Monday.com AI Customer Research: How the $7B Work OS Captures Voice of Customer in 2026

TL;DR#

Monday.com is a $7B+ Work OS that hit $1.23B in 2025 revenue with 250,000+ paying customers and 4,281 enterprise accounts at >$50K ARR. The company runs four productized verticals — monday work management, monday CRM (which crossed $100M ARR three years post-launch), monday dev, and monday service (GA in 2025, 215,000+ tickets resolved since January 2024) — and is now repositioning the entire platform as an "AI Work Platform" with native agents, sidekick, vibe, and workflows. That productization is also monday.com's biggest voice-of-customer problem: four verticals means four distinct buyer personas, four roadmaps, and four PMM teams competing for a finite research budget. Monday.com's published playbook leans on Customer Success Ambassadors and support-ticket tagging to surface signal, but that approach scales linearly with headcount — exactly the bottleneck conversational AI customer research is built to solve. Companies running multi-product Work OS strategies need persona-segmented voice-of-customer pipelines that scale with the catalog, not the research team.

Monday.com AI Customer Research: Why a $7B Work OS Needs Segmented Voice of Customer#

Monday.com is no longer one product. As of fiscal year 2025, the company reports $1.23B in revenue, 27% year-over-year growth, and a 14% non-GAAP operating margin. It serves more than 250,000 paying customers across 200+ industries. But the durable story isn't headline ARR — it's the shape of the catalog. Monday.com now sells:

• monday work management — the original horizontal Work OS
• monday CRM — sales-focused, hit $100M ARR in three years
• monday dev — engineering/product team workflows
• monday service — enterprise service management, GA in 2025
• monday AI — sidekick, vibe (which crossed $1M ARR within 2.5 months of pricing launch in October 2025), agents, and workflows

That's four distinct productized verticals plus a horizontal AI layer. Each one targets a different buyer (Head of Sales, Head of Engineering, Head of IT, Head of Operations), runs against different competitors (Salesforce, Jira, ServiceNow, Asana), and demands its own voice-of-customer signal. The customer research challenge isn't "what do monday.com users want?" — it's "what does the buyer of monday service want, in their own language, and how does that overlap or conflict with what the buyer of monday CRM wants?"

That is a problem the form-and-survey VoC stack was never designed to handle. It's also the problem this post is about.

The Productization Trap: One Brand, Four Roadmaps, One Research Team#

The productization trap hits every Work OS that successfully verticalizes. Monday.com's catalog expanded from one to four+ products in roughly five years, but the customer research function did not 4x in headcount. That's not a monday.com-specific failure — it's the default state of every horizontal SaaS company that productizes. We've seen the same pattern at the $5B work-management leader Asana, at ClickUp's all-in-one productivity platform, and at Atlassian's Jira-Confluence-Loom triangle.

The structural problem looks like this:

Each row in that table needs its own "why now," its own jobs-to-be-done, and its own competitive frame. A single annual customer survey blended across all five rows produces averages that are wrong everywhere. A monday service customer who wants better SLA reporting and a monday CRM customer who wants AI lead scoring are not the same data point — but the typical horizontal-saas VoC stack treats them as one.

How Monday.com Captures Voice of Customer Today#

Monday.com's published playbook for capturing customer signal centers on two mechanisms: Customer Success Ambassadors and structured support-ticket tagging.

The Customer Success Ambassador role embeds CS team members directly into product domains. As ProductLed has documented, these ambassadors "bring information from customers into product domains and have a clear understanding of the voice of the customer." They function as a translation layer between client-facing teams and R&D. Every support ticket gets tagged by product area and question type, giving Research & Development a quantitative read on where friction concentrates.

That model has real strengths. It builds tacit knowledge, it forces CS-product collaboration, and it produces a continuous trickle of qualitative signal. But it has three structural ceilings that get worse as the product catalog expands:

1. Ambassador headcount is the bottleneck. A Customer Success Ambassador can hold maybe 100–200 customer relationships in their head. With 250,000+ paying customers and four productized verticals, even a generous Ambassador-to-customer ratio leaves >99% of the base outside the signal loop.
2. Support tickets surface dissatisfaction, not aspiration. Tagged tickets tell you what's broken. They don't tell you what someone almost bought, what made them choose monday CRM over Salesforce Starter, or what would make them upgrade from work management to the full suite. The highest-leverage VoC questions never become tickets.
3. The signal is implicit, not transcribed. Ambassador insights live in Slack channels, one-off calls, and quarterly R&D syncs. There's no searchable corpus of "what 500 monday dev customers said about GitHub depth this quarter." That makes the signal hard to re-aggregate when a new PMM joins or a vertical pivots.

