Henry Schuck writes about go-to-market data and AI. Schuck argues that AI agents need connected, accurate company records—including account relationships and business history—and has announced ZoomInfo’s Agent Teams for go-to-market work.
Henry Schuck is most useful for explaining why AI sales workflows need reliable company records and context.
Best for
B2B SaaS brands helping sales and marketing teams prepare data for AI agents.
Not for
Consumer lifestyle brands seeking broad awareness or light entertainment endorsements.
Key takeaways
Duplicate Cisco records can make an AI agent confuse companies.
Missing subsidiary relationships can send a discount for non-customers to existing customers.
Agent Teams are presented as a way to bring AI workflows to a whole sales team.
Check first: Before contacting them, confirm sponsorship terms, audience breakdown and performance evidence.
Main ideas
AI go-to-market workflows break on the data underneath them; the fixes come in a set order · GTM context is the layer beneath AI, and owning it is the next vendor battleground · Agentic workflows only count once the whole team can run them, not just the engineer who built them
Henry Schuck writes about go-to-market data and AI.
Their central argument is that AI agents need reliable context: company records linked to the right subsidiaries, people, business history and activity.
They have also announced ZoomInfo’s Agent Teams, presenting them as a way to bring AI workflows to a whole sales team rather than leaving them with the person who built them.
Readers interested in making AI useful in sales and marketing will find practical examples of where poor data can lead it astray.
Schuck explains “GTM context” through concrete failure modes.
In one post, duplicate Cisco records illustrate why an agent may confuse companies; in another, missing subsidiary relationships lead a discount intended for non-customers to reach existing customers.
Their checklist for AI-ready go-to-market data: resolve company identities and hierarchies, connect first- and third-party data to the correct account, and make data definitions understandable to a model.
A separate post argues that even a strong account-planning workflow has limited reach if only its creator can use it.
Together, these ideas help readers assess both the data behind an AI workflow and whether a team can use it broadly.
Schuck’s posts describe ZoomInfo’s product response; ask for performance evidence before buying.
Across the three posts, Henry Schuck makes one argument in different registers: AI in go-to-market fails on the data layer, not the model.
Read more
Their priorities run in a consistent order — resolve entities and account hierarchies, stitch first- and third-party activity to each correctly resolved account, make the fields self-describing, and only then expect agents to produce anything usable. From there, they widen the claim: whoever owns that connected context graph holds the stronger position in the market, and single-engineer agent workflows only matter once an entire sales team can run them.
AI go-to-market workflows break on the data underneath them; the fixes come in a set order
Schuck argues that model capability is the wrong place to look when AI sales and marketing workflows disappoint.
Read more
The requirements they name are entity resolution, so the system knows which of several similarly named records is the real account; hierarchy resolution, so subsidiaries roll up under parents; first- and third-party data — calls, emails, product usage, opportunity history — stitched to the correctly resolved account; and metadata a model can read, described with Palantir's term "ontology."
The reasoning is that an agent only acts on records as they are stored, so an unresolved or drifting record produces a confident wrong answer at every step downstream.
They contrast this with what they call exhausting acronym churn: MCPs, CLIs, RAG, A2A.
Search "Cisco" in a scaled CRM and you get roughly 20 records — Cisco, Cisco Systems, Cisco WebEx, Cisco AppDynamics — with nothing to say which is real.
A team launched a 20% off promotion to "non-customers" and only after launch found it had discounted hundreds of existing customers, because subsidiaries such as Uber Freight had not been rolled up under Uber.
"Nvidia" transcribed as "In Video," or a press release about "Block, Inc" where the CRM says "Square," leaves the agent writing account plans about the wrong company.
On metadata, Schuck argues that go-to-market has its own ontology — accounts, leads, contacts, signals, champions, detractors, security review, procurement — and a model that does not know what a column means will answer anyway.
This post lays out the diagnosis. In the later context-graph post, Schuck keeps the same requirement list but reframes it as a competitive asset rather than a data-cleaning chore.
