Fuse AI growth playbook

Regenerate /machine-version from the same source as the pricing page and homepage: fuseai.com links, current data counts, current plans and discount, credit cost per action, and the one-line MCP install. Publish /llms.txt and /llms-full.txt at the paths the page promises. Then add a release check so that any change to pricing or headline numbers updates the machine page in the same edit.

For agent-led buyers the machine page is the pricing page. AI assistants answering "what does Fuse cost?" or "how much data does Fuse have?" draw on text like this. A generous, correct free plan (24,000 credits a year) is a far stronger hook than 2,000 one-time credits, and correct data counts back up the 850M+ claim on every comparison page.

[ ] tryfuse.ai and app.tryfuse.ai links -> fuseai.com and app.fuseai.com
[ ] "800M+ contacts across 20+ data providers" -> 850M+ contacts, 40+ providers
[ ] "Free ($0, 2,000 one-time credits)" -> Free ($0, 24,000 credits a year)
[ ] "annual billing 25% off" -> 20% off
[ ] Enterprise: match the pricing page (RBAC, SSO, audit logging, dedicated GTM engineer support)
[ ] Add credit cost per action
[ ] Add the one-line MCP install
[ ] Publish /llms.txt and /llms-full.txt
[ ] Add to the pricing-change checklist

• AI systems keep quoting the old figures for a while

  Keep the redirects, publish the llms files and recheck answers monthly.

• Numbers drift again after the next pricing change

  Generate the page from the pricing source and add it to the release checklist.

• Your websiteMachine version (/machine-version, reached from the homepage's Human/Machine toggle, published Nov 7, 2026): an llms.txt-style index for AI agents whose links all point to tryfuse.ai and app.tryfuse.ai (tryfuse.ai now 301-redirects to fuseai.com); it says "800M+ contacts across 20+ data providers", lists "Free ($0, 2,000 one-time credits)", Enterprise with "self-hosting" and "annual billing 25% off", and points to an llms-full.txt. fuseai.com/llms.txt and fuseai.com/llms-full.txt both return 404 (checked November 9, 2026).Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source
• Your websiteMCP page: Fuse works with popular AI assistants, coding agents "and other MCP clients"; setup is one line to paste into the agent, "Install and set up https://docs.fuseai.com/skill.md", with "no API key to paste"; MCP calls draw on the same credits as the app and "each call needs a balance of at least 100 credits"; tools are labelled "read-only or destructive" so the agent can ask before acting; the demos are labelled "Illustrative demo with fictional companies and people"; there is no customer proof.Source

Add a "Give this to your agent" block, with the one-line install and the free plan's 24,000 credits a year, in the middle and at the end of every agent-workflow post. Rewrite each post's main example as a prompt that runs with Fuse's tools, and keep the demo link as a secondary option for teams.

The reader is already in a do-it-yourself frame of mind. Replacing a meeting with a one-line action shortens the path from reading to first use, which the growth hire is measured on.

• Fewer demo bookings from the blog

  Keep the demo link and compare total pipeline, not bookings alone.

• Your websiteBlog: 102 posts and no newsletter signup; 7 of the 10 newest titles are guides to running prospecting and outbound from an AI assistant or coding agent, and older posts cover GTM engineering and "headless GTM". The co-founder's guide checked (byline July 19, 2026; page metadata Nov 7, 2026) ends with "Request access" and "Book a Demo" links to a 30-minute calendar booking and gives no MCP or skill install step.Source
• Your websiteMCP page: Fuse works with popular AI assistants, coding agents "and other MCP clients"; setup is one line to paste into the agent, "Install and set up https://docs.fuseai.com/skill.md", with "no API key to paste"; MCP calls draw on the same credits as the app and "each call needs a balance of at least 100 credits"; tools are labelled "read-only or destructive" so the agent can ask before acting; the demos are labelled "Illustrative demo with fictional companies and people"; there is no customer proof.Source
• HiringHead of Growth Marketing (open on Y Combinator's Work at a Startup board, checked November 9, 2026; San Francisco or remote US; $80K–$200K and 0.25%–2.00% equity): owns the inbound growth engine across partnerships, content and thought leadership, paid acquisition, SEO and programmatic content, events and community, product-led growth, lifecycle marketing and new growth experiments. "This is a hands-on role. You won't simply manage agencies, review dashboards, or create strategy decks." First-year goals: a predictable pipeline of qualified inbound opportunities, at least one major acquisition channel scaled, and high-impact distribution partnerships signed, with "clear reporting across awareness, acquisition, pipeline, conversion, and revenue".Source

Add a "What your credits buy" table under the plans showing contacts enriched, signals and research rows a month for each plan. Feed the same costs into the existing plan calculator, add a short pricing FAQ (rollover, top-ups, what happens at zero), and mirror the table in the help center and on the machine page.