Every Work OS hits this wall. Linear hit it with its feedback-strategy redesign. Notion hit it on its way to a $10B valuation. The fix isn't more ambassadors — it's a different research substrate.

What "AI-First Customer Research" Means Inside a Work OS#

In February 2025, monday.com announced its AI Vision built on three pillars — AI Blocks, Product Power-ups, and the Digital Workforce. The May 2026 announcement went further, repositioning the company from "Work Management Platform" to "AI Work Platform" with monday agents and one-click connectors to Claude, Microsoft 365 Copilot, and OpenAI's ChatGPT.

That positioning creates a consistency test: a Work OS pitching AI agents to its customers needs to be using AI-native research to understand those customers. AI-first customer research, applied to a Work OS, looks like this:

• Persona-segmented interview at scale. A continuous AI interview branches on vertical at minute one — "primarily using monday for project management, CRM, dev, or service?" — then probes vertical-specific JTBD for the rest of the conversation.
• Conversation-as-source-of-truth. Every conversation is transcribed and queryable. A PMM rolling onto monday service can ask "what did our last 300 service customers say about ticket triage" and get a real answer.
• Probe on uncertainty. When a monday CRM trial-er says "I'm not sure if this beats Salesforce Starter," the AI follows up — comparison frame, blocking feature, team size — rather than collapsing the highest-value moment into a radio button.

This is the bet Figma made with its AI customer research strategy and Airtable made with its template-library-driven conversational discovery. For monday.com, the strategic upside is larger because the catalog is broader.

The Conversational VoC Opportunity, Vertical by Vertical#

Each productized vertical has a distinct conversational VoC opportunity:

monday CRM. The question to answer is "what makes a trial-er upgrade from monday work management's CRM-shaped workflow into the dedicated monday CRM, and what makes them defect to HubSpot or Salesforce instead?" That answer lives in language like "my AE pushed back on the pipeline view" and "I can't justify another seat license." Conversational interviews capture that texture; multiple-choice surveys flatten it into nothing.

monday dev. The hard question is "where does monday dev win and lose against Jira and Linear in the engineering buyer's head?" The answer rarely fits a 1–5 scale. It looks like "we picked monday dev because the PM already lives in monday, but our staff engineers wanted Linear and we fought it out."

monday service. The newest vertical and highest-leverage VoC target. GA in 2025 with 215,000+ tickets resolved, but the IT-lead buyer is a different persona from the original work-management ICP. Monday.com needs a continuous interview pipeline capturing why customers chose monday service over ServiceNow or Zendesk, what ITSM workflows are missing, and where agent deflection actually moves the needle.

monday AI / monday agents. monday vibe hit $1M ARR within 2.5 months of pricing launch — signal exists, but the population is small and use cases are emerging. Forms ask "are you using monday agents?" and get a yes/no. AI interviews ask "walk me through the last agent you built and what you wanted it to do" and get the dataset needed to ship the next ten features.

A single VoC pipeline cannot serve all four verticals well. Persona-segmented conversational research can.

Why This Matters Now: The AI Work Platform Bet#

The May 2026 repositioning — Work Management Platform to AI Work Platform — is the largest strategic move in monday.com's history. Repositioning a $1.23B company around AI agents requires evidence that the new positioning resonates with the 250,000-customer base, with the 4,281 enterprise accounts, and with the prospective buyer who hasn't yet committed.

You don't generate that evidence from quarterly NPS. You generate it from sustained, persona-segmented conversation with thousands of customers — what they expected from "AI agents," what they actually used, what made them upgrade, what made them churn. That's a multi-vertical research operation, not a single survey, and it's the workload we built Perspective AI to handle.

McKinsey's State of AI 2024 report found 65% of organizations now use generative AI regularly — double the share from ten months prior — but the gap between adopting AI and capturing measurable business value remains the dominant pattern. The companies that turn AI from a feature line into a roadmap-shaping research substrate are the ones that compound.