GTM context is the layer beneath AI, and owning it is the next vendor battleground
Schuck picks up the phrase "GTM context" from analyst Adam Schoenfeld, whose argument Schuck credits and quotes (the phrase is Schoenfeld's; the extensions below are Schuck's).
Read more
Schuck's claim is that the next era is won not by whoever builds the most sophisticated campaign or demand workbook but by whoever builds and owns the connected context that people and agents then run motions on top of.
Companies assume that picture already lives in the CRM, and in Schuck's account the CRM is a mess: hierarchies, subsidiaries, employment histories, and rebrands change faster than stored records, and no amount of inference invents those links without a reference dataset.
The proposed layer is a "context graph" connecting entities, history, and activity.
Schuck then draws an analogy — every function running on AI needs such a layer — and names each function's incumbent before naming go-to-market's.
Google and Waymo sit as separate accounts with no link to Alphabet Inc.
SpaceX acquires Cursor but the account hierarchy never updates.
Travis Kalanick runs CloudKitchens, which is part of Atoms, while the marketing system still lists him at Uber.
Other functions already have their layer: legal has Harvey and Legora, engineering has Atlassian, enterprise search has Glean and Dust.
The post ends by naming ZoomInfo as go-to-market's equivalent, making the strategic framing vendor positioning for the company Schuck speaks for.
Read alongside the four-requirements post, this shifts from "here is what is broken" to "here is the layer that fixes it, and here is who owns it."
Agentic workflows only count once the whole team can run them, not just the engineer who built them
Schuck starts from an observation about the impressive Claude-built go-to-market workflows of the past year: dump in CRM notes, call transcripts, emails, and account data, and out comes a near-perfect account plan with who matters, what changed, what to do next, and what to say.
Read more
The problem, in their words, is that such a workflow has "a user base of one": usually the go-to-market engineer who built it, while everyone else keeps digging through Salesforce and starting from scratch. Their priority is the reliability of a workflow combined with the judgment of a team, delivered to every rep rather than to its author.
Schuck announced ZoomInfo's Agent Teams the day after ZoomInfo announced its acquisition of DoubleO AI, and describes Agent Teams as built on DoubleO's agentic technology.
Four motions are named: an inbound lead gets researched, qualified, routed, and sequenced; a just-closed deal triggers a search for lookalikes, prioritization, and drafted outreach; a champion changing jobs triggers a rebuilt relationship and drafted outreach; a shaky renewal triggers history pulled into a save plan and an account-owner briefing.
Ask for independent evidence that the product delivers these motions in practice.
Start here for the core thesis: why Schuck believes AI in sales and marketing keeps stalling at the data layer, and what they propose instead of another campaign tool.
Read more
Schuck picks up analyst Adam Schoenfeld’s phrase, “GTM context,” and argues that the next era of go-to-market competition will be won by whoever builds a connected picture of customers, not the most elaborate campaign.
Companies assume that picture already lives in the CRM, but the CRM is a mess: Google and Waymo sit as separate accounts with no link to Alphabet Inc.; SpaceX acquires Cursor, but the account hierarchy never changes; Travis Kalanick now runs CloudKitchens, itself part of Atoms, while the marketing system still lists them at Uber.
Schuck’s point is that no amount of inference can create those links: hierarchies, subsidiaries, employment histories, and rebrands are constantly in motion and need a reference dataset to anchor them.
The proposed fix is a context graph, a layer connecting entities, history, and activity so people and agents work from the full picture instead of isolated records.
To make the case concrete, they note that other functions already have this layer — legal has Harvey and Legora, engineering has Atlassian, enterprise search has Glean and Dust — and position ZoomInfo as the equivalent for go-to-market teams.
For brands planning a campaign
A vendor or agency pitching AI go-to-market tooling can borrow the infrastructure framing rather than the automation framing: lead with data readiness, not campaign speed.
The practical companion to the thesis: four checkable questions for a data team, with real failure modes to test against its own systems.
Read more
This is Schuck's most concrete piece: four requirements that have to be met before an AI agent can work on sales and marketing data, each illustrated with a specific failure.
