Enterprise Software: Partner checks point to resilient core CRM demand but an uneven, early path to enterprise AI monetization

Overview

This enterprise-software field-check report combines Salesforce Dreamforce observations, HubSpot partner conversations and a pre-quiet-period roundup. Its central message is that core application demand appears more durable than disruption narratives suggest, but AI adoption and monetization remain gradual, data-dependent and uneven across customers.

Core views

Dreamforce partner checks indicate that core CRM spending has moved off the sidelines after customers spent roughly last year through May or June in a status-quo mode focused on maintenance. Contacts described a material pickup in Salesforce activity over the most recent three to four months as customers concluded that AI-assisted or vibe-coded replacements were not sufficiently robust for complex CRM workflows. Existing customers are again pursuing digital-transformation projects, while Data Cloud and Agentforce have become the main areas of exploration after basic Sales Cloud and Service Cloud implementations. The Anthropic partnership was viewed as the most important Dreamforce development because it gives Salesforce a more tangible enterprise reasoning layer and reassures ecosystem participants that Anthropic is complementary rather than a route around Salesforce. However, Claudeforce remains new, initially focused on Sales rather than Service, and its economics are unclear. Partners expect customers to face Claude inference charges on top of Salesforce spending, creating potential concern about double payment. Whether customers accept that cost depends on demonstrable ROI, such as digital business-development agents generating incremental pipeline; contacts had not yet seen customers reach a mature renewal-and-invoice decision point. Rapid model changes also make credit consumption and exchange-rate assumptions difficult to forecast. The report argues that a headless, agent-driven interface does not necessarily undermine Salesforce's economic role. Human-seat trimming is occurring and may continue over time as users interact with agents instead of software interfaces, but contacts do not see a one-for-one connection between lower seats and lower contract value. Core CRM remains the system of record for customer data, workflows, governance and decision rules, while Data 360, Agentforce and inference could add monetization layers. In this view, Salesforce may increasingly operate as a back-end engine accessed through APIs or MCP, while the foundational CRM implementation remains necessary. Slack could become a digital front office, although partner deployment experience with Slack bots is still limited and Microsoft Teams remains a material competitive obstacle. Agentforce adoption is advancing but remains at an early stage. Contacts generally described experimentation with one or two agents, limited complex multi-agent deployments and many enterprise implementations still in pre-production. Production activity is reportedly increasing, especially beyond basic contact-center use cases, but enterprise readiness is constrained by foundational data work. One contact estimated average system-of-record accuracy at below 60%, illustrating why unreliable opportunity, customer-identity and governance data can prevent agents from delivering useful outcomes. Data 360 is positioned as a zero-copy, unified customer-data layer, but customers must assess whether the combined cost of Data 360, Agentforce and inference produces adequate returns. Salesforce's evolving consumption model creates both expansion potential and execution risk. Partners report free flex credits for experimentation and some three-year contracts that include Agentforce and Data 360. Yet swap clauses can allow customers dissatisfied with agent outcomes to reallocate flex credits toward seats. Token-based consumption could generate overage-style revenue as adoption grows, but it requires different sales incentives than conventional subscription selling, and packaging remains fluid. The report also notes that broad enterprise-agent adoption may still be years away because of security, compliance and permissioning requirements. HubSpot partner conversations show a mixed demand picture. Some partners reported strong inbound pipelines and practice growth of 20-25% year over year, while others described materially longer sales cycles because of macro uncertainty, executive-level scrutiny and AI confusion. EMEA contacts cited GDPR and data-privacy constraints as major barriers to broad AI adoption, particularly where customers cannot ensure data stays within the EU. Partner feedback suggests customer interest in AI remains high, but hesitation has increased versus the prior year. HubSpot AI credit-pack usage varies substantially by customer and use case. Customers are cautious about unexpectedly exhausting credits; initial implementation and automation setup can create a steep usage ramp, followed by a significant decline and then slower growth. Enterprise customers are more likely than SMB and mid-market customers to consume credits heavily, while poorly configured contact or company enrichment can burn large amounts unnecessarily. One partner characterized the customer base as bifurcated: about 10% exceed initial capacity while most use less than 5% of their credits. Partners view HubSpot as easier to use and less complex than Salesforce, supporting occasional migrations and faster adoption. Conversely, they also see product-depth limitations for complex enterprises, tougher competition from vertical software in SMB and mid-market, and a tendency for growing CRM users to evaluate Salesforce as they scale. HubSpot's multi-hub opportunity is broadening as integrations become more complex: one partner estimated customer attachment of 90% for Marketing Hub, 70% for Sales Hub and 40% for Service Hub, with Revenue Hub small but potentially reaching 60% of customers within 18 months. Revenue Hub was viewed as suitable for simple CPQ use cases but not yet adequate for complex billing, SKU and regional-discount requirements. HubSpot's data-retention policy episode generated frustration but, after the policy was withdrawn, partners reported no lasting impact on spending. The pre-quiet-period roundup identifies additional company-specific operating variables. Microsoft is described as retaining Commercial Cloud acceleration through E5/E7 upgrades, Copilot and new agentic consumption revenue. ServiceNow had exceeded $1B of AI ACV at the end of 2Q and remained on track for its raised $1.5B-plus CY26 target, although much current AI ACV reflects roughly 30% SKU-price uplift and consumption monetization is still early. Snowflake's AI contribution represented roughly half of its F2Q beat, while its migration opportunity remains large given FY26 Global 2000 average spend of about $2.4M versus roughly $34M for its top 25 product-revenue customers. Workday highlighted an AI-front-door architecture that retains core data and context in Workday, alongside a growth mix in which existing customers generate about 60% of incremental subscription revenue and new customers about 40%. CoreWeave's active-power outlook rose to 1.85 GW from 1.7 GW, characterized as a pull-forward from 2027. New-contract GPU pricing was running about 25% higher on a comparable-SKU basis, and management continued to see 5-10 points of margin expansion, although the timing of capacity costs and GPU monetization can produce quarterly volatility. Twilio reported messaging at about 60% of revenue and 18% growth excluding carrier fees in both Q1 and Q2, but retained conservative guidance because usage remains transactional. Dynatrace expects larger FY27 renewal cohorts to support a net-revenue-retention step-up from the current 110% level; logs consumption approaching $200M annualized is a key watchpoint. Paycom faces customer-acquisition weakness and near-term AI-related margin pressure, while monday.com's work-management platform remains about 88% of ARR and CRM about 10%, with larger customers accounting for 43% of ARR and growing faster than smaller accounts.

