Turn an AI idea into a decision you can act on

Preparing customer data for routing

Slow data preparation can cost a business customers and revenue. When prospects have to wait to see their own data working in a routing and scheduling solution, they can lose confidence and move on. For customers who've committed, delays hold up daily use and, where payment depends on going live, the start of revenue. A difficult onboarding experience also puts satisfaction and retention at risk.

The implementation team feels the pressure too. Each customer system describes and organizes vehicle and order information differently. Repeated custom work consumes time the team needs to get the next customer operating. As the customer base grows, that work can become a limit on growth.

This example shows how I'd use the canvas to address the problem. AI helps draft reusable instructions for converting customer data into the required format. People confirm the meaning and approve the instructions. Rules check the prepared data. The business aim is to help prospects reach a decision, get customers operating and paying sooner, and make onboarding more efficient.

About this public example: Some details have been generalized or omitted. It draws on earlier work and includes a proposed pilot; the business benefits described are aims to test.

Experience behind the example

In one customer implementation, I observed that the main data-preparation stage needed almost no ongoing manual setup once reusable conversion instructions were in place. The capability was released. That experience gives us a practical starting point for testing the approach with another customer.

Work through the canvas in this order: understand the problem → choose the approach → design the test → make the decision. Each section helps answer the next. Section 8 brings the reasoning together in a short decision summary.

1 Define the outcome

Workflow and people. Prepare customer vehicle and order data for repeated imports into a routing and scheduling solution. The implementation team coordinates the preparation. Sales needs prospects to see the solution working with their data, and customer operations teams need reliable information to start using it.

Outcome and reason to act. Get usable customer data into the solution sooner and reduce the custom work each source requires. Delays can stall a prospect's decision, postpone customer use and payment, and weaken confidence during onboarding. We need a preparation process that keeps pace with new customers and changes to their systems.

Starting evidence. I've observed preparation effort ranging from hours to weeks, depending on the source and the changes required. For the selected customer, we'll measure preparation and review time, how long data readiness holds up onboarding, and any effect on the planned start of operations or billing.

Who is responsible. Product leadership defined the capability and accuracy requirement in the earlier work. For this pilot, we'll name an owner who can coordinate implementation, sales and customer operations and take responsibility for the business result.

Before you move on: We've linked the preparation problem to customer progress and revenue. The pilot needs to show whether removing this delay helps customers reach the next milestone sooner.

2 Map the current work

This section describes the work before adding AI.

Start and finish. Work starts when a customer introduces a data source, repeats an import or changes a source format. It ends when vehicle and order information is ready for the routing system to use. A delay at this stage can hold up a demonstration, customer launch or daily operation.

Information and trusted sources. Customer applications, exports and spreadsheets supply the information. The customer confirms what each field means. The routing system defines the required format through its software interface, or API. We'll check both the format and the meaning of the prepared data.

Workload and consequences. Imports can recur in daily operations. We'll confirm the selected customer's frequency and volume. Incorrect preparation can block orders or cause operating errors. Repeated delays also consume customer time and reduce confidence in the solution.

Before you move on: Understanding and converting source formats is a likely bottleneck. We need to see where the team spends time and where the customer is left waiting.

3 Find what holds the work back

The main bottleneck. Each source needs its own field matching and conversion instructions before the data is ready. I've seen this custom work recur with prospects and paying customers, slowing progress at both stages of the relationship.

Why it happens. Customers use different structures, labels and conventions for similar information. Experienced people understand the meaning and the changes needed. Turning that knowledge into reusable instructions takes time, and source changes can send the work back through the same cycle.

Would solving it make a difference. Faster preparation could help prospects evaluate the solution sooner and customers start operating earlier. It could also free the implementation team to onboard more customers. We'll track the next customer milestone to see how much the improvement changes the overall timeline.

Simplest workable option to compare. A standard import template, a field-matching tool or an existing connection between systems may be enough for stable sources. We'll assess these options against the selected customer's needs, including how quickly each can be put into use and how much work it takes to maintain.

Problem statement: When a customer introduces or changes a data source, repeated interpretation and conversion work delays usable routing data. This slows customer progress, absorbs implementation capacity and puts sales momentum and customer confidence at risk.

Before you move on: We have a business reason to act. Compare how well each approach reduces the delay and repeated work while keeping the data reliable.

4 Decide where AI belongs

Compare the approaches. AI can turn source explanations and preparation instructions into a draft set of reusable conversion instructions. These instructions say where each source field belongs and how its values must change to fit the routing system. The opportunity is to shorten the custom preparation work that keeps customers waiting.

We'll use AI to create or update the instructions, then reuse the approved preparation method for later imports.

Limits and chosen approach. We'll focus AI on drafting conversion instructions. Source records and routing decisions stay with the existing systems and responsible people. A reviewer must approve changes before use. We'll confirm the tool permissions and test whether the checks catch errors that could disrupt customer operations.

