Lead generation ads, run from Claude and ChatGPT — with a lower cost per qualified lead
Opus Growth is a hosted MCP connector built for advertisers who are buying leads, not impressions. Paste one URL into Claude or ChatGPT and your assistant can work inside Google Ads (69 tools), Microsoft Advertising (30) and LinkedIn Ads (18), plus the measurement stack that decides whether those leads are counted correctly at all: Google Tag Manager (18), GA4 (10) and Search Console (6). Ask for cost per lead by campaign, strip out the search terms feeding junk form fills, repair a duplicated call conversion, launch a campaign targeted by job title. Every write action comes back as a dry-run preview and only executes when you approve it. 181 live tools, one connection, about 60 seconds to set up.
https://mcp.opus-growth.comWhat you can do
Move budget toward the leads that actually book
Pull cost per lead by campaign, ad group and keyword across Google Ads, Microsoft Advertising and LinkedIn, then act in the same conversation: raise daily budgets on the ad groups producing real enquiries, switch a campaign to Target CPA, cap the ones burning spend on form fills that never answer the phone. Budget, bidding-strategy and status tools write straight into the account once you approve the preview — no export, no second tab.
Fix the conversion tracking that inflates your lead count
Plenty of lead-gen accounts optimize on broken signals. List every conversion action, see which count as primary versus secondary, catch form and call conversions being double-counted, and correct counting settings. When the tag itself is the problem, 18 Google Tag Manager tools create the trigger, tag and variable and publish a new container version — with GA4 and Search Console data in the same chat to confirm the lead event actually fires.
Negative-keyword hygiene that protects lead quality
Junk leads usually start life as junk search terms: "jobs", "salary", "free", "do it yourself", a competitor's brand. Ask for search terms with clicks and zero conversions over the last 30 to 90 days, get them grouped by theme with a proposed negative list, and add it after you approve. Match types, keyword-level bids and pausing the losers all happen in the same conversation, so lead quality gets a weekly cleanup instead of a quarterly panic.
B2B targeting on LinkedIn without hand-copying URNs
Build Lead Generation campaign groups and Sponsored Content or text-ad campaigns targeting job title, job function, seniority, industry, skills, company size and geography — the connector resolves LinkedIn's targeting URNs from plain names like "Head of Demand Generation". Then pivot reporting by seniority, function, industry or company size and see precisely which segment produces your cheapest qualified lead, and which one only produces PDF downloads.
Form and call leads measured as leads, not clicks
In local lead gen the phone is the funnel. Manage call and sitelink extension assets, set ad schedules so ads run when someone can actually pick up, and adjust bids by device and location. Create or repair call and form conversion actions from the chat so phone enquiries are counted, then ask which hours and locations deliver answered calls rather than missed ones. The 7 Google Business Profile tools that keep your listings accurate are built but not live yet — they open when API access lands.
Microsoft's B2B intent plus a weekly report with no spreadsheet
Microsoft Advertising is live with 30 tools, and as the default on a lot of corporate desktops it often serves B2B queries at lower competition and cheaper clicks. Clone a winning Google search campaign — ad groups, keywords, RSAs — into Microsoft in one conversation. Then ask for spend, leads, cost per lead and week-over-week change by channel and get a single table across Google, Microsoft, LinkedIn and TikTok, plus GA4 and Search Console.
Set it up in 60 seconds
- 1
Add the connector
In Claude (claude.ai or Claude Desktop) or ChatGPT Developer mode, open Connectors and add https://mcp.opus-growth.com as a remote MCP server. Nothing to install, nothing to self-host, no config file to edit — it is a single URL.
- 2
Sign in with Google and pick your accounts
Hosted OAuth opens; sign in with Google and authorize the accounts you run leads for — Google Ads (including client accounts under an MCC), Microsoft Advertising, LinkedIn Ads and your measurement sources. No developer token, no API keys, no ETL pipeline.
- 3
Start asking in plain language
Type "show cost per lead by campaign for the last 30 days." Setup takes about 60 seconds end to end. Read questions answer instantly; anything that changes an account returns as a dry run you approve before it runs.
Ask in plain English
›Show cost per lead by campaign for the last 30 days across Google Ads, Microsoft Advertising and LinkedIn, and flag anything above $150.
