Intent engineering replaces prompt engineering for AI-driven revenue workflows

Aug 10, 2026 · Lenny's Podcast
🎧 PodShort 43 min squeezed to 3 AI SprinklerAS Sales Tech New
Episode artwork
Grace Clark
AI Teacher, Former Marketing Consultant at Grace Clark
Lenny's Podcast
43 min squeezed to 3
Full episode from Lenny's Podcast
Quotable Moments

Prompt engineering is dead, but intent engineering is where we need to be focusing our time.

Grace ClarkAI Teacher, Former Marketing Consultant at Grace Clark
00:00

If every touchpoint that you do is not AI first, then how can you convince people that every touchpoint they should do should be AI first?

Grace ClarkAI Teacher, Former Marketing Consultant at Grace Clark
09:00

No more prompting, be conversational, give it your outcome, and then let these powerful models help you and carry you and teach you how to work with them. They will rebuild Gmail for you, and you'll never have to go into Gmail ever again.

Grace ClarkAI Teacher, Former Marketing Consultant at Grace Clark
26:25
Key Insights
  • Prompt engineering is dead; intent engineering is where revenue leaders need to focus their time. This means communicating desired outcomes and problems to AI rather than prescriptive prompts.
  • The biggest hurdle to AI adoption is building the muscle memory of simply opening an AI app and defaulting to it for tasks. People need to learn to collaborate with AI rather than over-directing it.
  • AI democratizes work by enabling individuals to build custom, automated processes. This includes building personal AI agents to manage emails, generate proposals, and track personal tasks, freeing up significant time spent on manual admin.
  • Every touchpoint within your business, both internal and external, should be AI-first. If you, as a leader, aren't using AI for your own workflows and customizations, it's difficult to convince your team and clients to adopt it.
  • AI forces businesses to define 'ideal workflows.' By documenting how a process *should* ideally work step-by-step, even if done 'mechanically' before, AI can then execute and automate it, leading to a better performing business.
  • Rebuilding Gmail with AI is a universally applicable project. It allows for personalized, branded communication that learns from interactions, transforming email from a 'locked-away' system into a compounding learning asset for your AI.
  • The easiest way to get started with AI, especially when feeling overwhelmed, is to create a 'forcing function': set a reminder to screenshot a frustrating task and ask your AI (e.g., Claude) for help. This builds the habit of deferring to AI.
  • AI tools can handle complex workflows from initial problem statement to iterative refinement. By clearly stating the problem and desired outcome, the AI can 'reverse-engineer' solutions, allowing humans to focus on collaboration rather than prescriptive prompting.
Metrics Mentioned
  • 20 hours per week (Time spent on administrative tasks for teaching, before using AI automation.)
  • 45 minutes (Potential time to generate a sales proposal if AI use is optimized.)
  • 30-person team (The size of team where gathering qualitative feedback via traditional methods (Slack DMs, 1-on-1s) becomes unfeasible, highlighting the need for AI-powered questionnaires.)
  • 10 minutes of talking, 1 hour of Claude creating (Initial time spent to generate an interactive, password-protected artifact, with human feedback.)
  • Once an hour (Frequency at which Grace Clark's AI pipeline operator wakes up to check for tasks and manage client progression.)
  • Half an hour (Estimated time for anyone to rebuild their Gmail experience with AI, potentially longer for advanced customization.)
  • 4 agents in a single afternoon (A HyperAgent user built an entire outbound sales pipeline (prospecting, outreach, follow-ups, CRM updates) using four AI agents within one afternoon.)
  • $1,000 in free inference (Special offer for How I AI listeners to start building with HyperAgent.)

RevBots.ai View:

  • AI Sprinkler stage teams bolt on AI for tasks like email triage without full workflow redesign.
  • ARM stage orgs would use intent engineering to rebuild entire revenue workflows from first principles.
  • SaaS Hoarders collect point solutions for each task instead of building unified AI agents.
  • Tab Hoppers lack the technical literacy to even begin this level of AI integration.
🎧Full Episode:Lenny's Podcast →