Entity resolution: search "Cisco" in a scaled CRM and you get around 20 records — Cisco, Cisco Systems, Cisco WebEx, Cisco AppDynamics — and no way to tell which is real, so the agent picks one and is likely wrong.
Hierarchy resolution: one team launched a 20% off promotion to "non-customers" and only after launch discovered it had offered the discount to hundreds of existing customers, because subsidiaries such as Uber Freight had not been rolled up under Uber; account managers were furious and revenue was lost.
First- and third-party stitching: calls, emails, product usage, and opportunity history all have to attach to the correctly resolved account, or the agent writes account plans about the wrong company — Schuck's examples are "Nvidia" transcribed as "In Video" and a press release about "Block, Inc" when the CRM says "Square."
Readable metadata: borrowing Palantir's term "ontology," Schuck argues that if a model does not know what a column means it will give a confident, wrong answer, and that go-to-market has its own ontology — accounts, leads, contacts, signals, champions, detractors, security review, procurement.
For brands planning a campaign
Usable as an audit checklist for a data-readiness engagement: entity resolution, hierarchy roll-up, per-account stitching, and readable metadata map cleanly onto assessment phases.
Shows where the argument becomes a product: the clearest statement of how Schuck thinks agents should fit into a sales team’s day-to-day motions.
Read more
Schuck announces ZoomInfo’s Agent Teams for go-to-market work, the day after ZoomInfo announced its acquisition of DoubleO AI, and frames the launch as building on DoubleO’s agentic technology.
They open with an observation rather than a product pitch: the impressive GTM workflows people built in Claude this year — dumping in CRM notes, call transcripts, emails, and account data to generate a near-perfect account plan with who matters, what changed, what to do next, and what to say — have “a user base of one.”
That one user is usually the GTM engineer who built the workflow, while everyone else still digs through Salesforce and starts from scratch.
Agent Teams is presented as widening access to every rep, with four motions named: an inbound lead gets researched, qualified, routed, and sequenced; a just-closed deal triggers a search for lookalikes, prioritization, and drafted outreach; a champion changing jobs triggers a rebuild of the relationship and outreach; and a shaky renewal triggers a review of history and a save plan and account-owner briefing.
The claim is reliability of a workflow with the judgment of a team.
For brands planning a campaign
Relevant to anyone positioning against or alongside ZoomInfo: the differentiator claimed here is distribution and context, not model quality. Schuck presents this as a vendor announcement, not independently verified product performance; ask for performance data.
Henry Schuck argues that most AI go-to-market workflows fail because of the data beneath them, not the models.
Read more
The post names four requirements: entity resolution—the CRM that returns Cisco, Cisco Systems, Cisco WebEx and Cisco AppDynamics as separate records; hierarchy resolution—a team offered a 20% discount to "non-customers" and hit hundreds of existing customers because subsidiaries such as Uber Freight had not been rolled up under Uber; first- and third-party data stitched to each correctly resolved account, including calls, emails, product usage and opportunity history; and metadata a model can read, described as an ontology and credited to Palantir. They contrast this with the acronym churn they find exhausting: MCPs, CLIs, RAG, A2A.
Schuck uses named failure stories rather than benchmarks, with duplicate Cisco records and a mis-targeted discount illustrating what goes wrong when records remain unresolved.
For brands planning a campaign
A brand building a data readiness plan for AI agents can use Schuck's four-part sequence as an audit order: resolve entities and hierarchies first, connect usage and call data next, then make fields self-describing before expecting agent output to hold up.
GTM context: the layer beneath AI agents and a market position
Responding to Adam Schoenfeld's phrase "GTM context," Schuck argues that the next era will be won not by whoever builds the most sophisticated precision demand workbook, but by whoever builds and owns the context agents and people use to run motions.
Read more
In their account, CRM is a mess: Google and Waymo sit as separate accounts with no link to Alphabet, acquisitions never update account hierarchies, and people's current employers no longer match their records, as illustrated by Travis Kalanick at CloudKitchens and Atoms rather than Uber.
