Loading...

Explain Daily Route Changes

When a delivery plan changes before dispatch, the last-mile route planner has to explain what moved, why it changed, and what each driver should do next. This workflow compiles route changes, attributes likely operational causes, and produces a driver-ready narrative for a specific plan and date.

Planner judgment stays intact; the workflow removes the execution work of collecting changes, matching causes, and formatting instructions. The estimated impact is $27,300 per year in hard savings from compressing this work by 50%.

Interactive Demo

See the Decision Intelligence Workflow in Action

This workflow handles a compound communication task that planners usually do manually under time pressure. It identifies which stops changed from the original plan, links those changes to observable operating factors, and assembles a plain-language explanation that a driver can act on immediately. The result is a repeatable narrative instead of one-off calls, texts, and ad hoc notes.

👤
Designed for line-of-business leaders — specifically the Last-Mile Route Planner who is accountable for dispatch clarity, driver compliance, and exception follow-through. It addresses the repetitive explanation work that sits between route changes and clean field execution. Why this matters: What line-of-business leaders actually need from AI →
Decision Intelligence Playground · Driver Narratives for Route Changes
KWF Runtime
Start with an example
Want to see this workflow on your route planning data?
In a working session, we map the route plan and stop-factor data needed for the narrative, walk through the logic that identifies changes and likely causes, and show the output your planners and drivers would receive. We'll show you this workflow running on your Databricks environment. The session is focused on data fit, operating logic, and where the time recovery sits for your team.
Schedule a Databricks Session
How It Works

What the Decision Intelligence Workflow Does

This workflow handles a compound communication task that planners usually do manually under time pressure. It identifies which stops changed from the original plan, links those changes to observable operating factors, and assembles a plain-language explanation that a driver can act on immediately. The result is a repeatable narrative instead of one-off calls, texts, and ad hoc notes.

Click any step below to see the business logic, data query, and sample output for that step of the workflow.

Pre-specified logic, not runtime guessing — Most AI agent frameworks work by figuring things out on the fly. These Decision Intelligence workflows work differently. The Knowledge Work Foundry analyzes the cognitive labor pattern before deployment and encodes the decision logic directly into the configuration — which tables to query, which thresholds define a breach, how signals are ranked, and what the output artifact should contain. That analysis happens once. By the time the workflow runs, there is nothing left to figure out.
1
Detect Changes
Compares the original and adjusted route plan to identify which stops changed and how they changed.
2
Rank Causes
Looks up operational signals for the changed stops and ranks the factors most likely driving the adjustments.
↑ click a step to explore the logic, query, and output
1
Step Detail

                      

                    

The workflow produces a concise driver-facing route adjustment summary for a given plan and date. That output can be delivered in the driver mobile app and surfaced as an annotated update in the daily route planning dashboard.

What this workflow does NOT do: It does not decide whether a route should be redesigned for the day, choose which driver or carrier partner should absorb extra workload, determine whether to escalate repeated customer unavailability to account management, approve staffing or dispatch changes, or set the operating policy for how much delay is acceptable before a stop is reassigned.
Under the Hood

Data Warehouse Integration

The workflow depends on joining route plan data with stop-level operating signals for the same service date. It retrieves planned versus adjusted sequence and assignment changes, then connects those changes to measurable factors so the narrative is traceable to actual operating data.

That means the explanation is built across route planning data and stop-level factor data, not from a single operational table.

Source system: Daily route planning and dispatch output  ·  Domain: Last-mile Operations
Role in this workflow: This table is the core operational record for comparing the original route plan to the adjusted plan at the stop level. It provides the stop sequence, driver assignment, and execution status needed to identify whether a stop was reordered, reassigned, delayed, or left unchanged.
CREATE TABLE retail_dev.last_mile_ops.route_plans (
  plan_id VARCHAR,
  stop_id VARCHAR,
  plan_date DATE,
  driver_id VARCHAR,
  planned_sequence INTEGER,
  adjusted_sequence INTEGER,
  status VARCHAR
);

SELECT
  plan_id,
  plan_date,
  stop_id,
  driver_id,
  planned_sequence,
  adjusted_sequence,
  status,
  CASE
    WHEN planned_sequence <> adjusted_sequence THEN 1
    WHEN status <> string">'On Time' THEN 1
    ELSE 0
  END AS is_changed
FROM retail_dev.last_mile_ops.route_plans
WHERE plan_id = string">'{plan_id}'
  AND plan_date = string">'{plan_date}';
Source system: Operational signal feeds for traffic and customer availability  ·  Domain: Last-mile Operations
Role in this workflow: This table adds the contextual signals used to explain why a stop changed after the plan was adjusted. It contributes causal evidence such as elevated traffic conditions or customer unavailability, which the workflow uses to produce a ranked explanation rather than a raw list of route edits.
CREATE TABLE retail_dev.last_mile_ops.stop_factors (
  stop_id VARCHAR,
  factor_date DATE,
  traffic_index INTEGER,
  customer_available BOOLEAN
);

SELECT
  stop_id,
  factor_date AS plan_date,
  traffic_index,
  customer_available
FROM retail_dev.last_mile_ops.stop_factors
WHERE factor_date = string">'{plan_date}'
  AND stop_id IN (
    SELECT stop_id
    FROM retail_dev.last_mile_ops.route_plans
    WHERE plan_id = string">'{plan_id}'
      AND plan_date = string">'{plan_date}'
  );
Cognitive Labor Analysis

Where the Work Sits in the Labor Stack

Not all cognitive labor is equally automatable. The KWF analysis breaks the workflow into three layers — execution, judgment, and strategic — and maps each step to the layer it belongs to. Execution-layer work is automatable. Judgment and strategic work stays with the manager.

Execution 50% Judgment 35% Strategic 15%
Execution Layer
50%
The execution layer retrieves changed stops, evaluates route differences, attributes likely causes such as congestion or customer unavailability, ranks factors, and formats the driver-facing message.
Judgment Layer
35%
The planner still interprets whether the narrative fits field conditions, decides whether added outreach is needed, and handles exceptions with drivers or carrier partners.
Strategic Layer
15%
Leadership still owns broader decisions about dispatch policy, carrier management, staffing structure, and route planning standards.
Value Model

The Business Case for Automation

Time Recovered
30 min / day
One planner is assumed to spend 30 minutes each day explaining route changes for five drivers, six days per week.
Annual Savings
$27,300 / yr
This reflects the modeled hard savings from compressing that explanation work by 50% at a blended rate of $70 per hour.
Strategic Upside
Fewer Support Escalations
Consistent narratives reduce confusion, missed stops, and avoidable back-and-forth between drivers and planners.
Kill Question: Without this workflow, drivers understand daily plan changes through manual calls, ad-hoc texts, and inconsistent planner explanations.

Primary Valuation Metric: Reduction in average time spent by planners on daily route plan explanations per driver.

Next Step

Run your own value estimate for last-mile route planning — or talk to us about this driver communication workflow.

The Cognitive Labor Value Calculator models team size, role cost, and automation coverage to estimate the time and cost tied to manual route-change explanations. It takes under two minutes to complete.