The State of
GTM Connectors
How go-to-market teams and their agents are choosing data inside Claude.
The data decision moves to the connector.
What changed this quarter is not the data vendors; it is where their product gets used. A rep asks Claude to build a target list. The agent finds the accounts, enriches the contacts, drafts the outreach, and stages the sequence, and at no point does anyone open a vendor's interface. The agent gets its data by calling a connector.
That makes the connector the new point of choice. An agent cannot tell verified data from stale data; it acts on whatever the connector hands it. Which connector a team wires in has become the highest-leverage data decision in GTM, and, for the first time, a measurable one. This report benchmarks the 11 GTM data and enrichment connectors in the Claude directory across Q2.
What we found: ZoomInfo, Apollo.io, and Clay pulled ahead in the first weeks and never gave the lead back, separated from each other by about two points. Everyone else is fighting for fourth.
The leaders are outgrowing their lanes.
GTM connectors are not interchangeable. The category is splitting into intelligence, enrichment, signal, and execution layers, and an agent reaches all of them the same way: through a connector. The map arranges every first-party GTM data and enrichment connector in the directory by the job it does today.
The more important finding is that the lanes are already blurring at the top. Each premium-tier vendor ships a connector that under-represents its platform. ZoomInfo exposes five search-and-enrich tools while its platform sells intent data and engagement; Clay exposes find-and-enrich while its platform runs full enrichment workflows; Apollo.io has gone furthest, with a surface that already spans intelligence, enrichment, and execution. Further down the set, Lusha's surface already includes six signal tools. The pattern is consistent: vendors are extending their connectors toward their platforms' full footprint.
Expect the map to consolidate accordingly. Over the next two quarters, premium-tier connectors should converge from single-lane data access toward platform-shaped surfaces, and lane placement will become a trailing indicator of where each vendor already competes. The dashed lines on the map mark the expansions this report considers most likely.
A premium tier formed early and held.
ZoomInfo held #1 for all nine weeks; Apollo.io and Clay traded second beneath it, never more than a slot apart. Watched week by week, the top of the category looks less like a race than a settled hierarchy.
The real churn is in the middle band. Vibe Prospecting held #4 every single week of the quarter. G2 and Lusha climbed, then settled into #5 and #6. Sumble arrived in week seven and gained three places in two weeks. Rank is zero-sum: every climb here is someone else's slide.
How close the race is
Distance tells a different story than order. Two index points cover ZoomInfo, Apollo.io, and Clay; then comes a nine-point step down to Vibe Prospecting, and the field spreads through the sixties and eighties. This does not look like a one-company market. It looks like a three-way race ahead of a settling field.
What an agent can ask of each provider.
Tool count tells only part of the story. The tool names below show what an agent can actually ask each connector to do: search, enrich, create records, manage lists, read buyer intent, or move prospects into outbound workflows. This is the closest thing the directory has to a spec sheet, and it is worth reading before connecting anything.
A reading note: rank is a snapshot, popularity is a steadier signal, and trending is noisy. Tool count is useful, but it should not be read as quality by itself. Some connectors expose many small actions; others expose fewer, broader workflows.
The top cluster
The strongest current position in the set, separated from the field all quarter.
Claude can search contacts and companies, enrich a specific person or account, and look up GTM intelligence from a small set of broad tools.
search_contactsenrich_contactsearch_companiesenrich_companylookupClaude can search people, organizations, and existing contacts, enrich records, create or update contacts, and move prospects into or out of Apollo sequences.
apollo_enrich_organizationapollo_enrich_personapollo_bulk_enrich_peopleapollo_search_peopleapollo_search_organizationsapollo_search_contactsapollo_get_job_postingsapollo_create_contactsapollo_update_contactsapollo_search_sequencesapollo_add_contacts_to_sequenceapollo_remove_contacts_from_sequenceapollo_list_email_accountsClaude can find and enrich contacts or companies, add extra data points, check credits, and return to an existing Clay search.
find-and-enrich-contacts-at-companyfind-and-enrich-list-of-contactsfind-and-enrich-companyadd-contact-data-pointsadd-company-data-pointsget-credits-availableget-existing-searchThe specialist middle
Credible specialists whose strengths differ by use case.
