Skip to content
Graph of Plan

Market timing: Is the market ready for agent-native?

We’re betting on agent-native applications. But the market for autonomous AI agents may develop slower than we expect.

The Risk

Our thesis is that agent-native applications—software built around AI agents as first-class actors—will dominate how work gets done. But:

  • Most businesses are still figuring out basic AI (chatbots, copilots)
  • Enterprise adoption of autonomous agents is nascent
  • Trust in AI decision-making is still building
  • The “agent” concept is unclear to mainstream buyers

If the market isn’t ready, we’re building products people don’t want to buy yet.

Specific Threats

  1. Adoption lag: Enterprises take longer than expected to adopt autonomous agents
  2. Trust deficit: High-profile AI failures create backlash against autonomy
  3. Copilot dominance: “Human-in-the-loop” copilots win over autonomous agents
  4. Infrastructure oversupply: Everyone builds agent infrastructure, no one builds applications
  5. Buyer confusion: “Agent” becomes meaningless buzzword, buyers can’t evaluate

Mitigations

Product Positioning

  • Problem-first, not agent-first: Murphy is “delivery control,” not “delivery agent.” P4gent is “purchasing assistant,” not “procurement agent.”
  • Value, not technology: Sell outcomes (saved time, caught issues, reduced costs), not AI capabilities
  • Graduated autonomy: Start with suggestions and alerts, let users opt into more autonomy as trust builds

Market Strategy

  • Early adopter focus: Agencies and developers are ahead of mainstream; start there
  • Proof points: Every implementation becomes a case study that builds market confidence
  • Education: Content marketing that explains what agent-native means and why it matters

Business Model

  • Multiple products: If one market segment isn’t ready, others may be
  • Platform economics: Even if Murphy is early, SmartBoxes serves developers building their own agents
  • Flexible roadmap: Can pivot to copilot-style features if full autonomy adoption lags

Residual Risk

Timing is the hardest thing to get right. We’re betting the market is ready based on signals from AI infrastructure adoption, enterprise AI budgets, and the pace of agent framework development. If we’re wrong by a year or two, we may run out of runway before the market catches up.

Probability: Medium (signals are positive but not definitive) Impact: High (being early can be as bad as being wrong) Mitigation effectiveness: Moderate (hedges help, but can’t eliminate timing risk)