Should ServiceNow pivot from platform-led to agent-led?

Direct Answer

No — ServiceNow should EVOLVE, not pivot. A full pivot from platform-led to agent-led would torch the $1M+ club moat that took 15 years to build. The right move: become the "Agent Platform of Record" — keep the Now Platform as the durable substrate, but lead with Now Assist + AI Agent Studio in every net-new sales conversation. Four reasons full pivot is a mistake: (1) it orphans the 8,100+ enterprise customers who bought the platform-of-record story, (2) it collides with Pro Plus pricing (per-conversation vs per-seat), (3) it forces a 22,000-person sales-team retraining mid-FY, and (4) named acquisitions that pivoted too fast (Pega RPA, Appian's agent push) lost retention before they gained net-new. The three evolution moves: net-new = agent-led demo, expansion = platform-led extension, renewal = AI Agent Studio attach. McDermott's 2025 reorg already signals this; the Q4 FY26 earnings call should formalize it as the "Agent Platform of Record" narrative.

The Old Platform-Led GTM

• Now Platform as primary sale — land with ITSM or HRSD, the platform license is the anchor contract
• Modules layered over 18-36 months — ITSM → ITOM → CSM → HRSD → SecOps → App Engine, the classic 8-deep attach
• 8-deep attach defines the $1M+ club — 2,109 customers >$1M ACV (Q1 FY26), 80%+ run 3+ workflows
• Named $1M+ club examples — Deloitte (12 workflows), Siemens (9), AstraZeneca (8), JPMorgan (10) — all platform-of-record stories, agents are additive not foundational
• IT-buyer trust is the moat — CIOs buy the platform because it's the system of action across IT, HR, customer, employee — that trust does NOT transfer cleanly to a "buy our agent" pitch

The Pure Agent-Led GTM Risk

• Orphans existing platform customers — if McDermott says "we're an agent company now," the CIO at Deloitte who just signed a 5-year platform renewal feels sold a bill of goods
• Pricing-model collision — Now Assist Pro Plus is per-seat ($35-50/user/month uplift); AI Agent Studio is moving toward per-conversation/per-action; the two models confuse procurement and stall deals in legal review
• Sales-team retraining cost — 22,000 quota carriers trained on platform discovery, workflow mapping, attach math; agent-led requires use-case-first discovery, ROI-per-agent math, and a different champion (often the COO not CIO)
• Pega's RPA pivot (2019-2022) — pivoted hard from BPM to RPA, lost BPM mindshare to Camunda + Appian, RPA never matched UiPath, stock down 60% peak-to-trough
• Appian's agent push (2024-2025) — leaned hard into "AI process automation," alienated low-code customers, Q4 2024 net new ARR missed by 18%
• Salesforce's Agentforce 1.0 stumble (Sep 2024-Mar 2025) — Benioff's "don't hire SDRs, deploy Agentforce" line spooked CRM buyers; 2.0 reframe (Dec 2024) walked it back to "agents on top of Customer 360"

The Evolved "Agent Platform of Record" GTM

• Lead with Now Assist + AI Agent Studio in net-new accounts — the demo opens with "watch this agent resolve a P2 incident in 90 seconds," platform discussion comes in slide 12 not slide 2
• Lead with platform in expansion deals — existing customers get the "extend your Now Platform with 3 new agents on the workflows you already own" pitch — agents become the unlock for cross-workflow expansion
• Named agent-pilot motion — 30-day pilot with a single named agent (Resolution Agent, Change Agent, Onboarding Agent), success metric = % auto-resolved, conversion to paid = 60%+ per ServiceNow internal benchmarks
• Named-account swat teams — 50-person "Agent SWAT" team for top 200 accounts, parachutes in for 90-day agent deployments, owns Now Assist attach number separately from platform AE quota
• Platform-of-record narrative stays for IT board pitches — CIOs still need the "one system of action" story for board presentations; agent narrative is for COO/CHRO/CRO buyers
• Pro Plus + Agent Studio bundled at renewal — every renewal conversation includes an AI Agent Studio attach motion, target: 40% Pro Plus attach by end of FY27