Claude can match and enrich companies or prospects, fetch business and prospect events, inspect datasets, estimate cost, and export results.
autocompleteenrich-businessenrich-prospectsestimate-costexport-to-csvfetch-businesses-eventsfetch-entitiesfetch-entities-statisticsfetch-prospects-eventsget-datasetmatch-businessmatch-prospectsClaude can browse buyer intent, competitive intelligence, product and category data, and product reviews from G2.
browse_competitive_intelligencebrowse_product_buyer_intentlist_categorieslist_my_productslist_product_reviewslist_productsshow_categoryshow_productshow_product_reviewClaude can search and enrich contacts or companies, use prospecting filters, find lookalikes, inspect account usage, and pull signal data.
contacts_searchcompanies_searchprospecting_contact_searchprospecting_contact_enrichprospecting_company_searchprospecting_company_enrichprospecting_contact_filtersprospecting_company_filtersaccount_usagelookalike_companieslookalike_contactssignals_contacts_getsignals_contacts_searchsignals_contact_filterssignals_companies_getsignals_companies_searchsignals_company_filtersClaude can inspect available Common Room objects, list records, create or update objects, and send feedback back into the workflow.
commonroom_get_catalogcommonroom_list_objectscommonroom_create_objectcommonroom_update_objectcommonroom_submit_feedbackClaude can research accounts, search technologies, find and enrich organizations or people, manage account and contact lists, inspect jobs, and query saved tables. The largest tool surface in the set.
GetAccountInformationGetMyCompanyProfileSearchTechnologiesFindOrganizationsEnrichOrganizationMatchOrganizationsListOrganizationListsGetOrganizationListCreateOrganizationListAddOrganizationsToListListContactListsGetContactListCreateContactListAddContactsToListFindJobsGetJobDescriptionFindRelatedPeopleToJobFindPeopleEnrichPersonFindRelatedPeopleToPersonQueryListTablesLogOutThe emerging edge
Trailing on current rank, but each points at a GTM job that could matter more as agents become daily revenue tools.
Claude can search Sprouts data, refine queries, use saved templates, inspect search history, run ICP analysis, and move results into Sprouts.
sprouts_data_intelligencemove_to_sproutsquery_templatesquery_refinementsearch_historyicp_analysissprouts_user_preferencesClaude can pull technographics, firmographics, buyer intent, IT and cloud spend, contracts, install time series, and government opportunity data, plus search and enrich contacts.
company_technographiccompany_firmographiccompany_searchcompany_intentcompany_spendcompany_cloud_spendcompany_contractscompany_faicompany_operating_signalscompany_install_time_seriescompany_gov_opportunitiescompany_gov_relationshipscontact_searchcontact_enrichweb_searchClaude can find companies and subsidiaries and analyze workforce data: monthly headcount, employee growth, job functions, and seniority mix.
find_companiesget_company_subsidiariesget_monthly_headcountanalyze_employee_growthbreakdown_by_job_functionanalyze_seniority_levelsThe job decision is becoming a platform decision.
The practical takeaway is not "pick the leader." GTM connectors are splitting into layers, and the right one is a function of what your team's agents are being asked to do. Rank tells you where the market's defaults are forming; the job tells you where yours should form.
That guidance has a shelf life. If the premium tier converges on platform-shaped surfaces as this report expects, ZoomInfo, Apollo.io, and Clay will each cover several of these jobs from a single connector, and the specialist middle will have to differentiate on depth or be absorbed into a default. Choose for today's job, but weigh the trajectory: the connector a team standardizes on this quarter is likely to be the platform its agents are operating in by year end.
Defaults compound: the connector a team wires in today is the one its agents call next quarter, and the quarter after. In Q2, those defaults formed around a premium tier.
How this report is made.
Methodology
The set is the 11 first-party GTM data and enrichment providers in the Anthropic connector directory: ZoomInfo, Apollo.io, Clay, Vibe Prospecting, Lusha, G2, Common Room, Sumble, Sprouts Data Intelligence, Phoenix by HG Insights, and Aura. Community and unofficial listings are excluded.
Popularity is the directory's own popularity score, normalized within the provider set. Indexed values express each provider as a share of the set's top score. Weekly rank is mean popularity per ISO week, ranked within the set. Tool surfaces are snapshotted from each connector's directory listing at the end of the window.
The directory is polled every minute, and each connector's popularity and trending are snapshotted on change and at least every few hours: more than 100,000 snapshots across the window, 3,610 for these 11 providers alone. Window ≈ 25 Apr – 21 Jun 2026, nine weeks.
About Send
Send is where documents made with AI get shared. Turn a Claude conversation into a polished page, send it as a live link, and see who opened it, what they read, and when to follow up.
Send Research is our ongoing study of how work moves into agents, published quarterly. This report was built with the same data pipeline we run in production, and it is shared and tracked with Send.