customer — each running its own cycle, connected by six handoffs and a shared memory at the centre.123456SharedmemoryGTMLiveSalesEarly accessCustomerEarly access
Fig. 2Each circle runs its own run, measure, review, commit cycle. Numbered arrows are handoffs; the centre is the memory all three read from and write to.

What travels between loops

  1. 1GTM → SalesMessages that earn replies become call openers.
  2. 2Sales → GTMObjections heard on calls become content and FAQs.
  3. 3Sales → CustomerWhat was promised in the deal briefs the front desk.
  4. 4Customer → SalesRenewals, upgrades, and referrals become pipeline.
  5. 5Customer → GTMQuestions customers actually ask become pages AI engines cite.
  6. 6GTM → CustomerLive campaigns tell the front desk who is calling, and why.

The three loops

Loop
GTM
Live
Runs
Campaigns, posts, outreach, and AI-visibility fixes.
Measures
Citations, replies, and edits.
Keeps
Hooks, positioning, and content playbooks.
Loop
Sales
Early access
Runs
Prospect research, calls, sequences, and deal follow-up.
Measures
Connects, meetings booked, and objections heard.
Keeps
Talk tracks, objection handling, and buying signals.
Loop
Customer
Early access
Runs
An AI front desk that answers calls and chat, books, reminds, and follows up.
Measures
Missed calls, bookings, resolutions, and reviews.
Keeps
Answers, booking rules, and escalation paths.

The GTM loop is live. Sales and customer loops are in early access with design partners.

Join early access

§ 4What compounds

What it keeps.

Only what pays for itself.

Memory
Market memory
ICP, voice, objections, and promises — shared by every loop.
Skills
Playbooks
Winning hooks, talk tracks, and booking scripts, made reusable.
Harness
Quality checks
Repeated mistakes become pre-ship checks.
Routing
Routing policy
The right loop, channel, and model for each job.

Governed, not autopilot

Self-improving, never self-approving.

  • Risk-tiered review

    Brand, claims, pricing, and spend need a human.

  • Full ledger

    Every change is recorded with evidence.

  • Rollback

    Any asset can be retired when it stops paying.

§ 5Products

One system, three ways in.

Signal, execution, and the workflow that learns across GTM, sales, and service.

Learning loopDemo-led · Custom deployment

RSI Agent

Self-improving AI front office

The front office that improves itself.

Built on the NVIDIA DGX Spark champion agent team

The RSI loop: run, measure, review, commit, compound — then the next cycle starts with what was learned.cycle n → n+1starts smarter01 Run02 Measure03 Review04 Commit05 Compound
ExecutionLive · Self-serve

Launchvibes

Creator growth engine

Hooks and campaigns for founder-led growth.

Building in public? Join the Launchvibes creator cohort.

Apply to Creator Growth

§ 6Client proof

GTM systems for teams that need to be found, understood, and remembered.

From GTM, messaging, and content work.

01 · GTM proofHoneypot Finance
Florus helped turn complex product and launch ideas into clearer GTM messaging, partner-facing materials, and content direction.

The work focused on making a technical product easier to explain across launch, partner, and content surfaces.

02 · GTM proofSplatter Protocol
Florus helped package a technical protocol into a clearer market story, stronger website/product messaging, and a more repeatable GTM workflow.

The work connected protocol positioning, website clarity, and repeatable GTM execution.

03 · GTM proofToro / torodex.xyz
Florus helped sharpen the product narrative and identify what needed to be communicated more clearly before scaling GTM.

The work focused on clarifying the product story and surfacing the gaps that needed to be fixed before broader distribution.

04 · GTM proof1DAO
Florus helped clarify the vision, PMF narrative, and content priorities so the team could communicate the product more effectively.

The work helped turn early product direction into clearer GTM priorities and communication surfaces.

§ 8Careers

Help build GTM systems that compound.

A few focused roles. Direct contact.