Skip to content

Lesson 2 of 5

Working with Customers: Scoping and Integrating

Understand how an FDE scopes problems with customers and integrates solutions into the customer's existing systems.

5 minForward Deployed Engineering

After this lesson you will be able to scope a customer problem into a clear statement of work, map the customer's existing systems, and plan an integration that respects their constraints.

Scoping a customer problem

Scoping means turning a vague request into a concrete piece of work. A customer may say: "We want to use your product to speed up our claims process." That is not enough to build against. You need to find out what the process is today, where it is slow, what data is involved, and what a good outcome looks like.

Start by asking the customer to walk through the current process. Listen for the nouns: the people, systems, files and rules. Then ask what they want to change. Write down the problem as a single sentence: "When a claim arrives, an assessor manually checks three systems and re-enters data, which takes about twenty minutes per claim." That sentence names the trigger, the work, and the cost.

Then define success. Ask: "If we solve this, what will be different in three months?" The customer might say: "Assessors will handle forty claims a day instead of twenty-five." Turn that into a measurable goal. You now have a problem statement and a success measure. That is the scope.

A common mistake is to accept the customer's proposed solution as the scope. The customer may say: "Build us a dashboard that shows all claims." A dashboard may help, but the real problem is the manual re-entry. Your job is to find the problem behind the request. Ask why they want the dashboard. What decision will it help them make? The answer often reveals the true scope.

Mapping the customer's systems

Before you can integrate, you need to know what you are integrating with. Draw a map of the customer's systems. Start with the systems that touch the problem. For each one, note: what it does, who uses it, how it stores data, and how it talks to other systems.

You can draw this map on paper or in a tool. The important thing is to ask the right questions. For each system, ask:

  • What is this system called, and what does it do?
  • Who owns it inside the customer's organisation?
  • What data does it hold, and in what format?
  • How does it share data: an API, a file export, a database, or manual entry?
  • What are its constraints: uptime, security, compliance, or cost?

A simple map might look like this: a claims system receives claims from customers, stores them in a database, and sends a nightly file to the finance system. The assessor opens the claims system, copies data into a spreadsheet, then enters it into the finance system. Your integration will likely sit between the claims system and the finance system.

Do not try to map every system the customer has. Focus on the ones in the scope. If the scope is the claims process, you do not need to know about the HR system. But you do need to know about anything that touches the claims data.

Planning the integration

Integration means connecting your solution to the customer's systems so that data flows without manual work. The plan has three parts: what you will build, where it will run, and how it will be maintained.

First, decide what to build. Often the smallest useful thing is a connector: a piece of code that reads from one system and writes to another. In the claims example, a connector could read new claims from the claims system's API, transform the data into the format the finance system expects, and send it to the finance system's API.

Second, decide where it will run. The customer may have rules about where code can run. Some customers want everything on their own servers. Others allow cloud services. Ask about security and compliance early. If the customer says: "No data can leave our network," then your connector must run inside their network.

Third, plan for maintenance. Who will fix the connector if it breaks? Who will update it when the customer's systems change? Write down the answers. A good integration plan includes a simple runbook: a document that says what the connector does, how to start it, how to stop it, and what to check if it fails.

Here is a small example of a connector plan. The customer has a claims system with a REST API that returns JSON. The finance system accepts CSV files. The connector will run on a server inside the customer's network. It will run every hour, read new claims from the API, write them to a CSV file, and upload the file to the finance system's SFTP server. The runbook will say: check the log file for errors, and if the finance system is down, the connector will retry three times before alerting.

Writing a statement of work

A statement of work is a short document that captures the scope, the systems, and the integration plan. It does not need to be long. One page is often enough. It should contain:

  • The problem statement, in one sentence.
  • The success measure, in one sentence.
  • The systems involved, with their owners.
  • What you will build, in plain words.
  • Where it will run, and any constraints.
  • What is out of scope, so there are no surprises later.

Share this document with the customer before you start building. Ask them to confirm that it matches their understanding. This step prevents the most common failure in FDE work: building the right thing for the wrong problem.

Scoping and integration are not one-time events. As you build, you will learn more about the customer's systems. When that happens, go back to the statement of work and update it. A living document keeps everyone aligned.

Key takeaways

  • Scoping means turning a vague request into a problem statement and a measurable success goal.
  • Map only the customer systems that touch the scoped problem, and note their owners, data formats and constraints.
  • An integration plan says what you will build, where it will run, and how it will be maintained.
  • Write a one-page statement of work and confirm it with the customer before building.
  • Update the statement of work as you learn more during the build.

Try it

Choose a simple process you know well, such as ordering food in a restaurant or checking out a book from a library. Write a one-page statement of work for a small software improvement to that process. Include a problem statement, a success measure, a list of the systems involved (even if they are paper or human), what you would build, where it would run, and what is out of scope. Then ask a friend to read it and tell you if they understand what you will do and why. A good result is a statement that someone else can read and agree with without asking many questions.

Official documentation

Ready to use this in a job?

Rolebound applies to matching roles for you.

Start now

Written with AI and checked automatically. Tools change, so a command may behave differently in the version you install. Spotted a mistake? Tell us.

    Cookies on Rolebound

    We use necessary cookies to sign you in and keep the site working — those cannot be turned off. We would also like to set analytics and advertising cookies so we can see which pages help people get interviews and which adverts brought them here. Nothing optional is set unless you say yes, and you can change your mind at any time from “Cookie settings” in the footer. Read our Privacy Policy.