Key Takeaways

  • Developer Philip Jackson deployed the autonomous agent Muse to serve legal notices, compelling his opponent to offer a settlement within 20 minutes.
  • DoorDash reported on its earnings call that personal AI agents have generated only several hundred orders across its entire platform history.
  • DoorDash built official developer APIs and command-line interfaces to avoid resisting automated traffic, mirroring infrastructure built by Walmart and Shopify.
  • Autonomous agents struggle when replacing ingrained, zero-friction consumer habits like tapping an app to order dinner.

High Friction Wins While Food Delivery Stalls

Philip Jackson wanted to settle a dispute. Instead of spending thousands of dollars on billable legal hours, he set an autonomous agent named Muse loose on the problem. As John Coogan highlighted, Jackson reported: “I hired Muse to sue someone on my behalf. I won. Well, they caved. Within 20 minutes of being served papers, the company offered a settlement.”

Serving formal legal documents is miserable. It requires procedural precision, specific formatting, emotional stamina, and legal research. An AI agent does that work without flinching. The target recognized the exposure, did the math on legal fees, and wired the money.

Contrast that with consumer commerce. On its recent earnings call, DoorDash disclosed that personal agents like Muse have placed only several hundred orders across its entire network. People love the idea of delegating their daily routines to autonomous software, yet nobody actually uses bots to buy tacos.

Jordi Hays pointed out why: “I just think it's so incredibly easy to order food. And people are very habitual. It's possible that you could be a power user of Muse and not even care to use it to order Door Dash.”

When a user wants dinner, opening an app and tapping a reorder button takes four seconds. Delegating that order requires typing prompts, verifying addresses, setting delivery windows, and auditing payment details. The bot adds cognitive drag to a problem that smartphone interfaces solved a decade ago. Coogan put the problem bluntly: “Tell the robot to push the Door Dash the get burrito button is not a great use case.”

Companies Prepare for Bots Before Consumers Do

DoorDash is not ignoring autonomous agents just because consumers ignore them. The company built dedicated developer APIs and command-line interfaces specifically so scripts can query menus and place orders programmatically.

“My take on all things technology is you never want to fight gravity,” Coogan said. “That's why Door Dash has an API, a CLI. They're leaning in like a Walmart, like a Shopify. But at the same time, you do not want to be delusional.”

Platforms know that automated buyers will eventually arrive in volume. Fighting scrapers with rate limits and CAPTCHAs costs money and creates bad consumer experiences. Providing structured endpoints lets companies log the data, charge fees, and control distribution. DoorDash is preparing its pipes for machine traffic today, even while consumer volume remains practically zero.

If you build software agents, stop targeting habitual, single-click consumer habits. Build for legal notices, supplier disputes, insurance appeals, and billing corrections. Target tasks where human dread and friction run high.

What to Do With This

Audit your current product backlog this week and identify any feature meant to automate a low-friction user habit. If your tool replaces an action that already takes less than thirty seconds on a smartphone, kill it. Reallocate those engineering hours toward an unpleasant, paperwork-heavy workflow where your customer currently delays action for days out of frustration.