Saar Arbel

Role

Builder/owner of “the Brain” — an LLM-maintained company context OS being prototyped inside monday. Currently in early/exploratory phase (“bosr” — immature) but with conviction the underlying problem is real and the market is willing to spend.

Affiliations

• monday — building the Brain prototype

Notes

• Hosted the 2026-05-06-brain-os-strategy-brainstorm session.
• Working theory: the Brain must be agent-native from day 1, since within 1–2 years most work will be done by agents on behalf of humans. The product is not “do work for you” — it’s “make your agents better at doing your work.”
• Acknowledges the prototype currently won’t scale to a large org; has been thinking about scale/rags solutions overnight but is confident the engineering is solvable.
• Believes the differentiation challenge is sharpening the value proposition vs. competitors like glean, not the engineering.
• Updated framing as of 2026-05-27: explicitly stepped off the agentic-org horizontal pitch (“everyone is approaching that”) and onto vertical use-case-led Brain. Anchor anecdote: a recent meeting with ari-leshno (an ophthalmologist) where Brain-as-context plausibly transforms clinic operations. See vertical-use-case-led-brain.
• Sceptical of the prediction-markets-regulatory-play tangent — argued winner-toto is “a well-oiled machine, statisticians and all” — but did not block exploration.
• Heavy critic of agent slop: the monday-agents-month catalog is “mostly slop”; the value test is whether a real painful workflow gets unlocked.
• Wedding upcoming (date not captured); referenced in the meeting as light agenda context.
• 2026-06-20: forced the discussion back from Brain architecture to problem discovery, arguing that the Brain is a tool and that the week should be about finding concrete pains.
• 2026-06-20: reported speaking with roughly 15 monday people and seeing many Brain use cases, each specific to a local workflow rather than a universal product.

Mentioned in

• 2026-05-06-brain-os-strategy-brainstorm — host; presented the prototype and led discussion on agent-native positioning.
• 2026-05-19-saar-leshno-glaucoma-data-tracking — met ari-leshno (intro by Ohad) to scope the glaucoma visual-field-tracking opportunity; the conversation became the anchor anecdote one week later.
• 2026-05-27-directions-vertical-pivot-and-prediction-markets — co-driver of the vertical-use-case-led pivot; introduced the ari-leshno anecdote.
• 2026-05-31-directions-connect-businesses-and-brain-scaling — demoed a Brain-on-Claude-project ideation tool with a 10-criteria scorecard for strategic directions; will provision Guy + Nizan as Brain repo collaborators and pair-install with Guy this week. Reinforced that vertical determines the data model.
• 2026-06-02-directions-account-management-vertical — demoed the daniela AM Brain (built from her last ~100 call transcripts); explained the index.md + graph-traversal architecture that keeps performance tractable at current scale; explained prior revenue-intelligence work at Monday; began the Harmony AM pilot.
• 2026-06-13-directions-revenue-cycle-and-construction-poc — introduced the broader revenue-cycle-brain thread, described Monday/Harmony closed-loop GTM experiments, and pushed the construction design-partner path toward a demoable POC.
• 2026-06-14-directions-from-home-strategy-brain-dogfooding — pushed the problem-first discipline: Brain is a powerful capability, but the team should not force it as the product before validating real AM, revenue-cycle, or construction pain.
• 2026-06-15-directions-ofer-polivoda-am-operator-interview — joined the ofer-polivoda operator interview; probed AM coverage, unmanaged/scaled accounts, and prep-time distribution.
• 2026-06-16-directions-ran-levy-fiverr-cs-operator-interview — joined the ran-levy / fiverr CS interview; framed the team’s AM/CS exploration and surfaced the post-call question of whether Fiverr’s model is a different, scaled-CS pattern.
• 2026-06-16-directions-ofri-avivi-renewals-operator-interview — joined the ofri-avivi renewals interview; asked about account coverage/QBR scale and helped ground renewals as another AM/CS workflow archetype.
• 2026-06-20-directions-planning-the-week-18-30 — pushed problem-first discipline, proposed ~10 weekly user interviews, and took ownership for updating the Founder-Market Fit table.
• 2026-06-21-directions-roy-keren-salesforce-gtm-ai-interview — joined the roy-keren operator interview used to pressure-test the AM/revenue-team AI wedge against the Salesforce/Slack incumbent baseline.