Credit pricing is a common objection for metered sales tools. Concrete units let buyers pick a plan in seconds, and showing them that Free covers real work makes the plan more convincing.

[ ] Free: 24,000 a year = about 2,000 a month = 20 to 100 enriched contacts, or 20 signals
[ ] Solo ($60): 480K a year = 40,000 a month = 400 to 2,000 enriched contacts
[ ] Team ($200): 1.2M a year = 100,000 a month = 1,000 to 5,000 enriched contacts, shared by 5 seats
[ ] Same figures in the plan calculator, help center and machine page

• Buyers find enrichment expensive at the high end

  Show the range and explain why found contacts cost more than misses.

• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• Your docsAgent skill file: prospect search costs 1 credit per result; enrichment 20–100 credits per contact found; AI research columns 20 or 100 credits per row; signal agents 100 credits per result delivered; CRM push is free and goes to HubSpot, Salesforce, Zoho, Attio, Pipedrive and GoHighLevel; when the balance is too low the call returns "402 INSUFFICIENT_CREDITS" and bills nothing; campaigns stay drafts until the user approves sending.Source
• Your websiteHelp center: the index lists 10 topics, including "Getting Started" and "Billing, Plans & Credits", with no article titles and no credit costs; "Contact Support" opens a founder's 30-minute calendar booking.Source

Start with the five comparison pages that bring the most signups. Give each one a labelled two-column table, a price row, a source link and a "last checked" date for every claim about the rival, an honest "when to choose them" note, a switching guide (CSV or CRM import, first campaign in a day) and a customer quote once one exists (idea 9).

Comparison searches are high intent. Claims a buyer can check survive their scrutiny and the rival's sales team, and they are more likely to be cited by AI assistants.

• Rival claims go out of date

  Date every claim and recheck quarterly.

• Disputes from rivals

  Link to their own public pages for each statement.

• Your websiteComparison pages (21, each dated "Nov 7, 2026"): the Apollo page claims a "3.5 × larger contact database" (850M+ vs 240M+) and says "Apollo's LinkedIn sequence steps require manual completion"; on the Clay page the "at a glance" list has no labelled columns, so a reader cannot tell which value belongs to which product. Neither page has a price comparison, a customer quote, sources or a "last checked" note, and the stats block ("PIPELINE GENERATED", "REP HOURS SAVED") shows no readable figures to a text reader.Source
• Competitor researchRivals' own homepages: Apollo, "Grow revenue with one connected GTM system", 240M+ contacts, 600,000+ companies, its own MCP and free signup; Clay, "500K+ GTM teams", named customer results (Intercom +140% outbound pipeline, Rippling 2x demos from cold email) and an agent plugin to "Paste in your coding agent"; Unify, "Outbound agents for every rep", "$786,008,510 in pipeline powered" and named results (Spellbook $2.5M pipeline in 7 months).See the research Source
• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source

Keep one numbers sheet (contacts, companies, providers, signals, accuracy, customers) and update every page from it. Fix the title, align the meta description with the H1, and add a share image for the homepage, pricing, MCP and comparison pages.

Search and link previews are the first thing most visitors see. Numbers that disagree undercut a data company's main claim, accuracy.

• Pages drift again

  Link the sheet in the publishing checklist.

• #1 Fix the page Fuse writes for AI agents: it still quotes the old domain, 20+ providers and a 2,000-credit free plan

• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source
• Your websiteSignals page: "Identify up to 30% of anonymous website visitors"; funding, news, social activity, job changes and "20+ other buying signals" (the homepage and comparison pages say 50+); claims of "< 24 H" from signal to alert, "10 H+" saved per rep weekly and "3x" pipeline, with no customer or source named.Source
• Your websiteMachine version (/machine-version, reached from the homepage's Human/Machine toggle, published Nov 7, 2026): an llms.txt-style index for AI agents whose links all point to tryfuse.ai and app.tryfuse.ai (tryfuse.ai now 301-redirects to fuseai.com); it says "800M+ contacts across 20+ data providers", lists "Free ($0, 2,000 one-time credits)", Enterprise with "self-hosting" and "annual billing 25% off", and points to an llms-full.txt. fuseai.com/llms.txt and fuseai.com/llms-full.txt both return 404 (checked November 9, 2026).Source

Point each "Learn More" button at the matching product page (/prospect, /engage, /signals, /automations). Make "Start for free" the main button on the Founders and Marketing pages, with a ready-made starter workflow for that persona, and keep "Contact Sales" as the main button on the RevOps page for Team and Enterprise buyers.

Each button then matches what the visitor clicked for. Founders are the most self-serve buyers Fuse has, so a meeting is the wrong first step for them.

