What does By application mean?

FORKOFF runs 5 engagements per quarter. Every engagement passes a 4-step intake: ICP fit check, founder bandwidth check, lane match, and outcome floor agreement. We turn down briefs that fall outside the lane rather than soften the operating model to fit them.

What is the engagement floor across /for ICPs?

Pilot floor varies by service mix and ICP. Most ICP routes carry a 2k to 5k pilot floor (AEO, GEO, founder funnel). Foundation Model and pre-TGE pilots run higher because launch-arc work covers more ground. None of the /for ICP routes carry the 0.003 CPQV clipping wedge price by default. Clipping is the specialist product at clips.forkoff.xyz, priced separately.

Talk to FORKOFF. We screen the brief against ICP fit, founder bandwidth, and lane match before quoting an engagement. The first conversation is free; the diagnostic that follows is the first paid step.

ICP hub · 8 cohorts · By application

FORKOFF runs marketing forAI startups.

FORKOFF for ICPs is the cohort-routed marketing engagement for AI and Web3 founders. Eight ICP cohorts on the same 4-Layer Operating Model (Narrative, Long-Form, Clipping-Led Distribution, Ecosystem Access) with different anchors per cohort. Selective on ICP. Five engagements per quarter. Dubai HQ.

Talk to FORKOFF

See the services

By application · 5 engagements per quarterDubai HQ · Operating from DIFCAI plus Web3 native · Same operator across cohorts

9ICP profiles served

By applicationSelective on ICP

4-LayerOperating model

Dubai HQOperating from DIFC

Eight ICP cohorts

Pick the cohort.
FORKOFF runs the cadence.

Each cohort routes into a tuned variant of the 4-Layer Operating Model. Same operator, same dashboard, different anchors and receipts per cohort.

AI startups

Company axis · Pre-Seed → Series B

AI-native company cohort. AEO citation across ChatGPT and Gemini, founder-led distribution, audit-ledger receipts. The largest /for cohort.

Open cohort

AI agents

Product axis · Agent SDK + vertical agent builders

Marketing for vertical SDR, coding, ops, voice, and multi-agent platforms. AEO + GEO citation, DevRel motion, audit-ledger per dollar.

Open cohort

Foundation models

Model-lab axis · Frontier + open-weights labs

Pre-launch cluster activation, narrative system, eval-page corpus, generative-engine citation. Built for the launch arc model labs run.

Open cohort

SaaS companies

B2B SaaS axis · PLG · sales-assisted · vertical · enterprise

Founder-funnel cadence, channel marketing on the operator spine, GEO surface for category-shortlist queries, qualified-inbound ledger.

Open cohort

Web3 protocols

Protocol axis · L1 · L2 · appchain · mainnet protocols

Cluster activation across crypto-twitter and Telegram, KOL coordination, ecosystem access at ETHCC and Token2049, audit-ledger per arc.

Open cohort

DeFi protocols

DeFi axis · DEX · lending · perps · stablecoin protocols

Long-form moments on the protocol thesis, KOL drops with bot-screened roster, founder funnel anchored on TVL receipts and trader recall.

Open cohort

Pre-TGE protocols

Pre-token axis · Mainnet-soon + tokenless protocols

16-week pre-TGE cadence covering narrative spine, KOL warm pool, ecosystem footprint, and the launch-day cluster activation playbook.

Open cohort

Dev tools

Tooling axis · SDKs · libraries · infra primitives

Marketing for dev-tooling teams. Builder trust through DevRel motion, docs polish, AI Engineer Summit access, and the audit-ledger receipts on contributor signal.

Open cohort

DePIN networks

Network axis · DePIN protocols + sensor / compute networks

Cluster activation for DePIN protocols. Operator + supply-side narrative, KOL coordination, ecosystem access on the Solana plus Helium plus IoTeX clusters.

Open cohort

ICP fit · Reject ledger

Five briefs we turn down.
We do not soften the lane.

