The rise of consumer AI agents that perform tasks end-to-end is setting a new standard. B2B products must evolve from simply providing insights ('help me do this') to executing entire workflows ('do this for me') to meet rising user expectations.
Unlike software with discrete feature releases, AI capabilities are updated continuously in the background. Product teams must build mechanisms to constantly re-educate users on what the tool can now do, as its evolution is invisible to them and requires overcoming the 'blank page problem' repeatedly.
Contrary to assumption, Toast discovered its least tech-savvy customers were quickest to adopt its AI assistant. Natural language interfaces allowed them to bypass learning complex software workflows, making them power users simply by asking questions in plain English.
To de-risk launching AI into its mature platform, Toast created a high-feedback loop with a small group of 'design partners' in a WhatsApp group. This allowed for rapid, contained iteration for nearly a year, ensuring the product was valuable and stable before scaling.
Toast's VP of Product argues that building with AI is evolving so rapidly that leaders will become ineffective if they don't stay hands-on. She intentionally takes on Individual Contributor (IC) roles to remain close to the changing development process and tooling.
When a new interaction model appears in a consumer AI app, a clock starts. B2B users will soon expect the same capabilities in their work tools. Product teams must treat these consumer trends as a timer, needing to build similar functionality to stay relevant and meet expectations.
Unlike AI-native startups that can ship experimental features, established platforms like Toast must meet a high bar for quality and accuracy from day one. Existing users have established workflows and trust, which can be easily burned by unreliable AI additions.