LAST MILE Where agents actually land

Remediation playbooks

Fix the signals that stop agent journeys.

The Index shows where the route breaks. These playbooks connect each measured failure to the least expensive credible fix, then show what completion should look like.

Playbooks
8
Signals covered
6
Remediation tiers
3
Evidence
versioned, cited
Input affordances Controls accept agent input. Failure transparency Errors explain the recovery. Consent wall Blocking gate measured.
Tier 0Remove the blocker
Tier 1Publish the facts
Tier 2Offer the tool
Verified route restored Agent journey completed.

Find a playbook.

Filter the repair ledger by the signal, maturity level, or operating status you need.

Showing 8 playbooks.

01 · Input affordances

Fix inputs so agents can start and finish tasks.

Forms, date pickers, and controls must accept the same values an agent can actually provide.

Open the Index →
Tier 0 · Remove the blocker Production

Replace hostile date widgets with agent-readable inputs

Remove readonly/hidden date fields and custom datepicker overlays that reject programmatic input; use a standard <input type="date"> or publish the fact as readable text so an agent never needs to fight the widget.

Before Hostile date widget

Overlay rejects programmatic input

readonly + hidden input
After Agent-readable input

A native control accepts a deterministic value

<input type="date">
Expected result
Agents can enter dates without fighting the widget.
Maturity lift
Understanding

02 · Failure transparency

Turn silent dead ends into recoverable errors.

When an action fails, the page must say what happened and give the agent a usable next step.

Open the Index →
Tier 0 · Remove the blocker Production

Surface real validation errors instead of silent dead-ends

When a form submission fails, return a visible, machine-readable error message — role="alert" or aria-invalid — so the agent knows the attempt failed and why.

Before Silent dead end

Submit fails without a usable explanation

submit → no response
After Visible validation error

The field and correction are surfaced together

role="alert"
Expected result
Agents know what failed and how to recover.
Maturity lift
Understanding

03 · Consent wall

Keep consent from becoming the journey-ending gate.

Consent controls should remain accessible without intercepting focus or blocking the core task.

Open the Index →

04 · Discoverability

Make the right entry point possible to find.

Semantic navigation and capability indexes reduce the guesswork before an agent begins the task.

Open the Index →
Tier 1 · Publish the facts Production

Add semantic markup and structured navigation to expose entry points

Mark up navigation landmarks with ARIA roles, use descriptive link text matching common task keywords, and add structured data (JSON-LD) so agents identify the right entry point without guessing.

Before Ambiguous entry points

Generic links force agents to crawl and guess

click here · learn more
After Structured navigation

Landmarks and descriptive links expose the task

<nav aria-label="Primary">
Expected result
Agents reach the right capability with less search.
Maturity lift
Discoverability
Tier 1 · Publish the facts Emerging

Publish an llms.txt capability index

Serve /llms.txt describing the site's key pages and capabilities in Markdown so agents can orient without crawling.

Before Crawl to orient

Capabilities are scattered across the site

crawl → infer
After Capability index

Key pages and functions are named in one place

/llms.txt
Expected result
Agents can orient before entering the interface.
Maturity lift
Discoverability

05 · Structured path

Offer a supported path beyond blind DOM driving.

Typed tools give agents a stable, machine-usable route through high-value tasks.

Open the Index →
Tier 2 · Offer the tool Emerging

Expose WebMCP tools for true agent workflows

Implement window.__webmcp (or a link[rel='mcp'] manifest) exposing typed tool definitions for core tasks (search, booking) so agents invoke structured operations instead of driving the DOM blindly.

Before Blind DOM driving

The agent infers controls from presentation

inspect → guess → retry
After Typed agent tool

A declared operation returns structured output

tool(searchFlights)
Expected result
Agents invoke a supported path instead of guessing.
Maturity lift
Interaction

06 · Task completion

Put the answer where an agent can verify it.

Readable facts and deterministic outcomes keep the journey measurable from request to result.

Open the Index →
Tier 0 · Remove the blocker Production

Publish core info as agent-readable content

Ensure the answer to the task is available as readable static content (a text table, a paragraph) so an agent can extract it without navigating a transactional flow.

Before Answer behind interaction

The fact exists only after a brittle flow

answer → hidden state
After Readable fact

The answer is present as stable page content

text · table · paragraph
Expected result
The deterministic check can verify the answer.
Maturity lift
Understanding

Future frontier

Transaction comes after the route is clean.

Payment protocols are gaining traction, but they do not compensate for blocked inputs, opaque failures, or missing facts. Treat this as a horizon, not a baseline.

Tier 2 · Offer the tool Emerging

Support agentic payment rails (x402 · MPP · UCP · ACP)

Adopt an emerging agent-payment protocol so an agent can complete a purchase under delegated authority. Frontier — earns its place only after Tier 0/1 are clean.

BeforeAuthority stops at checkout

No delegated payment rail can complete purchase

purchase → handoff
FutureDelegated transaction path

An emerging protocol carries scoped authority

UCP · ACP · x402
Expected result
A future agent can complete the paid journey safely.
Maturity lift
Transaction
Find the break See which airline signals fail in verified agent runs. Open the Index → Verify the method See how signals, access gates, and outcomes are scored. Read the methodology →