Selective on ICP is not a slogan. Five engagements per quarter forces every brief through a four-step intake. The patterns below are the ones we route out, not in.

fk_audit · for_reject_log.csv

#Reject reasonAudit detailFORKOFF fix

Row 0101
Reject reasonOutside FORKOFF lane
Audit detail
B2C consumer brands, e-commerce stores, and small-business marketing requests sit outside the AI plus Web3 lane. The 4-Layer Operating Model is built for technical buyers, not retail shoppers.
FORKOFF fix
Fix
Routed to specialist agencies on intake. We do not soften the lane to fit a brief that does not match the operator stack.
Row 0202
Reject reasonNo founder bandwidth
Audit detail
FORKOFF runs founder-led distribution. When the founder cannot give 60 to 90 minutes per week to the cadence, the long-form spine never compiles, and the clipping layer compounds against an empty pool.
FORKOFF fix
Fix
Pilot is gated on founder availability for the spine. If the founder cannot show up for the cadence, we recommend the in-house operator path, not a FORKOFF retainer.
Row 0303
Reject reasonPre-product distribution ask
Audit detail
Pre-product founders asking for paid amplification on a prototype with no usable receipts. Distribution against an empty foundation makes the gap louder, not quieter.
FORKOFF fix
Fix
Marketing Foundation pilot first. Lock positioning, narrative, and one publishable receipt before the distribution layer turns on.
Row 0404
Reject reasonRetainer-only intent
Audit detail
Buyer wants a flat monthly retainer with deliverable activity but no measurable outcome. FORKOFF prices on outcome (qualified view, qualified inbound, citation share), never raw activity.
FORKOFF fix
Fix
Sandbox pilot with a refund floor on the outcome metric. If the floor is not hit, the engagement does not graduate to the embedded retainer.
Row 0505
Reject reasonWrong stage for the wedge
Audit detail
Series C and later companies asking for clipping wedge pricing. The wedge is built for pre-Series-B compounding recall, not late-stage performance scaling.
FORKOFF fix
Fix
Routed to the Fractional CMO operating model or the channel marketing retainer. Clipping wedge stays priced for the cohort it serves.

5 / 5 patterns auditedSource: 200+ FORKOFF intake conversationsPre-engagement diagnostic

Frequently asked questions

Pre-Series-B AI-native company building authority for the next round routes to /for/ai-startups. Vertical agent SDK or product builder routes to /for/ai-agents. Frontier or open-weights model lab routes to /for/foundation-models. B2B SaaS company across PLG, sales-assisted, vertical, or enterprise routes to /for/saas-companies. Mainnet L1 or L2 routes to /for/web3-protocols. DEX, lending, perps, or stablecoin protocol routes to /for/defi-protocols. Pre-TGE protocol routes to /for/pre-tge-protocols. Individual founder pre-fundraise or post-Series-A routes to /for/ai-founders.

L1 Narrative · L2 Long-Form · L3 Clipping-Led Distribution · L4 Ecosystem Access. Same framework, different anchors per ICP. AI startups anchor on AEO citation plus YC plus a16z access. Web3 protocols anchor on cluster activation plus Telegram plus ETHCC. SaaS anchors on category-shortlist queries plus operator-driven inbound. The operating model is constant. The receipts and channels shift per cohort.

Most adjacent profiles route into one of the eight cohorts above. AI founders running individual operator funnels route into /for/ai-startups (axis-merge: individual founders sit inside the AI-native company cohort). YC companies route into /for/ai-startups or /for/saas-companies depending on stack. ICPs outside AI plus Web3 typically fall outside the lane and we route them to specialist agencies rather than weaken the operating model.

Apply for the diagnostic

Pick the ICP.
FORKOFF runs the cadence.

Eight cohorts, one operating model. Talk to FORKOFF for the ICP fit conversation. Five engagements per quarter, Dubai HQ, by application.

Talk to FORKOFF

See the services

From the FORKOFF blog

Receipts, deep dives, and playbooks.

Read all

ETH NYC 2026 Activation Playbook: The 3-Week Motion

ETH NYC 2026 hits June 8 to 14 (ETHConf + ETHGlobal NY). The 3-week activation playbook that lifted sponsor day-30 dev activations 8.9x across 11 audited teams.

By FORKOFF

Read