• Fewer sales meetings

  Keep Contact Sales visible and compare paid conversions.

• Your websitePersona pages: /founders ("Founder-led sales made easy") and /rev-ops ("Replace multiple tools with the ultimate all-in-one platform") lead with "Contact Sales", and their three "Learn More" buttons each open the same 30-minute calendar booking rather than a product page; no customer names appear.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• Your websiteSitemap (245 URLs, no last-modified dates, checked November 9, 2026): 102 blog posts, 43 data-provider pages and 7 provider categories, 50 help pages, 21 "<rival>-vs-fuse" comparison pages plus /alternatives, persona pages for founders, marketing, RevOps and sales, and /mcp, /labs and /machine-version. No case-study, customer, integrations, security/trust, careers or changelog pages.Source

Profile the slowest interactions (menu, Human/Machine toggle, tabs, calculator). Make the cookie notice lightweight, lazy-load demo media and animations below the fold, trim hidden sections that inflate the element count, and remove unused JavaScript. Target INP under 200 ms in the field.

A page that responds slowly to taps feels broken. Fixing INP also lifts the page out of Google's failing group.

• Site-builder limits

  Start with content and media changes that the builder allows.

• PageSpeedGoogle PageSpeed (homepage, November 9, 2026): mobile score 62 (desktop 91). Real visitors pass on loading (LCP 2.5 s) and layout shift (CLS 0.01) but fail Core Web Vitals on responsiveness (INP 345 ms at the 75th percentile). In the mobile lab test the largest element is the cookie notice text, LCP is 6.1 s, the page downloads 18.9 MB in 103 requests (mostly site-builder assets), with 4,084 page elements, 5.3 s of main-thread work and about 0.8 s of unused JavaScript.Source

Rewrite the post as a real list of 8-10 platforms scored on the same six categories, with a clear note that Fuse makes one of them, honest best-for notes for each rival, links to each comparison page and a date. Fuse can still come out on top for its own use case, signal-based multichannel outbound from one tool.

List posts win the "best X" searches, and AI assistants tend to quote balanced, disclosed comparisons over self-promotion. A fair list also earns links that a pitch does not.

• Sending readers to rivals

  Make Fuse's best-for case clear and link to the comparison pages.

• Your website"Best AI Sales Platforms for Outbound in 2026" (by a Fuse co-founder, byline Aug 30, 2026): evaluates six capability categories but profiles only Fuse and ranks no other product; the page title calls Fuse "#1 AI Platform for B2B Sales Teams"; its call to action asks readers to evaluate FuseAI "against these six categories".Source
• Your websiteComparison pages (21, each dated "Nov 7, 2026"): the Apollo page claims a "3.5 × larger contact database" (850M+ vs 240M+) and says "Apollo's LinkedIn sequence steps require manual completion"; on the Clay page the "at a glance" list has no labelled columns, so a reader cannot tell which value belongs to which product. Neither page has a price comparison, a customer quote, sources or a "last checked" note, and the stats block ("PIPELINE GENERATED", "REP HOURS SAVED") shows no readable figures to a text reader.Source

In six weeks, publish five short stories, one each for a founder, an SDR team, RevOps, marketing and an agent/MCP user, each built around one number (pipeline, reply rate, hours saved, tools replaced). Offer credits in return. Put logos on the homepage and pricing, quotes on the matching comparison and persona pages, and ask happy users for Product Hunt and review-site reviews.

For a young company, proof is the biggest gap between a visit and a signup. One named number does more than the 3x and 10-hours claims combined.

• Customers will not be named

  Use role and company size with the number, and logos only where allowed.

• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source
• Your websiteSignals page: "Identify up to 30% of anonymous website visitors"; funding, news, social activity, job changes and "20+ other buying signals" (the homepage and comparison pages say 50+); claims of "< 24 H" from signal to alert, "10 H+" saved per rep weekly and "3x" pipeline, with no customer or source named.Source
• Competitor researchRivals' own homepages: Apollo, "Grow revenue with one connected GTM system", 240M+ contacts, 600,000+ companies, its own MCP and free signup; Clay, "500K+ GTM teams", named customer results (Intercom +140% outbound pipeline, Rippling 2x demos from cold email) and an agent plugin to "Paste in your coding agent"; Unify, "Outbound agents for every rep", "$786,008,510 in pipeline powered" and named results (Spellbook $2.5M pipeline in 7 months).See the research Source
• Product HuntTwo launches: "The Agentic Sales Platform" (April 14, 2026, #5 of the day) and "Build a Custom GTM Stack. One SDK. One MCP." (October 6, 2026, #4 of the day); 1.2K followers; "No reviews yet". The listing says the average team uses 5-10 tools to run outbound.Source

Publish the benchmark method: sample size, roles and regions, how emails and phones were verified, and the date. Rerun it every quarter. Add a "Test it on your list" path: upload up to 100 contacts on the free plan and get Fuse's find rate and verified rate next to the buyer's current tool, as a shareable report.

