AI Firms Rebuild Customer Success Playbooks Across B2B

At the Forward Deployed / Customer Success track of SaaStr AI 2026, leaders from Lovable, Harvey, and Assembly AI described large-scale changes to post-sales motions, arguing existing SaaS-era customer success playbooks are now liabilities, according to reporting by SaaStr. SaaStr reports the CSM role grew more than 700% through Q2 2022 then flatlined for four years, while forward deployed engineering rose over 1,000% in the same window. Specific examples in the session: Ryan Seams of Assembly AI said technical buyers reacted poorly to the title "head of customer success," and that renaming the role to "forward deployed engineer" produced two candidates in 2.5 months where the older title did not, per SaaStr. Monica Perez of Lovable said she stopped leading with AI entirely. Tom Ronen of Harvey described using heavy on-site EBRs and change-management frameworks at an $11B company and cited a Harvard study of 1,515 startups showing a single change-management document correlated with 2x revenue, as reported by SaaStr.
What happened
SaaStr reported that the Forward Deployed / Customer Success track at SaaStr AI 2026 featured speakers from Lovable, Harvey, and Assembly AI arguing that the conventional post-sales playbook built during the SaaS era is now a liability. Per SaaStr, the CSM role expanded by more than 700% through Q2 2022 then flatlined for four years, while forward deployed engineering increased by over 1,000% in the same period. The session included on-stage examples: Ryan Seams (Assembly AI) described technical buyers going defensive when hearing "head of customer success," and said renaming the role to "forward deployed engineer" yielded two candidates in 2.5 months versus a fuller pipeline for the older title, according to SaaStr. Monica Perez (Lovable) said she stopped leading with AI entirely. Tom Ronen (Harvey) described running executive business reviews and on-site change management at an $11B company and cited a Harvard study of 1,515 startups where providing one change-management document was associated with 2x revenue, as reported by SaaStr.
Editorial analysis - technical context
Companies selling AI into enterprises face a different adoption surface than traditional SaaS. Industry-pattern observations: AI deployments often require workflow redesign, security and compliance work, and forward deployment of engineers to integrate models into business processes. Observers note that role names and hiring signals influence recruiter pipelines and buyer trust in technical sales interactions. Replacing generic CSM language with more technical or delivery-focused titles is one tactic seen across high-growth AI vendors, per the session anecdotes.
Industry context
Observers tracking B2B AI adoption should see the session as highlighting a broader shift from usage metrics to change-management outcomes. Industry-pattern observations: metrics like NPS and simple activity scores can break down when well-used systems do not alter underlying business processes. Vendors and customers increasingly emphasize adoption tied to measurable workflow change rather than seat utilization alone.
What to watch
For practitioners and buyers, watch hiring posts and job titles (for example, increases in "forward deployed engineer" roles) and whether customer-facing teams embed formal change-management artifacts into deployments. For product and success teams, track whether vendors substitute traditional health metrics with outcome-oriented measures and whether post-sales orgs increasingly include engineering-heavy, on-site engagement models.
Source attribution
All reported facts and quotes in this summary are drawn from the SaaStr session report published on saastr.com.
Scoring Rationale
The story documents a notable operational shift in how B2B AI products are delivered and supported, which matters to product, success, and deployment teams. It is a significant practitioner-level trend but not a frontier research or platform release.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems
