Why Depth Wins: Dom Dinardo's Case for Vertical AI
Key points
Key takeaways from a YouTube interview with Aforza Founder & CEO Dom Dinardo (Septemberer 2026):
Transcript unavailable — summary based on title and description only.
Three software waves. Dinardo, who joined Salesforce in 2002 before "cloud computing" was a widely used term, sees modern software history as three waves: horizontal SaaS dominated the 2000s, vertical SaaS emerged in the 2010s,and he expects the 2020s to belong to vertical AI — products built around the workflows, data and expertise of individual industries.
Vertical intimacy. His time at Veeva taught him why going deep into a single industry can produce a completely different and far more powerful software business — a concept he calls "vertical intimacy".
The $30B consumer goods bet. That insight is why he chose the roughly $30B consumer goods market for Aforza, founded in 2019, betting that AI built around industry data and workflows will reshape how FMCG brands operate.
From SaaS to vertical AI. Aforza evolved its platform from traditional SaaS into vertical AI applications,and he argues generic AI copilots often struggle to demonstrate ROI because they lack the workflows, data and domain expertise of a specific industry.
Pricing without tokens. He discusses pricing vertical AI without selling tokens, anchoring value in workflow outcomes rather than per-seat chat usage.
Beyond the chat interface. The best enterprise AI, he argues, may use photos, voice and actions instead of chat alone — embedding AI into how work actually happens in an industry.
Enterprise-scale deployment. The episode covers the enterprise workflow AI can completely change,and how to scale vertical AI deployments across 40+ markets while making implementations actually work.
Threatened incumbents. He also considers which SaaS companies are most threatened by AI,and reflects on the biggest lesson learned at Veeva: going deep on one industry rather than spreading broadly.
Read more: YouTube · Verticals: A Weekly Biz Show