A claim the buyer can rerun on their own list becomes a demo they run themselves. The method page also gives journalists, reviewers and AI assistants something solid to cite.

• The rerun scores below 95%

  Publish the real number; a documented lower figure is still more persuasive than an undocumented high one.

• Rivals dispute the method

  Publish the sample design and invite them to rerun it.

• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source
• Your websiteProviders directory: "40+ DATA PROVIDERS" (43 cards in 7 categories, from web search and company signals to contact enrichment and verification); each card has a name, one line and a category; "Provider access varies by plan"; most run on Fuse's own accounts and a few need the customer's own API key; no coverage, hit-rate or price data.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source

Add anonymised, aggregate stats to each provider page, refreshed monthly: email find rate, phone find rate and verification pass rate, by region and seniority, plus the plans it is available on and its typical credit cost through Fuse. Add a "how Fuse uses it in the waterfall" section and links to similar providers and comparison pages.

People searching a provider's name plus "coverage", "accuracy" or "alternative" are buyers. Unique data makes the pages worth ranking and citing, and shows the waterfall advantage instead of just claiming it.

• Provider contracts restrict publishing

  Publish only what terms allow and aggregate across categories where needed.

• Small samples mislead

  Show stats only above a minimum sample size.

• Your websiteProviders directory: "40+ DATA PROVIDERS" (43 cards in 7 categories, from web search and company signals to contact enrichment and verification); each card has a name, one line and a category; "Provider access varies by plan"; most run on Fuse's own accounts and a few need the customer's own API key; no coverage, hit-rate or price data.Source
• Your websiteSitemap (245 URLs, no last-modified dates, checked November 9, 2026): 102 blog posts, 43 data-provider pages and 7 provider categories, 50 help pages, 21 "<rival>-vs-fuse" comparison pages plus /alternatives, persona pages for founders, marketing, RevOps and sales, and /mcp, /labs and /machine-version. No case-study, customer, integrations, security/trust, careers or changelog pages.Source
• HiringHead of Growth Marketing (open on Y Combinator's Work at a Startup board, checked November 9, 2026; San Francisco or remote US; $80K–$200K and 0.25%–2.00% equity): owns the inbound growth engine across partnerships, content and thought leadership, paid acquisition, SEO and programmatic content, events and community, product-led growth, lifecycle marketing and new growth experiments. "This is a hands-on role. You won't simply manage agencies, review dashboards, or create strategy decks." First-year goals: a predictable pipeline of qualified inbound opportunities, at least one major acquisition channel scaled, and high-impact distribution partnerships signed, with "clear reporting across awareness, acquisition, pipeline, conversion, and revenue".Source
• Your docsAgent skill file: prospect search costs 1 credit per result; enrichment 20–100 credits per contact found; AI research columns 20 or 100 credits per row; signal agents 100 credits per result delivered; CRM push is free and goes to HubSpot, Salesforce, Zoho, Attio, Pipedrive and GoHighLevel; when the balance is too low the call returns "402 INSUFFICIENT_CREDITS" and bills nothing; campaigns stay drafts until the user approves sending.Source

Make the out-of-credits response helpful: a plain message the agent can show, with credits left, what the next plan includes and a one-click upgrade or top-up link. Warn at 20% remaining and send a lifecycle email that shows what the user's agent achieved with the credits (lists built, contacts enriched).

The moment a user runs out mid-task is when they value the product most. Putting the offer in the same place they are working removes the trip to a pricing page.

• Agents hide or reword the message

  Keep it short, factual and in a structured field, and back it up with email.

• #3 Show what a credit buys on the pricing page: 24,000 free credits is 240 to 1,200 enriched contacts

• Your docsAgent skill file: prospect search costs 1 credit per result; enrichment 20–100 credits per contact found; AI research columns 20 or 100 credits per row; signal agents 100 credits per result delivered; CRM push is free and goes to HubSpot, Salesforce, Zoho, Attio, Pipedrive and GoHighLevel; when the balance is too low the call returns "402 INSUFFICIENT_CREDITS" and bills nothing; campaigns stay drafts until the user approves sending.Source
• Your websiteMCP page: Fuse works with popular AI assistants, coding agents "and other MCP clients"; setup is one line to paste into the agent, "Install and set up https://docs.fuseai.com/skill.md", with "no API key to paste"; MCP calls draw on the same credits as the app and "each call needs a balance of at least 100 credits"; tools are labelled "read-only or destructive" so the agent can ask before acting; the demos are labelled "Illustrative demo with fictional companies and people"; there is no customer proof.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source

Build a listing kit (one-line install, five sales use cases, read-only and destructive tool labels, the free credits) and submit Fuse to the main MCP registries, skill and plugin marketplaces and assistant connector directories. Give each listing its own tracking so installs can be credited, and ask directory owners for featured slots in sales categories.