Analysis framework

JPMorgan synthesizes qualitative conversations with Salesforce implementation and CRM-ecosystem contacts, HubSpot partners and company-management commentary. It tests demand, AI adoption, pricing, product positioning and competitive dynamics through partner observations, then uses company-specific operating metrics and management frameworks to identify near-term variables ahead of quiet periods.

Methodology notes

• Other

  Partner and ecosystem channel checks

  The report uses conversations with systems integrators, CRM ecosystem contacts and HubSpot partners to assess customer demand, adoption behavior, deal pipelines, competitive positioning and product economics.

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  Enterprise AI adoption and consumption-monetization transmission

  The report follows how data quality, customer readiness, product packaging and usage behavior affect adoption of AI agents and ultimately subscription, inference and credit-pack revenue.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Salesforce

  Core CRM demand is recovering and AI/data products may add monetization layers.

  Strengths

  Complex CRM workflows, system-of-record data, governance and underlying metadata remain difficult to replace.

  Weaknesses

  Human-seat erosion, fluid packaging and uncertain Claudeforce economics.

  Comparison

  Viewed as more robust for complex enterprise needs than simpler CRM alternatives; HubSpot is regarded as easier to use.

  Risks

  Weak Agentforce ROI, poor customer data readiness, credit swaps and privacy or compliance constraints.

• HubSpot (HUBS)

  Partner checks address pipeline trends, AI-credit usage and multi-hub expansion.

  Strengths

  Ease of use, native product design and expanding Sales, Service and Revenue Hub adoption.

  Weaknesses

  Less product depth for complex needs and limitations in complex CPQ and billing use cases.

  Comparison

  Can win from Salesforce on usability, while larger or more complex customers may need Salesforce's robustness.

  Risks

  Longer sales cycles, GDPR-related AI hesitation, credit-use uncertainty and vertical-software competition.

• ServiceNow (NOW)

  AI ACV and future consumption monetization are highlighted.

  Strengths

  AI contributed roughly half of the F2Q beat; AI ACV exceeded $1B at the end of 2Q.

  Weaknesses

  Current AI ACV is predominantly driven by SKU upgrades rather than consumption.

  Comparison

  Identified as a durable back-end enterprise platform alongside Salesforce and Workday.

  Risks

  Consumption revenue is early and more material monetization is expected to be a 2027 story.

• Snowflake (SNOW)

  AI and migration demand are identified as growth drivers.

  Strengths

  AI contributed roughly half of the F2Q beat; substantial customer-spend expansion opportunity remains.

  Weaknesses

  Large-customer consumption behavior provided limited incremental clarity.

