What is ServiceNow's right org structure in 2027 — verticals or horizontals?

Direct Answer

The right answer for ServiceNow in 2027 is the same answer it has today, just sharpened: a horizontal product org with a vertical sales overlay — and a new horizontal AI org sitting on top of both. Now Assist, AI Agent Studio, and the Workflow Data Fabric are platform-level capabilities; bending them into vertical P&Ls would shatter the velocity advantage McDermott has spent four years building.

But the buyers in Healthcare, FSI, Telco, and Public Sector pay 1.4–1.8x ACV uplift for vertical fit and named-industry references, which is too much margin to surrender to Salesforce Industries or Veeva-style pure-plays. The four design principles: (1) horizontal product, (2) vertical go-to-market, (3) AI as its own horizontal org reporting to the CPO, (4) a thin Industry Solutions packaging team that wraps horizontal modules into vertical SKUs.

The one organizational risk McDermott has to manage: matrix coordination tax — the same disease that slowed IBM, HPE, and pre-Nadella Microsoft. If the AE, the Vertical Specialist, the Industry Solutions PM, and the Now Assist PM all need to align before a deal closes, the AI velocity story dies in a Slack thread.

The Current Hybrid Structure

• Horizontal product orgs own the platform modules — ITSM, HRSD, CSM, IRM, App Engine, SecOps, and Now Assist — each with its own GM, PM org, and engineering line
• Vertical Industry Solutions teams (Healthcare, FSI, Telco, Manufacturing, Public Sector, Retail) own packaging, reference architectures, and named industry workflows on top of the horizontal modules
• Sales is a hybrid — territory + segment AEs (Enterprise, Commercial, SMB, Public Sector) with a vertical specialist overlay that parachutes into deals above $500K ACV
• Customer Success and Solution Consulting are organized horizontally by product, with industry SC pods inside Healthcare, FSI, and Public Sector
• The 2025 reorg consolidated Now Assist, Workflow Data Fabric, and AI Agent Studio under a single AI product GM reporting to CPO Pat Casey — a deliberate horizontal AI play

The Pure Horizontal Argument

• Now Assist crosses every module — IT, HR, customer service, security — so building it inside any single product org creates an instant tax on every other org that needs it
• Faster product velocity — horizontal orgs ship features once and propagate them across all SKUs, instead of N teams reimplementing the same capability for their vertical
• Named precedent: Microsoft Cloud + Copilot — Nadella collapsed Azure, M365, and Dynamics into a horizontal cloud + AI org and added 700 bps of operating margin while accelerating AI shipping cadence
• Less coordination cost — a horizontal org has one roadmap, one release train, one PM hierarchy; vertical orgs require N×N alignment meetings to agree on shared platform investments
• Talent gravity — top AI/ML and platform engineers want to work on horizontal capabilities used by 8,000 customers, not vertical features used by 400
• Pricing power on the platform layer — Now Assist Pro and AI Agent Studio are sold as platform add-ons across every module, which is only possible if AI is owned horizontally

The Pure Vertical Argument

• Regulated industries demand specialization — Healthcare needs HITRUST, HIPAA, FedRAMP High, EHR integrations; FSI needs SOC 2 Type II, FFIEC, model risk governance — generic horizontal teams ship these too slowly
• Named precedent: Salesforce Industries (Vlocity) and Veeva — both built $1B+ revenue lines on the thesis that vertical packaging beats horizontal customization, and Veeva is now a $30B market cap pure-play in life sciences
• Faster sales cycles — a vertical AE who has closed 40 hospital systems sells faster than a generalist with one healthcare deck, because the buyer hears their language back
• Better named-customer fit — boards buy the logo of a peer in their industry; "3 of the top 5 US health systems run on us" is a vertical narrative that horizontal orgs can't credibly assemble
• Vertical pricing power — Healthcare ACVs run 1.4–1.8x baseline because the buyer is comparing to Epic services, not to a horizontal SaaS list price
• Defensibility — vertical workflows (claims processing, telco order management, public sector case management) are stickier and harder to rip out than horizontal IT workflows

👉 Quick Call with Kory White, Fractional CRO · See Kory on LinkedIn · CRO Syndicate

The Hybrid Compromise (What ServiceNow Has Now)

• Product is horizontal, sales is vertical — the cleanest org-design pattern in enterprise software, used today by Microsoft, Adobe, and (partially) Oracle
• AE + Vertical Specialist overlay — the AE owns the relationship and the quota, the Vertical Specialist owns the industry credibility and pulls in Industry Solutions resources for the packaging story
• Product → vertical packaging cycle — horizontal product orgs ship the platform capability, then a thin Industry Solutions team wraps it into a vertical SKU within 60–90 days (the Now Assist for Telco template is the canonical example)
• Shared P&L on platform, vertical P&L on Industry SKUs — horizontal orgs are measured on platform ARR, vertical orgs on Industry SKU attach rate and pricing premium
• One Now Platform, many industry skins — the architectural commitment that the data model, workflow engine, and AI layer are never forked per industry, only configured