These directories are where agent users discover tools, and listing in them is low cost. Each listing is a backlink, and agents can find Fuse there too.

• Listings bring installs that never activate

  Pair with the activation path (idea 16).

• #1 Fix the page Fuse writes for AI agents: it still quotes the old domain, 20+ providers and a 2,000-credit free plan

• Your websiteMCP page: Fuse works with popular AI assistants, coding agents "and other MCP clients"; setup is one line to paste into the agent, "Install and set up https://docs.fuseai.com/skill.md", with "no API key to paste"; MCP calls draw on the same credits as the app and "each call needs a balance of at least 100 credits"; tools are labelled "read-only or destructive" so the agent can ask before acting; the demos are labelled "Illustrative demo with fictional companies and people"; there is no customer proof.Source
• Product HuntTwo launches: "The Agentic Sales Platform" (April 14, 2026, #5 of the day) and "Build a Custom GTM Stack. One SDK. One MCP." (October 6, 2026, #4 of the day); 1.2K followers; "No reviews yet". The listing says the average team uses 5-10 tools to run outbound.Source
• Competitor researchRivals' own homepages: Apollo, "Grow revenue with one connected GTM system", 240M+ contacts, 600,000+ companies, its own MCP and free signup; Clay, "500K+ GTM teams", named customer results (Intercom +140% outbound pipeline, Rippling 2x demos from cold email) and an agent plugin to "Paste in your coding agent"; Unify, "Outbound agents for every rep", "$786,008,510 in pipeline powered" and named results (Spellbook $2.5M pipeline in 7 months).See the research Source
• HiringHead of Growth Marketing (open on Y Combinator's Work at a Startup board, checked November 9, 2026; San Francisco or remote US; $80K–$200K and 0.25%–2.00% equity): owns the inbound growth engine across partnerships, content and thought leadership, paid acquisition, SEO and programmatic content, events and community, product-led growth, lifecycle marketing and new growth experiments. "This is a hands-on role. You won't simply manage agencies, review dashboards, or create strategy decks." First-year goals: a predictable pipeline of qualified inbound opportunities, at least one major acquisition channel scaled, and high-impact distribution partnerships signed, with "clear reporting across awareness, acquisition, pipeline, conversion, and revenue".Source

Publish one integration page per CRM (what syncs, setup in minutes, free push, a recipe), then apply to each CRM's app marketplace, starting with the one where most current customers already are. Co-market the listing with the CRM's partner team where possible.

Marketplace listings bring buyers who already have budget and a CRM, and integration pages catch searches the blog does not.

• Marketplace reviews take time

  Ship the integration pages first; they work without the listing.

• Your docsAgent skill file: prospect search costs 1 credit per result; enrichment 20–100 credits per contact found; AI research columns 20 or 100 credits per row; signal agents 100 credits per result delivered; CRM push is free and goes to HubSpot, Salesforce, Zoho, Attio, Pipedrive and GoHighLevel; when the balance is too low the call returns "402 INSUFFICIENT_CREDITS" and bills nothing; campaigns stay drafts until the user approves sending.Source
• Your websiteSitemap (245 URLs, no last-modified dates, checked November 9, 2026): 102 blog posts, 43 data-provider pages and 7 provider categories, 50 help pages, 21 "<rival>-vs-fuse" comparison pages plus /alternatives, persona pages for founders, marketing, RevOps and sales, and /mcp, /labs and /machine-version. No case-study, customer, integrations, security/trust, careers or changelog pages.Source
• HiringHead of Growth Marketing (open on Y Combinator's Work at a Startup board, checked November 9, 2026; San Francisco or remote US; $80K–$200K and 0.25%–2.00% equity): owns the inbound growth engine across partnerships, content and thought leadership, paid acquisition, SEO and programmatic content, events and community, product-led growth, lifecycle marketing and new growth experiments. "This is a hands-on role. You won't simply manage agencies, review dashboards, or create strategy decks." First-year goals: a predictable pipeline of qualified inbound opportunities, at least one major acquisition channel scaled, and high-impact distribution partnerships signed, with "clear reporting across awareness, acquisition, pipeline, conversion, and revenue".Source

Write the 15 articles that answer the most common support questions (credit costs, first campaign, connecting LinkedIn and email, CRM push, approving agent drafts) and list their titles on the help index. Add a support email or form with a stated response time, and keep calendar calls for Enterprise and onboarding.