Schuck calls a context graph the missing layer beneath AI, connecting entities, history and activity.
They argue every function running on AI needs one, pointing to legal (Harvey, Legora), engineering (Atlassian) and enterprise search (Glean, Dust) before naming ZoomInfo for GTM.
Schuck argues by analogy: if other functions have a context layer, GTM needs one too. Drift examples show how hierarchies and employment histories can change faster than stored records.
The post closes with a link to ZoomInfo, positioning GTM context as a contested battleground for the company the post speaks for.
For brands planning a campaign
For a brand planning an AI data foundation, Schuck's strategic argument is that ownership of context can be a competitive question. Their Alphabet and Kalanick examples offer simple tests for whether a brand's account records have drifted out of date.
Announcing ZoomInfo's acquisition of DoubleO AI and the launch of Agent Teams, Schuck describes a familiar pattern: a GTM engineer builds an impressive workflow in Claude that turns CRM notes, call transcripts, emails and account data into a near-perfect account plan, but only one person uses it.
Read more
Agent Teams is presented as the fix, putting agents into every rep's day for named jobs: researching, qualifying, routing and sequencing an inbound lead; finding lookalikes and prioritising those worth pursuing after a closed-won deal; rebuilding a relationship and drafting outreach when a champion changes jobs; and pulling the history and building a save plan when a renewal looks bleak. The pitch combines the reliability of a workflow with the judgment of a team, available to everyone rather than just its builder.
For brands planning a campaign
For a brand weighing whether to build or buy internal GTM agents, this list defines the rep-facing tasks vendors are targeting and the gap between a working prototype and a tool an entire team can rely on.
The posts connect Schuck most clearly to ZoomInfo, where they announce product and acquisition news as “we.”
Read more
They also mention DoubleO AI as the acquired technology behind ZoomInfo’s Agent Teams. Other company names in the posts are examples of GTM data problems, not documented brand relationships.
ZoomInfo
Disclosed partnership
Schuck posts as “we” when announcing ZoomInfo’s acquisition of DoubleO AI and the launch of Agent Teams for GTM, and presents ZoomInfo as the GTM context layer beneath AI.
Read more
For a paid partnership, ask ZoomInfo to confirm their formal role and ownership.
Henry Schuck's posts are short, argument-first pieces about the data layer beneath AI in go-to-market work.
Read more
Start with the four-requirements post on entity resolution, hierarchy resolution, account-level stitching of first- and third-party data, and readable metadata.
It shows how they reason: through named failures and specific examples, not benchmarks.
Read the GTM context post next for the strategic version of the same case: hierarchies, acquisitions and job changes move faster than stored records, and owning the connected picture is the market position worth holding.
The Agent Teams post is the product-facing counterpart.
Schuck writes from a software company's perspective, so weigh that vendor interest when reading the post, which names the equivalent data layer in legal, engineering and enterprise search before naming go-to-market's.
For a practical sequence, use the four requirements as an audit order; for the argument about where the industry is heading, start with the context post.
Go-to-market dataAI agents in go-to-marketEntity resolutionAccount hierarchy and subsidiariesContext graphsGo-to-market technologyLinkedIn postsRevOps and GTM engineers
Tatulea makes the case through the same examples: Google and Waymo sitting apart from Alphabet, an acquisition that never updates the hierarchy, and Travis Kalanick recorded at Uber while running CloudKitchens.
Read more
The conclusion is the same: whoever gives an agent the best understanding of the account wins, not whoever gives it the most tools. Tatulea also quotes Schuck directly, calling their line about reasoning without verified data “fluent guesswork.”
Tatulea writes from the operator's side, laying out ABM as four layers — a named account list, named buyers, proven conversion signals and account context — and explaining how teams stitch them together and how they go stale. Schuck writes from the software company's side, and their posts carry the vendor positioning that comes with it.
Robbins uses the same format: first-person leadership posts that argue about enterprise AI adoption through accounts of specific customer, partner and government meetings.
Read more
Both write as executives explaining where AI progress depends on something beyond the model. For Robbins, it’s partnerships, infrastructure and skills; for Schuck, resolved and connected account data.