›Pull search terms from the last 60 days with 3+ clicks and zero conversions, group them by theme, and draft a negative keyword list — preview it before adding anything.
›Audit my conversion actions: which count as primary, are form and call conversions double-counting, and is the lead event firing in Tag Manager and GA4?
›Create a LinkedIn Lead Generation campaign group and a Sponsored Content campaign targeting Heads of Operations at 201–1,000 employee manufacturers in Germany at $150/day — show the dry run first.
›Most of our phone leads close between 9am and 6pm — set the ad schedule to match and adjust the call extension hours.
›Weekly summary: spend, leads, CPL and week-over-week change by channel, then tell me the three ad groups to cut and where to move that budget.
Frequently asked questions
It works from whatever your ad platforms count as a conversion. If your qualified-lead stage is already imported back into Google Ads as an offline conversion action, Opus Growth can report on it and move bidding to optimize toward that instead of raw form fills. If it isn't, the fastest route to lead quality is cleaning the inputs — search terms, negatives, match types, targeting and tracking hygiene — which is exactly what the write tools do. Opus Growth works inside your ad and measurement platforms; it does not read your CRM.
It writes. For lead gen that means creating search, Performance Max and Demand Gen campaigns, adding keywords and negatives, switching to Target CPA, shifting daily budgets, editing responsive search ads, managing call and sitelink extension assets, setting ad schedules, creating and correcting conversion actions, and building LinkedIn campaign groups, campaigns and creatives. Reporting is included, but real write actions inside the account are the point — that is the difference between an ad connector and a data pipeline.
Every write tool runs as a dry run first. The assistant shows exactly what will change — account, campaign, field, old value, new value — and nothing reaches Google, Microsoft or LinkedIn until you confirm. Guardrails block runaway budget increases, every confirmed change is written to an audit log, and each customer's connection stays isolated. You can put the connector in a junior buyer's hands without worrying that one misread sentence multiplies a daily budget.
Most of it, it can fix. List conversion actions and see which are primary versus secondary, spot duplicate form and call conversions, and change counting settings. When the tag itself is wrong, 18 Google Tag Manager tools create triggers, tags and variables and publish a new container version, while GA4 and Search Console data help confirm the event fires. It is still worth submitting one test lead afterwards — then ask the assistant to pull the numbers and prove it landed.
Job title, job function, seniority, industry, skills, company size and geography, including location exclusions. You describe the audience in plain words — "Heads of Demand Generation at 201–1,000 employee SaaS companies in DACH" — and the connector resolves LinkedIn's targeting URNs so you never hand-copy an ID. Reporting pivots the same way, so you can compare cost per lead by seniority or job function and cut the segments that only generate content downloads.
Frequently, yes. Microsoft Advertising is live with 30 tools, and because it is the default search on a large share of corporate desktops, B2B queries there often face less competition and cheaper clicks. The low-risk test: clone a winning Google search campaign — ad groups, keywords, RSAs — into Microsoft in a single conversation, run it two weeks, then compare cost per lead side by side in the same chat. If it underperforms, you pause it the same way you launched it.
Start with a 7-day free trial: 50 actions, no card required. Pro is $49.99/month with 5,000 actions and covers every live platform. Agencies should take Agency at $99.99/month — unlimited actions plus multi-account and MCC-safe management, so one connection covers every client whose leads you are responsible for. Yearly plans are $499.99 and $999.99. Billing is in USD through Lemon Squeezy, and you can cancel anytime.
It runs in Claude and ChatGPT. Live platforms today: Google Ads (69 tools), Microsoft Advertising (30), TikTok Ads (25) and LinkedIn Ads (18), alongside Google Tag Manager (18), GA4 (10), Search Console (6) and YouTube (5) — 181 live tools out of a 221-tool catalogue. Two integrations are complete but not live yet: Meta Ads (27 tools) is in platform review, and Google Business Profile (7 tools) is waiting on API access. On the app side: Claude at claude.ai and Claude Desktop, ChatGPT via Developer mode connectors, and any MCP-compatible app such as Cursor. Sign-in is hosted OAuth — no developer token, no API keys, no self-hosting.
Explore more
Connect your lead-gen accounts to Claude or ChatGPT in about 60 seconds: paste https://mcp.opus-growth.com, sign in with Google, and ask for cost per qualified lead by campaign. 7-day free trial, 50 actions, no card, no API keys.