Self-serve answers keep free and Solo users moving, free founder time, and can be read by the user's agent too.

• Articles go stale as the product ships fast

  Add help updates to the release checklist.

• #3 Show what a credit buys on the pricing page: 24,000 free credits is 240 to 1,200 enriched contacts

• Your websiteHelp center: the index lists 10 topics, including "Getting Started" and "Billing, Plans & Credits", with no article titles and no credit costs; "Contact Support" opens a founder's 30-minute calendar booking.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• HiringGrowth & Operations Summer Associate (San Francisco, $4.25K–$8.5K a month): launch growth experiments across digital channels, partnerships and community; help run onboarding flows, email campaigns and in-product experiments; build presence on social channels and online communities; explore new distribution channels, viral loops and product-led growth experiments.Source

Define activation as a first approved campaign or a first live signal agent within seven days. Build a guided start for each persona: a pre-filled search to preview, a ready-made signal agent, a starter knowledge hub and a draft sequence to approve. Back it with a short email series tied to the step each user reached.

Free users who reach a sent campaign see replies, and replies sell the paid plan. Measuring one activation event also tells the growth team which channel brings real users.

• Pushing sends before users are ready harms deliverability

  Keep sending behind approval and the sending policy.

• #6 Point the persona pages' "Learn More" buttons at the product, not a sales calendar

• HiringGrowth & Operations Summer Associate (San Francisco, $4.25K–$8.5K a month): launch growth experiments across digital channels, partnerships and community; help run onboarding flows, email campaigns and in-product experiments; build presence on social channels and online communities; explore new distribution channels, viral loops and product-led growth experiments.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• Your docsAgent skill file: prospect search costs 1 credit per result; enrichment 20–100 credits per contact found; AI research columns 20 or 100 credits per row; signal agents 100 credits per result delivered; CRM push is free and goes to HubSpot, Salesforce, Zoho, Attio, Pipedrive and GoHighLevel; when the balance is too low the call returns "402 INSUFFICIENT_CREDITS" and bills nothing; campaigns stay drafts until the user approves sending.Source

Give the referrer and the friend a block of bonus credits (for example 50,000 each) once the friend activates, not just signs up. Put a share link in the app, in the low-balance warning (idea 12) and in the lifecycle emails, and show a simple referral count in the workspace.

Credits cost Fuse far less than their face value, the reward is something users already want, and paying only on activation keeps fraud and empty signups down.

• Self-referral abuse

  Pay only on activation and block same-domain referrals.

• #16 Define activation for free users and build the path to the first approved campaign

• HiringGrowth & Operations Summer Associate (San Francisco, $4.25K–$8.5K a month): launch growth experiments across digital channels, partnerships and community; help run onboarding flows, email campaigns and in-product experiments; build presence on social channels and online communities; explore new distribution channels, viral loops and product-led growth experiments.Source
• Your websiteSitemap (245 URLs, no last-modified dates, checked November 9, 2026): 102 blog posts, 43 data-provider pages and 7 provider categories, 50 help pages, 21 "<rival>-vs-fuse" comparison pages plus /alternatives, persona pages for founders, marketing, RevOps and sales, and /mcp, /labs and /machine-version. No case-study, customer, integrations, security/trust, careers or changelog pages.Source
• Your websiteBlog: 102 posts and no newsletter signup; 7 of the 10 newest titles are guides to running prospecting and outbound from an AI assistant or coding agent, and older posts cover GTM engineering and "headless GTM". The co-founder's guide checked (byline July 19, 2026; page metadata Nov 7, 2026) ends with "Request access" and "Book a Demo" links to a 30-minute calendar booking and gives no MCP or skill install step.Source

Create /recipes on the main site: one page per workflow (for example "funding signal to three-step sequence" or "job change to LinkedIn and email"), each with the agent prompt, the credit cost, the expected output and a "Run in Fuse" button. Generate variants for each persona and signal type as programmatic pages.

A recipe page answers "how do I do X" searches with something the reader can run right away, and it gives agent users prompts that use Fuse by name.

• Thin programmatic pages

  Only publish variants with a distinct prompt and example output.