Robbins focuses on connectivity, cybersecurity and country-level AI programmes, illustrated by visits, launches and training commitments rather than a data-model argument. Schuck stays inside go-to-market systems, showing what breaks in a CRM record; Robbins rarely gets that technical and never pitches a schema or sequence of fixes.
Champoux describes building the same kind of layer Schuck argues for: a system that connects the tools a business already uses, reconstructs the full context in a structured graph, then runs agents on top to capture, enrich, reconcile and act — including scoring leads and routing them into the CRM automatically.
Read more
Both write build-and-launch posts with concrete operational numbers.
Champoux focuses on startup finance — bookkeeping, invoices, receipts and receivables — and writes mostly about company progress from an accelerator cohort, rather than where go-to-market data breaks. Schuck's posts analyze a market they sell into; Champoux's record a product taking shape.
Henry Schuck reaches a substantial, self-selected LinkedIn following, but the posts reveal more about its interests than its composition.
Read more
The only published platform figure is roughly 103,000 LinkedIn followers on 8 October 2026.
The three dated posts argue that AI agents fail when account records are unresolved, hierarchies are stale and metadata is unreadable; one announces ZoomInfo's Agent Teams for GTM work.
This is a narrow, operationally literate register for people who own CRM hygiene, GTM data pipelines or go-to-market engineering decisions.
It aligns with a buyer evaluating data readiness, though that is an inference from the posts, not a measured read of the follower base.
For a B2B SaaS company building a go-to-market data foundation for AI agents, ask for a documented audience breakdown—follower demographics from LinkedIn, if they can provide them, or an audience survey reporting job function, seniority and geography—before committing budget.
Henry Schuck's three posts in scope cover product and company announcements: the launch of ZoomInfo Agent Teams for go-to-market work, GTM data context and entity resolution, and account hierarchy and CRM data quality.
Read more
They do not describe sponsored placements or disclosed brand collaborations.
No official sponsorship policy, rate structure, audience specification, or partnership inquiry route appears in the posts.
Buyers should request first-party confirmation of partnership terms before proceeding.
Good fit: A B2B SaaS company helping sales and marketing teams prepare their data for AI agents.
Read more
Schuck’s posts address this problem directly, laying out four practical checks: resolve duplicate entities, roll subsidiaries into the right account, connect first- and third-party activity to that account, and make metadata understandable to agents.
The Cisco duplicates they cite and the discount mistakenly sent to existing customers make the risks concrete.
A campaign could help buyers assess their data foundation, provided the company is comfortable with a point of view grounded in Schuck’s public ZoomInfo product positioning.
The posts support this subject fit.
Confirm sponsorship format, audience makeup, and partnership availability before planning a paid placement.
When to choose a specialist
Poor fit: A consumer lifestyle brand seeking broad awareness or a light entertainment endorsement.
Read more
The posts focus narrowly on go-to-market data, CRM records, and AI workflows; they show no lifestyle content or consumer-focused format.
The available audience information is a LinkedIn follower count, which does not establish consumer demographics or buying interests.
This category calls for a different creator with demonstrated relevant work and audience information.
Confirm Schuck’s sponsorship terms and whether they accept this kind of work before considering a partnership.
A worked campaign example
Campaign need: A B2B SaaS company wants prospects to assess whether their account data is ready for AI-assisted sales workflows.
Recommendation: Consider Schuck for a focused educational campaign, subject to confirming partnership availability, audience fit, and commercial terms.
Campaign angle: “Before you add an AI agent, check whether it knows which account it is working on.”
Build a short diagnostic around the four checks in their post: (1) search for duplicate or conflicting company records, using the Cisco example as a prompt; (2) test whether subsidiaries roll up to the correct parent, using the Uber Freight discount mistake as the caution; (3) check whether calls, emails, product usage, and opportunity history attach to the resolved account; and (4) ask whether field definitions are clear enough for an agent to interpret.
The company could offer a worksheet asking readers to record one failure or gap for each check, then direct them to a data-readiness consultation.
