Salesforce admits AI rollout was messy, quota stays flat at Mangomint
Three revenue leaders got honest about AI in sales at SaaStr AI London. Greg Beltzer (CCO, Salesforce Agentforce), Ashley Wilson (COO, Momentum), and Marchelle Mooney (VP Sales, Mangomint) came from wildly different contexts: Salesforce has tens of thousands of sellers, Momentum is a 50-person startup, Mangomint's AEs close 20 SMB logos monthly. But they kept landing on the same reality checks.
AI rollouts are rough, even at Salesforce
Beltzer was blunt: Agentforce deployment at Salesforce itself was not smooth. Data was not clean, processes were not ready, ways of working got rewritten on the fly. If Salesforce's CRM data was not AI-ready, yours probably is not either.
The insight: AI deploys easier on service and ops than sales. Service already has structured data and documented workflows. Sales relies on individual sellers, and most of what they know lives in their heads or text messages.
The 40% of leads nobody touches
Salesforce gets massive inbound volume. A shockingly low percentage got follow-up. Reps cherry-picked leads, the rest got fake loss reasons with no notes. Post-SaaStr Annual, hundreds of leads from people who wanted to sponsor just sat there. Not because reps are bad, because humans prioritize and the bottom 40% always gets ignored.
Beltzer's number: leads worked by agents led to direct revenue they would not have had, because nobody was working those leads at all. This is the easiest AI use case in sales right now. Put an agent on the leads nobody is working.
Your CRM data is worse than you think
Mooney rolled out Momentum and realized her Salesforce instance was a disaster. Deals closed with no notes, no call logs, no contact records. Her top rep closed 35 logos monthly (against a 20-logo quota), absolute machine, but everything lived in text messages. Another rep had to explain to his wife why he had so many women in his texts (he was new to selling to salon owners).
If data never made it into your CRM, your AI agent cannot reconstruct it. First step is not deploying agents, it is deploying tools that ensure data flows into Salesforce going forward.
Quota is staying flat
Mooney announced she is not raising quota in 2026. Her top performers hit 35 logos against a 20-logo target. That is a 7x ARR-to-OTE ratio in SMB. But she is keeping quota flat and using AI to lift middle performers toward top-performer levels.
Her logic: we are not at the point where AI is truly native in sales workflows. Raise quota now, you are asking reps to do more before you have given them systems to do more. Wilson backed this up from Momentum research: companies winning with AI in sales are investing heavily in ops and rev ops headcount to run agents, manage data, and architect systems. The efficiency gains only happen if the business supports the reps.
Worth noting: Salesforce has major ANZ presence through local offices, Momentum and Mangomint have no disclosed ANZ headcount. The lessons apply regardless of market, but comp and quota expectations will differ locally.