• Your docsDocs (docs.fuseai.com): Quickstart, Authentication, Guides (a list from search, custom columns, exporting, campaigns), Workflows ("end-to-end recipes" chaining search, enrichment, AI columns, signal agents and campaigns), an API reference with a playground, and an MCP server guide.Source
• Your websiteBlog: 102 posts and no newsletter signup; 7 of the 10 newest titles are guides to running prospecting and outbound from an AI assistant or coding agent, and older posts cover GTM engineering and "headless GTM". The co-founder's guide checked (byline July 19, 2026; page metadata Nov 7, 2026) ends with "Request access" and "Book a Demo" links to a 30-minute calendar booking and gives no MCP or skill install step.Source
• Your websiteSitemap (245 URLs, no last-modified dates, checked November 9, 2026): 102 blog posts, 43 data-provider pages and 7 provider categories, 50 help pages, 21 "<rival>-vs-fuse" comparison pages plus /alternatives, persona pages for founders, marketing, RevOps and sales, and /mcp, /labs and /machine-version. No case-study, customer, integrations, security/trust, careers or changelog pages.Source

Publish /security with the certifications and how to request reports, the agent guardrails (tool labels, approval before sending, per-user rate limits), the sending policy, where data comes from (the provider directory), a DPA and subprocessor list, and a contact for security questions.

Enterprise deals stall in security review, and agent-led selling raises the obvious question of what the agent can do on its own. One page answers both before the call.

• Overstating compliance

  State exactly which report type exists and its date.

• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source
• Your websiteSitemap (245 URLs, no last-modified dates, checked November 9, 2026): 102 blog posts, 43 data-provider pages and 7 provider categories, 50 help pages, 21 "<rival>-vs-fuse" comparison pages plus /alternatives, persona pages for founders, marketing, RevOps and sales, and /mcp, /labs and /machine-version. No case-study, customer, integrations, security/trust, careers or changelog pages.Source
• Your websiteMCP page: Fuse works with popular AI assistants, coding agents "and other MCP clients"; setup is one line to paste into the agent, "Install and set up https://docs.fuseai.com/skill.md", with "no API key to paste"; MCP calls draw on the same credits as the app and "each call needs a balance of at least 100 credits"; tools are labelled "read-only or destructive" so the agent can ask before acting; the demos are labelled "Illustrative demo with fictional companies and people"; there is no customer proof.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source

Offer a free two-week visitor report: add a snippet, get a daily list of the companies that visited, with their buying signals, and one click to build a list and draft a sequence. At the end, show what it would cost to keep it running on Solo or Team. Identify visitors at company level by default, with clear guidance on consent and privacy.

Buyers see their own pipeline in Fuse before paying, and the snippet keeps them coming back daily. It also makes Fuse's signals concrete in a way no demo can.

• Privacy rules vary by region

  Company-level only by default, legal review and clear consent guidance.

• Low-traffic sites see little

  Set expectations by traffic level before install.

• #16 Define activation for free users and build the path to the first approved campaign

• Your websiteSignals page: "Identify up to 30% of anonymous website visitors"; funding, news, social activity, job changes and "20+ other buying signals" (the homepage and comparison pages say 50+); claims of "< 24 H" from signal to alert, "10 H+" saved per rep weekly and "3x" pipeline, with no customer or source named.Source
• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• HiringHead of Growth Marketing (open on Y Combinator's Work at a Startup board, checked November 9, 2026; San Francisco or remote US; $80K–$200K and 0.25%–2.00% equity): owns the inbound growth engine across partnerships, content and thought leadership, paid acquisition, SEO and programmatic content, events and community, product-led growth, lifecycle marketing and new growth experiments. "This is a hands-on role. You won't simply manage agencies, review dashboards, or create strategy decks." First-year goals: a predictable pipeline of qualified inbound opportunities, at least one major acquisition channel scaled, and high-impact distribution partnerships signed, with "clear reporting across awareness, acquisition, pipeline, conversion, and revenue".Source

Test a usage-only plan for API and MCP users: no seats, monthly credit packs or pay-as-you-go top-ups, higher rate limits by tier, and app access as an add-on. Offer it to new signups who come from the MCP page, the docs and agent directories, and compare revenue per account with Solo and Team.

Builders and GTM engineers dislike paying for seats nobody logs into. A plan that matches how they use Fuse removes a buying objection and grows revenue with usage.

• Existing Team customers downgrade

  Offer the plan only to new agent and API signups at first.

• #3 Show what a credit buys on the pricing page: 24,000 free credits is 240 to 1,200 enriched contacts
• #12 Turn the out-of-credits error inside the agent into the upgrade moment

• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• Your websiteMCP page: Fuse works with popular AI assistants, coding agents "and other MCP clients"; setup is one line to paste into the agent, "Install and set up https://docs.fuseai.com/skill.md", with "no API key to paste"; MCP calls draw on the same credits as the app and "each call needs a balance of at least 100 credits"; tools are labelled "read-only or destructive" so the agent can ask before acting; the demos are labelled "Illustrative demo with fictional companies and people"; there is no customer proof.Source
• Product HuntTwo launches: "The Agentic Sales Platform" (April 14, 2026, #5 of the day) and "Build a Custom GTM Stack. One SDK. One MCP." (October 6, 2026, #4 of the day); 1.2K followers; "No reviews yet". The listing says the average team uses 5-10 tools to run outbound.Source

Launch a partner program: multi-client workspace management, a revenue share or credit bonus for each client workspace, a public partner directory, a short certification on Fuse workflows, and co-marketing (shared recipes, case studies). Start by inviting the agencies and freelancers who already run several Fuse workspaces.

