Giving Context

By the Syncfusion DevTutorials team · Last updated October 20, 2026

Give the AI the background it needs: who, what, when, where and why. Name the reader, mark off pasted text, and let the AI interview you so nothing is guessed.

Arun's Vendor Visit

A paper supplier is visiting Kestrel next month. Arun, the project coordinator, asks the AI for a plan using just four words.

The plan looks professional at first glance, but read it carefully.

• It includes a factory tour, even though Kestrel does not have a factory.
• It includes a contract negotiation, even though that is not the purpose of the visit.
• It assumes a full-day schedule from 9 a.m. to 4 p.m.
• The supplier's name and date are missing.

The AI was not trying to be wrong. It simply did not know enough about the situation, so it filled the gaps with assumptions.

In simple words: Context is the C in CRAFT. Without it, AI does not know your situation, so it creates one of its own.

TL;DR

• Context is everything AI cannot know unless you tell it.
• Good context answers who, what, when, where, and why.
• The why is often the most important part because it tells AI what matters most.
• Better context leads to better answers because the AI has fewer gaps to fill.
• Tell AI who the audience is. Customers, managers, and coworkers usually need different messages.
• When pasting text, clearly separate your instructions from the content.
• If you are unsure what information to provide, ask AI to interview you first.

A Real-World Example

Imagine you hand a new colleague a box of files and say, "Sort these."

They might sort them by date, customer, project, or department because they are guessing what "sorted" means in this situation.

Now imagine you say:

"These are last year's supplier invoices. Finance needs them grouped by supplier before Friday's audit."

Suddenly the task is clear. The second instruction is not better because it is longer. It is better because it contains the information needed to do the job correctly.

AI works the same way.

What You Will Learn

In this lesson, you will learn:

1. The five questions that create useful context.
2. Why the goal is often the most important piece of context.
3. How the audience changes the answer.
4. How to paste content so AI treats it as source material rather than part of your instruction.
5. How to get AI to ask questions before it begins.

The Five Context Questions

Who, What, When, Where, Why

Before sending a prompt, ask yourself these five questions:

• Who is involved?
• What is happening?
• When is it happening?
• Where is it happening?
• Why does it matter? What is the goal?

You do not need a long explanation for each one. Usually a short sentence or even a short phrase is enough.

Give It the Background

Here is Arun's prompt again, but this time it includes answers to all five questions.

Notice what changed.

The goal was to review late deliveries, so the AI focused the visit plan around delivery performance rather than factory tours or contract discussions. The AI used the goal to decide what mattered most.

In simple words: Facts tell the AI what to include. The goal tells it what to focus on.

Name the Reader

Same Information, Different Audience

The same information often needs very different wording depending on who will read it.

• Customers want to know how something affects them.
• Employees want to know what actions they need to take.
• Managers often want the key points and next steps.

The box below describes a new order cut-off time.

Send the prompt as it is.

Then start a new chat and specify that the message is for customers.

Start another new chat and make it for the warehouse team.

The information stayed the same. Only the audience changed, yet the wording, structure, and focus changed dramatically.

That is why naming the reader is an important part of context.

Paste Text the Right Way

Mark It Off

Many workplace tasks involve pasting content such as:

• Emails
• Meeting notes
• Policies
• Reports
• Customer messages

AI needs to know which text is the instruction and which text is the material you want it to work on.

A simple approach is:

1. Write your instruction first.
2. Add a label such as Email:.
3. Put the content between """ marks.

For example:

Summarise this email in two sentences.

Email:
"""
[pasted email]
"""

Try it below.

When the content is clearly separated from your instructions, AI is much more likely to perform the correct task. Without those boundaries, it may think the email is a message that needs a reply rather than something that needs summarising.

Let the AI Interview You

The "Interview Me" Pattern

Sometimes you know what you want but you do not know which details the AI needs.

In those situations, let AI gather the information first.

Add this sentence to the end of your prompt, before you send it:

"Before you begin, ask me any questions you need answered."

Instead of immediately producing an answer, the AI will usually ask for the missing information. You answer the questions, and then it creates the final result. This reduces guessing and improves accuracy.

Try It With Arun's Visit

Start with Arun's original four-word prompt.

Then add:

"Before you start, ask me questions about anything you need to know."

Notice what happens. The AI asks for the details instead of inventing them, and once those questions are answered, the resulting plan becomes much more useful.

Common mistakes

Exercise

Complete each task below and then press Check.

What to Remember

• Context answers who, what, when, where, and why.
• The goal is often the most important piece of context.
• Better context leads to better answers because the AI has fewer gaps to fill.
• Different audiences need different messages.
• Put your instruction first and clearly separate pasted content.
• If you are unsure which information matters, ask the AI to interview you first.

Key Takeaway

Good context reduces guessing.

The more clearly you explain the situation, the goal, and the audience, the less the AI has to assume and the more useful the answer becomes.

Next: Output Formatting