What Monday.com Should Build Into Its 2026 Research Stack#

1. One conversational interview pipeline per productized vertical. Work management, CRM, dev, service, and AI each get their own always-on AI interviewer aimed at trial-ers, new customers, expanders, and churners — five pipelines, persona-segmented, running continuously.
2. AI-segmented at minute one, vertical-specific from minute two. The interviewer branches on "what's the primary product you're evaluating" and runs vertical-specific JTBD probes thereafter. PMMs get a vertical-tagged corpus instead of a blended average.
3. Replace the post-trial NPS form with a 5-minute conversational debrief. Trial-to-paid is monday.com's most important conversion event. Replacing the exit survey with a conversation that probes "why now" and "why not" also generates qualitative training data for the sidekick layer itself.
4. Amplify the Customer Success Ambassador workflow. Ambassadors don't go away — they gain a searchable conversation corpus instead of relying on memory and quarterly syncs.
5. Make the research itself a product power-up. monday service IT leads and monday CRM RevOps leads also need conversational customer research inside their own organizations. The same pipeline that serves monday.com's internal PMMs is, at scale, a use case for the AI Work Platform to host.

The throughline: the AI Work Platform pitch only works if monday.com itself runs an AI-first customer research function.

Frequently Asked Questions#

What is monday.com's AI strategy in 2026?#

Monday.com's 2026 AI strategy is to reposition from "Work Management Platform" to "AI Work Platform," anchored on four AI Work Capabilities — monday sidekick, monday vibe, monday agents, and monday workflows — plus one-click connectors to Claude, ChatGPT, and Microsoft 365 Copilot. The May 2026 announcement is the largest strategic shift in the company's history, betting that native AI agents become the primary mode of work across the Work OS.

How does monday.com capture voice of customer today?#

Monday.com captures voice of customer through Customer Success Ambassadors embedded in product domains and through structured support-ticket tagging that maps every ticket to product area and question type. This approach surfaces continuous qualitative signal but is bottlenecked by Ambassador headcount, biased toward dissatisfaction rather than aspiration, and stored as implicit knowledge rather than a searchable corpus.

Why is multi-product VoC harder than single-product VoC?#

Multi-product VoC is harder because each productized vertical — monday work management, CRM, dev, service, and AI — targets a different buyer persona, competes against a different incumbent, and demands a separate research operating cadence. A single blended survey across all verticals produces averages that misrepresent every individual vertical. Persona-segmented conversational research is the only approach that scales with the catalog rather than with research-team headcount.

What does monday CRM compete with?#

Monday CRM competes primarily with HubSpot, Salesforce, and Pipedrive in the SMB and mid-market. It hit $100M ARR three years post-launch by leveraging the existing monday work management customer base and offering a lighter-weight CRM experience than Salesforce's enterprise stack. The strategic VoC question is what drives a monday work management customer to adopt monday CRM versus a competitor.

How does Perspective AI fit a Work OS like monday.com?#

Perspective AI runs always-on AI-powered customer interviews segmented by persona and vertical, transcribing every conversation into a searchable corpus that PMMs and product managers can query directly. For a multi-product Work OS, that means one pipeline per productized vertical, AI-segmentation at minute one, and a corpus that scales with the catalog rather than with Customer Success Ambassador headcount.

Conclusion: monday.com AI Customer Research at $7B Scale#

Monday.com is a case study in productization-driven VoC complexity. The company that hit $1.23B in revenue and four productized verticals by 2025 cannot continue to run its research operation on Customer Success Ambassadors and tagged support tickets alone. The May 2026 repositioning to AI Work Platform raises the stakes further: a company telling its customers to adopt AI agents needs to be using AI-native research to understand whether that pitch is landing.

The monday.com AI customer research stack of 2027 won't be "one survey across 250,000 customers." It will be five persona-segmented conversational interview pipelines — one each for work management, CRM, dev, service, and AI — running continuously, transcribed and tagged, queryable by PMMs, and feeding directly into the product roadmap. That's what conversational voice-of-customer at Work OS scale looks like.

If you're building the research operating model for a multi-product SaaS company and want to see what AI-first persona-segmented customer interviews look like in production, start a Perspective AI research project, explore how Perspective AI works for product teams, or meet the AI interviewer agent. The catalog scales. The research substrate has to scale with it.