Compare the full setup, review and recurring effort with the simplest workable alternative. The chosen approach needs to save enough time to improve customer progress and team capacity.

Before you move on: We've defined AI's role and the checks needed to protect the customer. We can now plan a small test of its business value.

5 Define the pilot scope

The owner must agree to the plan before the pilot starts.

Pilot question and scope. For one customer implementation, can AI help create accurate, reusable conversion instructions quickly enough to reduce preparation effort and bring forward data readiness? We'll test one source format with representative imports. Other customer sources, routing decisions and unreviewed changes to live operations are outside the pilot.

Approved information and access. We'll use customer-approved samples and preparation instructions. Before starting, we'll confirm the tool, permission to share information with it, who needs access, where the information will go and how long it may be kept.

Checks and responsible reviewer. We'll check required fields and formats, then compare converted values with reviewed examples to confirm their meaning. A named reviewer will approve the instructions and resolve uncertain cases. We'll agree on the checklist and when results can be used so the customer can proceed with confidence.

When things go wrong. We'll set aside affected imports when checks fail or meaning is unclear. The reviewer will resolve the issue, with the existing preparation process available to keep work moving. We'll name who can pause the pilot and keep the source versions, instructions, check results and approvals needed to investigate problems.

Test cases and resources. We'll include typical imports, missing fields, unclear values, difficult conversions and a changed source format. The owner will agree on the number of cases, duration, time or cost limit and reviewer availability. We'll fit the test around the customer's next milestone so it supports timely onboarding.

Before you move on: The test has a clear task and customer purpose. Confirm approvals, resources and limits promptly so the team can start once the requirements are met.

6 Decide how to measure success

We'll measure both preparation performance and customer progress. Time saved matters when it helps customers move forward or gives the implementation team capacity for more work.

Include tool and implementation costs in the decision. Compare the current and proposed approaches using the same source and similar cases, including difficult ones. Ask the customer whether preparation is easier and whether they're able to move forward as expected.

Proposed decision rules

• Use more widely when results meet the agreed quality, total effort falls enough to justify the approach, and the improvement helps customer progress or team capacity. The owner must confirm that safeguards work.
• Improve and retest when conversion issues, repeated problems or review effort reduce the benefit and there's a practical fix worth testing.
• Stop when permissions are unavailable, serious errors escape the checks, problems disrupt customer operations, or a simpler approach meets the need more effectively.

Enough evidence to decide. We'll agree on the cases, targets, people collecting evidence and review date before starting. The pilot will give us early evidence about effort, quality and customer progress. As the approach is used more widely, we'll track its contribution to customer satisfaction, retention and revenue.

Before you move on: We've defined the results needed to justify the next decision and connected them to the customer and business outcomes that matter.

7 Check whether you are ready

For a new customer pilot, we'll confirm the following before starting.

Readiness decision: Revise the plan and resolve the open items before approving the pilot. Customer delays give us a reason to move promptly through these checks.

Open questions: First, name the person responsible for the business result and decision to proceed. That person will assign the permission and source checks, confirm the review process and agree on measures of success. Each action needs an owner and a date tied to the customer's next milestone.

Before you move on: The next step is to close the readiness gaps and make a timely decision on the pilot.

8 Write the decision summary

Decision and reason. Prioritize this problem and resolve the remaining requirements for a customer pilot. Preparation delays put prospect conversion, customer satisfaction and revenue at risk. We need a faster, reliable onboarding process that supports growth.

Problem and intended improvement. Repeated custom preparation keeps customers waiting and consumes implementation capacity. We want prospects to reach a decision sooner, customers to start operating and paying earlier, and a smoother experience that helps prevent churn.

Approach and smallest useful test. Test one approved customer source. AI drafts reusable conversion instructions; people confirm meaning and approve use; rules check the results. Include typical and difficult imports within an agreed time limit.

Main risk and safeguard. Incorrect data can disrupt operations and damage trust. Require checks and review before use, with the existing preparation process available when problems arise.

Evidence and open questions. We'll look for reliable results, less total effort and faster customer progress. Confirm permissions, targets and reviewers before starting. Track satisfaction, retention and revenue as use expands.

Who is responsible and what happens next. Name the pilot owner, assign the remaining checks and set a decision date around the customer's next milestone. Start when section 7 is complete.

---

The canvas turns an operating problem into a decision and a practical next step. The summary makes the business case clear; the earlier sections show the work, checks and evidence needed to act on it.

Your experience and close understanding of the problem may already lead you to a sound decision. The canvas adds a systematic way to cover the relevant aspects, check your assumptions and weigh the evidence. This helps you stay objective and improve your chances of success.

If you'd value help working through an operating problem or planning a test, contact me.

Jose Santos | Operating Problem and AI Pilot Canvas | September 2026