Partners sell, set up and support clients Fuse could not reach with a team of 12, and they are often the people the client asks which tool to buy.

• Partners send low-quality outbound

  Require the sending policy and suspend partners who breach it.

• #17 A credits-for-credits referral loop for founders and GTM engineers

• Your websitePricing page: Free at $0 with "24,000/yr" credits and 1 seat; Solo at $60/mo with 480K credits/yr; Team at $200/mo with 1.2M credits/yr, 5 seats, unlimited AI research and buying signals, and parallel dialing; Enterprise at a custom price with "RBAC, SSO and audit logging", dedicated GTM engineer support and priority Slack support; an "Annual 20% Off" toggle; a plan calculator based on prospects, website visitors and team size. The page shows no credit cost per action, no self-hosting and no FAQ.Source
• Your websiteBlog: 102 posts and no newsletter signup; 7 of the 10 newest titles are guides to running prospecting and outbound from an AI assistant or coding agent, and older posts cover GTM engineering and "headless GTM". The co-founder's guide checked (byline July 19, 2026; page metadata Nov 7, 2026) ends with "Request access" and "Book a Demo" links to a 30-minute calendar booking and gives no MCP or skill install step.Source
• HiringHead of Growth Marketing (open on Y Combinator's Work at a Startup board, checked November 9, 2026; San Francisco or remote US; $80K–$200K and 0.25%–2.00% equity): owns the inbound growth engine across partnerships, content and thought leadership, paid acquisition, SEO and programmatic content, events and community, product-led growth, lifecycle marketing and new growth experiments. "This is a hands-on role. You won't simply manage agencies, review dashboards, or create strategy decks." First-year goals: a predictable pipeline of qualified inbound opportunities, at least one major acquisition channel scaled, and high-impact distribution partnerships signed, with "clear reporting across awareness, acquisition, pipeline, conversion, and revenue".Source

Replace the counters with real platform totals rendered as plain text and updated daily (or remove them). Then publish a monthly "Fuse on Fuse" post: the lists Fuse built, the signals that triggered outreach, reply and meeting rates, credits spent and what the team changed. Link it from the homepage and comparison pages.

Real numbers beat empty labels, and a company showing its own funnel in a market full of claims stands out. Each monthly post is also new, linkable content with real data.

• Early numbers look small

  Show rates and month-on-month change rather than only totals.

• #5 One set of numbers and a corrected page title across the site

• Your websiteComparison pages (21, each dated "Nov 7, 2026"): the Apollo page claims a "3.5 × larger contact database" (850M+ vs 240M+) and says "Apollo's LinkedIn sequence steps require manual completion"; on the Clay page the "at a glance" list has no labelled columns, so a reader cannot tell which value belongs to which product. Neither page has a price comparison, a customer quote, sources or a "last checked" note, and the stats block ("PIPELINE GENERATED", "REP HOURS SAVED") shows no readable figures to a text reader.Source
• Your websiteHomepage: H1 "Grow revenue with AI agents built for reps"; "850M+" contacts, "40M+" companies, "40+" data providers and "90%+ accuracy"; a self-reported accuracy chart (Fuse 95%, ZoomInfo 82%, RocketReach 78%, Apollo 74%) with no method shown; "Used by sales & marketing professionals from 1,000+ startups and enterprises worldwide" but no customer logos, names or testimonials; "BACKED BY Y COMBINATOR"; SOC 2 Type I, GDPR, CCPA and CASA Tier 2. The page title reads "Fuse | #1 AI Sales Platform Fast Growing Companies", the meta description calls Fuse "sales superintelligence for modern revenue teams", and there is no social share image.Source
• Your websiteCompany page: a manifesto ("Our philosophy is rooted in human augmentation, not human replacement"; "Our story is still being written") with no team, customers or careers link; footer "Made with ... in San Francisco".Source
• HiringEngineering: a Senior Software Engineer (San Francisco or remote in India, Brazil or Argentina; React, Node.js, MongoDB, AWS, microservices and Chrome extensions; "Experiment with AI agent architectures") and a Senior Frontend Engineer (remote; Next.js, React, TypeScript, SSR and caching, agent-driven interfaces). Fuse describes itself there as an agentic operating system for GTM teams; founded 2024, Y Combinator W25, team of 12.Source