Keep the piece educational rather than presenting Schuck as an independent product reviewer: their post on GTM context ends by positioning ZoomInfo, and their Agent Teams post is a ZoomInfo launch announcement.
Validate the worksheet, consultation, and placement with the brand; these are proposed campaign elements, not confirmed offerings.
Pick Tatulea when the campaign wants the data-readiness argument from someone running ABM programs, not an executive at a data vendor.
Read more
The posts describe ABM plays scaled past $1.2bn and set out four layers a team needs in one place: a named account list, the named buyers at each account, signals that predict conversion, and the existing context on each account.
They also warn that an agent chains steps, so a wrong field at step one becomes the assumption every later step builds on.
That practitioner framing suits a hands-on workshop, a diagnostic session, or a peer case discussion for sales and RevOps teams more than a product-led message, and fits the same GTM data-foundation brief.
Choosing Tatulea may reinforce rather than diversify the argument.
The examples overlap almost exactly with Schuck's: separate Google and Waymo records with no link to Alphabet, SpaceX acquiring Cursor without the hierarchy changing, and Travis Kalanick still listed at Uber while running CloudKitchens.
One post credits them directly with the line that reasoning without verified data is fluent guesswork.
A buyer choosing between the two gets a similar case, not an independent second view.
Only two posts were reviewed; ask for audience makeup and available partnership formats before deciding on commercial fit.
Consider Robbins when the need shifts from data hygiene to enterprise AI infrastructure: national-scale AI programs, sovereign security operations, workforce training commitments, and audiences in senior technology, government, and service-provider roles.
Read more
The posts cover AI infrastructure and partnership work in India and Indonesia, including an AI Center of Excellence with IOH, NVIDIA and the communications ministry, a sovereign security operations platform using Splunk and managed services, a research partnership with the national innovation agency, and a training goal of 500,000 more students by 2030.
They also cover meetings with connectivity providers at MWC Barcelona and government discussions in Switzerland and Italy.
That suits an awareness or thought-leadership campaign about deploying AI infrastructure at scale with public and institutional partners.
Robbins's posts do not address campaigns aimed at buyers who need CRM entity resolution, subsidiary rollups, account hierarchies, or sales data prepared for agents.
A brand that needs buyers to act on data readiness gets no usable material.
The posts offer narrative accounts of country visits and announcements, not method, diagnosis, or a repeatable process.
Robbins is also a much larger institutional subject, and the content is not tailored to a B2B SaaS data-readiness campaign.
Choose Champoux when the campaign should show an operator building a GTM data engine from scratch rather than an executive explaining why the problem matters.
Read more
The posts document the build in specifics: 46 features a month against 7 at the start, a go-to-market engine processing 3.4 million signals and scoring more than 1,600 leads across three tiers, and a context graph that connects the tools a business already uses and deploys agents to capture, enrich, reconcile and act on the data. That fits a founder-to-founder or build-in-public angle about assembling GTM data infrastructure, or a startup finance and operations context around automating invoices, receipts, and financial questions.
The setting is different.
Champoux's material focuses on early-stage startup finance and the company's own product build, not an established sales organisation fixing its account records.
It does not address entity resolution, subsidiary rollups, or CRM hygiene, so it would not answer the original data-foundation brief.
Some content is launch promotion, including a new website, logo, and vision statement, which limits its value as teaching material.
Ask directly about available sponsorship formats, audience makeup, and partnership availability.
This profile draws on the complete text of three public LinkedIn posts from Henry Schuck's account, covering four requirements for AI-ready go-to-market data, the case for a connected "GTM context" layer, and the launch of ZoomInfo's Agent Teams.
The posts arrived without publication dates, so no timeline is claimed for how the arguments developed, and nothing here is drawn from outside these three posts.
Comparisons with other creators elsewhere on the page rest on public posts from their accounts, also read in full.
Statements about ZoomInfo products, acquisitions and data problems are reported as Schuck's own posts describe them; they were not independently verified, and the posts do not establish a formal role, sponsorship formats, audience composition or partnership terms.