What Has To Change For 2027

• Now Assist + AI Agent Studio + Workflow Data Fabric become a single horizontal AI org with its own GM and a P&L that sits above the product modules — the same move OpenAI made by separating its model org from its applications org
• Vertical AI agents (Healthcare AI, FSI AI, Public Sector AI) get built as packaging on top of the horizontal AI org, never as forks of the AI stack — owned by Industry Solutions, not by the AI product team
• Named exec-sponsor pattern — every Industry Solutions GM gets a named C-suite sponsor (Casey for product, Wookey for sales, McDermott for the top 10 strategic accounts) to break matrix deadlocks within 48 hours
• The McDermott "one company" narrative — quarterly all-hands and earnings commentary explicitly reinforce that there is one platform, one AI strategy, and many industry GTM motions, to prevent vertical orgs from drifting into product-org behavior
• Compensation realignment — vertical specialists get paid on Industry SKU attach and ACV uplift, not on raw deal close, to stop them from competing with the core AE on incentives
• Kill the second-tier verticals — Retail and Manufacturing under-index on ServiceNow's strengths; consolidate them into a single "Commercial Industries" team and reinvest the headcount in Healthcare, FSI, and Public Sector

The 1 Organizational Risk

• Matrix coordination cost slows AI velocity — the same disease that ate IBM, HPE, pre-Nadella Microsoft, and Cisco in the 2010s; the symptom is a 6-week delay shipping a Now Assist feature because Healthcare, FSI, and Telco all want a different default prompt
• Named-pattern failure mode: dotted-line PMs — when Industry Solutions PMs have a dotted line into the AI product org, no one owns the roadmap and the AI features ship late and watered-down
• Vertical-specialist turnover — matrix structures burn out senior ICs because they have two managers, two priorities, and one quota; ServiceNow's vertical specialists turn over at ~22% annually vs. 14% for AEs, and the replacement ramp is 9 months
• Earnings-call risk — analysts will keep asking "is ServiceNow a horizontal platform or a vertical industries company?" — McDermott has to answer this in one sentence every quarter, or the multiple compresses
• Acquisition integration risk — every vertical-flavored M&A (Element AI for FSI, Lightstep for Telco, Hitch Works for HR) creates a new matrix node; without a clear horizontal-product vs. Vertical-GTM rule, every acquisition relitigates the org design

Org Design Tradeoffs Table

Org Decision Flow

Bottom Line

Keep the hybrid. Sharpen the horizontal AI org, kill the second-tier verticals, fix the comp conflict, and make McDermott's "one platform, many industries" sentence the single answer to every analyst question. The risk is not picking the wrong structure — it is letting the matrix tax the AI velocity that is the entire 2027 thesis.

(see also: q1632, q1638, q1639)

FAQ

Why not just fold Now Assist into each vertical P&L like Healthcare or FSI? Now Assist crosses every module — IT, HR, customer service, security — so embedding it in a single product org taxes every other org that needs it. Bending platform-level AI into vertical P&Ls would shatter the velocity advantage McDermott spent four years building.

The recommended structure keeps AI as its own horizontal org reporting to CPO Pat Casey.

How much pricing premium do vertical buyers actually pay? Buyers in Healthcare, FSI, Telco, and Public Sector pay a 1.4–1.8x ACV uplift for vertical fit and named-industry references. Healthcare ACVs run high specifically because the buyer is comparing to Epic services, not to a horizontal SaaS list price.

That margin is too large to surrender to Salesforce Industries or Veeva-style pure-plays.

What is the matrix coordination tax the article warns about? It is the risk that the AE, Vertical Specialist, Industry Solutions PM, and Now Assist PM all need to align before a deal can close. The same disease slowed IBM, HPE, and pre-Nadella Microsoft. If alignment dies in a Slack thread, the AI velocity story dies with it.

What did the 2025 reorg actually change? It consolidated Now Assist, Workflow Data Fabric, and AI Agent Studio under a single AI product GM reporting to CPO Pat Casey — a deliberate horizontal AI play. The 2027 target sharpens this into a single horizontal AI org with its own GM and a P&L sitting above the product modules, mirroring OpenAI separating its model org from its applications org.

How fast can a horizontal capability become a vertical SKU? A thin Industry Solutions team wraps a horizontal platform capability into a vertical SKU within 60–90 days, with the Now Assist for Telco template cited as the canonical example. The architectural commitment is that the data model, workflow engine, and AI layer are never forked per industry, only configured.

Sources

ServiceNow Q1 FY26 earnings transcript, ServiceNow Industry Solutions product page, Salesforce Industries page, Microsoft cloud + Copilot org analysis (Stratechery), Workday org structure analysis, Bessemer State of the Cloud 2026, McDermott interviews on org design (Fortune, CNBC).