The 3 Evolution Moves

• Net-new sales motion = agent-led — Now Assist demo first, named-agent pilot in week 2, platform license signed in month 3 as the substrate (not the headline). Discovery question shifts from "what workflows do you run?" to "what work do you want a digital employee to do?"
• Expansion sales motion = platform-led — "You already own the Now Platform; here are 3 agents that unlock workflows you haven't touched yet (CSM, FSM, Legal Ops)." Agents become the wedge for cross-workflow attach, not a replacement for the platform pitch.
• Renewal motion = AI Agent Studio attach — every renewal conversation includes an Agent Studio sizing exercise (named agents per workflow, projected resolution rate, ROI). Renewal AE owns the Pro Plus + Agent Studio attach number; target: $250K incremental ACV per renewal in the $1M+ club.

Named Precedents

• Salesforce evolved CRM → Platform → Agentforce (1999-2024) — never pivoted, always layered: SFA → Force.com → AppExchange → Einstein → Agentforce. Each layer reinforced the prior. Agentforce 2.0 explicitly positions agents as "on top of Customer 360," not replacing it.
• Workday evolved HR → FSM → Extend (2005-2024) — kept HCM as the moat, layered Financials, then opened the platform via Extend. Never said "we're a finance company now" — said "the same trusted system, more workflows."
• Microsoft evolved Licensing → Subscription → Copilot (2010-2024) — Nadella never said "we're an AI company, forget Office." Said "Copilot is in every product you already own." M365 Copilot attach is the model ServiceNow should copy: $30/user uplift on existing seats, no SKU disruption.
• Adobe evolved Creative Suite → Creative Cloud → Firefly (2013-2024) — kept the Creative Cloud subscription as substrate, layered Firefly as additive — agents/AI as feature within the platform, not replacement for it.
• The pattern ServiceNow should borrow — "the same trusted platform, now with agents" — not "we're an agent company." McDermott has the storytelling chops to land this; the risk is over-rotating in earnings calls because Wall Street rewards AI narrative.

What McDermott Should Say On The Q4 FY26 Earnings Call

• Frame the narrative explicitly — "ServiceNow is the Agent Platform of Record. The Now Platform is the substrate. AI agents are the new unit of work that runs on it." Repeat verbatim 3+ times.
• Now Assist attach as the flagship metric — break out Now Assist Pro Plus attach % alongside the $1M+ club count; target: 35% attach in $1M+ club by end of FY26, 50% by end of FY27
• AI Agent Studio revenue line breakout — disclose Agent Studio ACV as a separate line item starting Q1 FY27 (~$400-600M run-rate projection); signals to Wall Street it's a real business, not a feature
• Named-customer reference patterns — 5-7 named customers with specific agent deployments + $ savings (e.g., "Deloitte deployed 12 agents, automated 1.2M tickets, saved $42M"); the "$1M+ club" frame extends to a "100M+ Agent Hours Automated" club
• Avoid the Benioff trap — do NOT say "don't hire IT staff, deploy agents." Do say "agents make your IT team 3x more productive." CIO buyers are the moat; don't insult them.

GTM Motion Comparison

GTM Evolution Flow

Bottom Line

EVOLVE, do not pivot. ServiceNow becomes the "Agent Platform of Record" — net-new conversations lead with Now Assist + AI Agent Studio, expansion conversations lead with the platform, renewals attach Pro Plus + Agent Studio. McDermott has the storytelling range to land this on the Q4 FY26 earnings call without spooking the $1M+ club. The Sandwich Stack frame from q1650 holds: outcomes on top, agent in the middle, platform as the toolset substrate — and ServiceNow owns more of that stack than Salesforce, Workday, or Microsoft do in their respective domains. (see also: q1613, q1649, q1650)