  Comparison

  Data consolidation may occur in Snowflake Cortex, Salesforce Data 360, Databricks or other platforms.

  Risks

  Lower consumption at its largest customer and execution against migration monetization.

• CoreWeave (CRWV)

  Capacity, pricing, contract structure and margin progression are central variables.

  Strengths

  Higher active-power outlook, roughly 25% higher comparable-SKU new-contract pricing and favorable demand.

  Weaknesses

  Quarterly results can be volatile as capacity costs precede GPU monetization.

  Comparison

  Shorter two- to three-year contracts broaden customer opportunities relative to five-year commitments.

  Risks

  Deployment timing, contract-renewal downside and current deal margins nearer the lower end of the 25-35% range.

• Twilio (TWLO)

  Messaging remains the main revenue driver.

  Strengths

  Messaging is about 60% of revenue and grew 18% excluding carrier fees in both Q1 and Q2.

  Weaknesses

  Transactional usage limits visibility despite minimum commitments.

  Comparison

  Recent strength was broad-based rather than linked to one use case or one-time item.

  Risks

  Management does not consider recent 5%+ quarterly beats a new normal.

• Dynatrace (DT)

  Renewals, consumption and sales execution underpin expected NRR improvement.

  Strengths

  TTM net-new ARR growth reached 17% in 1Q and logs consumption is approaching $200M annualized.

  Weaknesses

  Logs remain a relatively small contributor to current ARR.

  Comparison

  Consumption growth above 20% is running ahead of ARR growth.

  Risks

  Renewal conversion and deal timing remain key variables.

• monday.com (MNDY)

  The report highlights product-mix and larger-customer expansion.

  Strengths

  Work management remains about 88% of ARR; customers above $50K ARR represent 43% of ARR and grow faster.

  Weaknesses

  CRM remains a smaller portion at roughly 10% of ARR.

  Comparison

  Sales resources are shifting toward larger customers and installed-base expansion.

Key data

• Salesforce partner pipelineMajor uptick over the last 3-4 monthsReported after business was flat from last year through June.
• CRM system-of-record accuracyLess than 60%A partner's average estimate illustrating the data-quality constraint on agent performance.
• HubSpot credit-pack power users~10% of customersPartner estimate of customers going well above initial capacity; most reportedly use less than 5% of credits.
• HubSpot multi-hub attachment90% Marketing Hub; 70% Sales Hub; 40% Service HubRough partner estimates; Revenue Hub could reach 60% of customers in 18 months.
• ServiceNow AI ACV$1B+ at the end of 2Q; $1.5B+ CY26 targetThe $1B figure is cumulative ACV, with Moveworks contributing more than $100M of the 1H increase.
• CoreWeave active power outlook1.85 GWRaised from 1.7 GW and described as capacity pulled forward from 2027.
• Dynatrace annualized logs consumptionApproaching $200MReached after crossing $100M two quarters earlier.
• monday.com work-management ARR share~88%CRM is tracking to roughly 10% of ARR.

Impact & implications

The report's evidence supports a view that core enterprise platforms remain embedded despite AI-disruption concerns, while AI may add consumption and data-platform revenue over time rather than immediately replace application spend. The pace and durability of that monetization depend on proof of ROI, data quality, transparent packaging, customer privacy requirements and the timing of production deployments.

Risks

• Salesforce AI products may not demonstrate enough ROI to justify additional inference, Data 360 and Agentforce spending.
• Poor data quality, governance and access may limit enterprise-agent performance and delay production deployment.
• Salesforce packaging, flex-credit exchange terms and consumption economics remain unsettled.
• HubSpot faces lengthening sales cycles, customer AI hesitation and GDPR-driven data-privacy restrictions in Europe.
• CoreWeave results can be volatile because capacity spending and GPU monetization do not always align by quarter.
• Twilio's usage-led messaging model remains inherently variable despite larger-customer commitments.

What to watch

• Customer Claudeforce invoices, renewal behavior and evidence that agent-led workflows generate measurable ROI.
• Salesforce Agentforce production deployments, data-foundation progress and the final form of packaging and credit economics.
• HubSpot deal-cycle duration, AI-credit consumption patterns and adoption of Revenue Hub and other non-Marketing Hubs.
• ServiceNow's transition from SKU-driven AI ACV to consumption monetization and progress toward its $1.5B-plus CY26 AI ACV target.
• CoreWeave capacity delivery, GPU-contract pricing, margin expansion and deployment timing.
• Dynatrace's second-half FY27 renewal cohorts, logs-consumption conversion and NRR progression from 110%.