The Last Mile: Inside Monday.com’s Shift to Embedded AI Engineering

In late July 2026, Monday.com executed the deepest cut in its history, laying off roughly 20 percent of its workforce (about 620 employees, including 350 at its Tel Aviv headquarters). In its second-quarter financial report, the company recognized $21.4 million in restructuring charges to build what Co-CEOs Roy Mann and Eran Zinman termed a “faster, flatter company” built around its AI Work Platform.

Three months later, that new organization is taking shape… as can be seen through their hiring boards.

During October, active job requisitions at Monday.com more than doubled. But Monday is not backfilling the self-serve product and middle-management roles it eliminated over the summer.

Instead, the longtime textbook case study of Product-Led Growth (PLG) is building an enterprise army of Forward Deployed Engineers (FDEs).

Monday earned that reputation by mastering frictionless software adoption. While legacy giants like Oracle or SAP required long sales cycles and consultants, Monday flooded web feeds with self-serve ads. Teams put down a credit card, set up a colorful project board, and invited coworkers without ever talking to a salesperson.

Now, that hands-off model has hit a wall. To sell autonomous AI agents to large enterprises, SaaS companies are discovering that self-serve software is not enough.

The 29 Percent Integration Wall

For a decade, cloud software economics were defined by near-zero marginal cost. You built code once in Tel Aviv, hosted it on AWS, and let customers onboard themselves. Gross profit margins sat comfortably near 90 percent.

Autonomous AI agents break that playbook entirely.

On October 21, Monday published a global study of 1,701 senior managers across seven countries, conducted with Nielsen: “Squinting in the Sun”. The report revealed a glaring organizational disconnect:

• 96 percent of enterprise leaders use AI at work, with 73 percent using it daily.
• 58 percent expect to manage more autonomous AI agents than human direct reports within twelve months.
• Only 29 percent have integrated AI into a central operating system with shared data libraries and cross-departmental workflows.
• 57 percent reported that the number of standalone software tools their teams use has actually increased since adopting AI.
• 46 percent worry about losing control over autonomous agent actions.

A human employee can adopt self-serve software alone because humans absorb organizational friction. When data is messy, people copy and paste across browser tabs.

Autonomous agents cannot run on copy and paste. When an enterprise deploys Monday’s AI Workflows and its new fleet of autonomous AI agents (such as Pipeline Guardian for CRM governance, Dependency & Risk Mapper for cross-team project risks, or Invoice Reconciliation Agent for accounting audits), those bots require direct access to core enterprise systems. Chief Information Officers, worried about compliance breaches and data leaks, will not grant autonomous code access to financial databases through a self-serve credit-card signup.

Big Tech firms like Google, Meta, and Apple have long employed customer-facing technical specialists, often titled solutions engineers or partner engineers. But modern enterprise AI platforms like Palantir and Wonderful AI have demonstrated that when autonomous agents meet messy corporate infrastructure, standard pre-sales consulting falls short. Dedicated engineers must embed directly with clients to own the technical outcome from discovery to production.

Monday’s October hiring surge reveals that new deployment machine in action. Roughly one third of open requisitions are for Forward Deployed Engineers on Monday’s careers portal dedicated to just one or two enterprise accounts, tasked with building custom agents, vibe apps, and wiring Model Context Protocol (MCP) integrations. Another third are infrastructure and backend engineers in Tel Aviv and London scaling the underlying MCP framework to ensure external models like Claude and ChatGPT can interact with Monday boards within enterprise security boundaries. The final third are Agentic CRM go-to-market teams and recruiters focused on onboarding global systems integrators.

The Unit Economic Tradeoff

Deploying engineers to hand-hold one or two accounts at a time presents a classic scaling dilemma for a software company whose official Q2 earnings report showed overall Net Dollar Retention at 109 percent, while retention for enterprise customers spending over $50,000 and $100,000 in ARR held at 115 percent (and the number of paid customers spending over $500,000 in ARR grew 68 percent year-over-year to 114). Monday may be betting its new FDEs will act as field technicians and sensors: solving customer-specific roadblocks today, and feeding empirical evaluations back to Tel Aviv to turn those solutions into reusable platform features tomorrow.

When reached for comment regarding the strategic deployment shift, Monday stated that it had nothing to share beyond what has already been publicly disclosed. Investors and industry observers will be watching closely when Monday reports third-quarter financial results on November 9, as similar high-touch enterprise pivots begin to ripple across the SaaS world. In the agentic era, machines do not buy software on credit cards, and winning the enterprise means sending human engineers into the room to wire the pipes by hand.