GTM & RevOps
01Mostly Metrics
Anthropic's 85% Gross Margin Hides a 21-Point Gap
Reclassify channel and free-tier costs and Anthropic's 85% gross margin drops to roughly 64%.
What happened
- Anthropic reports an 85% gross margin with no COGS line broken out in its P&L.
- Cloud marketplace channel costs, about 50% of recent-quarter revenue routed through AWS, Google and Azure, sit in S&M, not COGS — roughly $2.7B in H1 on $16.1B gross revenue.
- Free-tier inference, arguably a cost of goods, also sits in S&M — about $600M in H1.
- Reclassifying both into COGS drops gross margin from about 85% to roughly 64%.
Why it matters
Reported gross margin for AI-native and channel-heavy companies can overstate unit economics by 20-plus points, so the number investors and board decks lean on may not reflect the real cost of delivering the product.
The JayOh take →
When building a client's RevOps reporting layer, push for a fully loaded margin view that captures channel and free-tier costs in COGS, not S&M, or leadership ends up managing to a number that won't hold up to a PE-style diligence pass.
02PPC Land
ChatGPT Work Use Falls as Writing Loses Top Spot
ChatGPT's share of work-related requests fell from 51.6% to 32.2% as decisions and advice overtook writing.
What happened
- OpenAI data shows work-related requests fell from 51.6% to 32.2% of consumer ChatGPT use.
- Writing dropped to 21.4% of requests and is no longer the top use case.
- Decisions and advice now lead US work-related use, about 1 in 5 requests.
- ChatGPT Ads adds a new monetization link inside the product.
Why it matters
As AI shifts from a writing tool to a decision engine, the content and campaigns B2B marketers optimize for need to answer 'what should I do' questions, not just rank for keywords, since buyer research increasingly happens inside the chat rather than on the page.
The JayOh take →
We're advising clients to audit their content for direct, decision-oriented answers like pricing comparisons, build-vs-buy frameworks and ROI math, since that's what wins inside AI answers now, not long-form SEO copy.
03GTM Strategist
One Person Runs a $2.4M AI Partnership Program
An AI-native startup's partnership OS lets one person manage 1,200 creators driving $2.4M in spend.
What happened
- Viktor, an AI employee that lives in Slack and Teams, raised a $75M Series A and passed a $50M run rate in under seven months.
- The company signed 70,000-plus teams and charges up to $60,000 a month per AI employee.
- One person runs its partnership program: 1,200 creators have driven $2.4M in spend.
- The company built an AI-native partnership OS to manage thousands of relationships without adding headcount.
Why it matters
A PLG, AI-native company proved a single-person partnership function can scale to thousands of relationships when recruiting, enablement, payouts and tracking are built as software instead of headcount.
The JayOh take →
For clients running partner or affiliate motions with a 1:1 ratio of managers to partners, this is the model to study: automate the operational layer first, then let one strategist own relationships at scale instead of hiring a team to do admin.
04Exponential View
Microsoft Offers Buyouts to 27% of Its Workforce
Microsoft offered voluntary exits to about 60,000 employees as AI capex climbs toward $170B.
What happened
- Microsoft offered a voluntary exit to around 60,000 employees across the US, UK, Canada and India.
- That's about 27% of its 223,000-person workforce; even VPs are eligible.
- Microsoft's capital spending is heading toward $170B, up from $23B five years ago.
- The move is not framed as layoffs, though it may precede a reduction round.
Why it matters
The scale of AI infrastructure spend is now reshaping headcount at the largest tech employers, not just entry-level roles, which signals how aggressively enterprise buyers may push AI-driven efficiency into their own GTM and ops functions.
The JayOh take →
Expect enterprise buyers who watch Microsoft's moves to accelerate their own AI-headcount tradeoffs, so RevOps teams should build the business case for AI-assisted roles now, before a budget mandate forces a rushed version.
Platform & Tool Releases
05HubSpot
Claude Now Runs Growth Workflows Inside HubSpot
HubSpot's Claude connector adds automated lead response, pipeline management and campaign tracking.
What happened
- HubSpot's Claude for Small Business integration adds new growth workflows.
- Covers lead response, pipeline management, CRM updates and campaign tracking.
- By default, Claude waits for user approval before actions go out.
- Announced alongside HubSpot's refreshed Starter plan for small businesses.
Why it matters
This pushes agentic AI further into day-to-day CRM execution for the SMB segment HubSpot's Starter tier targets, and the approval-gate design is the detail worth watching as larger accounts ask for the same workflows.
The JayOh take →
Pilot this inside a sandboxed HubSpot portal before letting it touch live pipeline; the approval gate is good, but any client using it needs a clear audit trail for what Claude touched and when.
06HubSpot
HubSpot Notetaker Adds Smart Deal Progression
HubSpot's September rollup ships Notetaker with Smart Deal Progression and new Help Desk QA scoring.
What happened
- Notetaker now pairs with Smart Deal Progression to suggest next steps from call notes.
- Customer Agent gets an auto-setup flow.
- Help Desk adds Customer Experience and Quality Assurance scores.
- Agent Builder is now available on the Starter tier, alongside dashboard caching improvements.
Why it matters
Smart Deal Progression turns passive call notes into an active pipeline signal, one more step toward HubSpot auto-populating the next best action instead of reps guessing at it.
The JayOh take →
Test this against a client's existing deal-stage criteria before trusting it; AI-suggested next steps need to be validated against the stage-exit rules already built, not treated as a replacement for them.
07Ruben ten Brinke
HubSpot Lets Admins Set a Ticket Reopen Window
HubSpot Help Desk now lets admins set a time window for whether a reply reopens a closed ticket or starts a new one.
What happened
- Admins can configure a time window for what happens when a customer replies to a closed ticket.
- A reply within the window reopens the original ticket.
- A reply after the window creates a new, linked ticket instead.
- Automated replies are logged without reopening or creating anything; email only for now, more channels coming.
Why it matters
This fixes a real reporting problem: old threads reopened months later were skewing reopen rates and SLA metrics for support teams, and the new window keeps that noise out of the data.
The JayOh take →
Set this window to match each client's actual SLA cycle rather than leaving it on a default, since it directly affects reopen-rate and first-response metrics in board-level support reporting.
Catch up on past briefings in JayOh University.
Turn these signals into your GTM roadmap.
We engineer the CRM, RevOps and AI systems behind B2B revenue teams.
Talk to JayOh