Complex Tasks

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

Handle big jobs in small steps: outline first, then one section at a time; messy notes into an action table with TBD for gaps; one step per message; and role-play to practise hard conversations.

A Big Task in One Line

Arun, the project coordinator, needs a guide for new suppliers. It should explain how to set up an account, book a delivery, and send an invoice. He asks for the whole thing in one line.

Eight generic sections, blanks everywhere, and nothing about how Kestrel actually works. The bigger the task, the vaguer a one-line answer gets.

In simple words: Break a big task into small steps and check each step before moving to the next one.

TL;DR

• For long documents, ask for an outline first. Check it, then generate one section at a time.
• For messy notes, ask for a table with named columns and tell the AI to write TBD when information is missing.
• For multi-part jobs, work one step at a time: understand first, then act.
• For difficult conversations, use AI as a role-play partner.
• Check each piece before moving on.

A Real-World Example

Imagine asking a new colleague to "organise the team offsite." A week later, they come back with the wrong venue, the wrong date, and a budget twice what you have available.

Now imagine you say, "First, give me three venue ideas. Once I choose one, draft the agenda." Each step is small, and you can correct the direction before any real work is wasted.

Big tasks with AI work exactly the same way. Small steps, checked as you go.

What You Will Learn

In this lesson, you will learn:

1. How to build a long document from an outline.
2. How to turn messy notes into an action table without invented owners.
3. How to split a job into manageable steps.
4. How to practise a difficult conversation with AI.

Outline First

Why an Outline Helps

An outline is a list of headings that defines the structure of a document before any text is written.

It takes seconds to review and seconds to fix.

If the outline is wrong, you change a few lines. If a ten-page document is wrong, you have to rework everything.

Build It Section by Section

Change Arun's message so it asks for an outline only. Once you approve the outline, ask for one section by number, for example:

"Looks good. Write section 3."

You checked the structure before any content was written. Each section came back short enough to review properly, and you can request the next section when you are ready.

Notes Into an Action Table

Name the Columns

Meeting notes are often messy. Actions, owners, dates, and decisions are mixed together.

A table makes the information easier to review by putting each action into its own row.

As you learned earlier, always name the columns you want.

For this task, use:

• Action
• Owner
• Due Date
• Status

Don't Let It Invent Owners

Some actions have no owner yet. Others do not have a due date.

If you do not tell AI what to do when information is missing, it may invent names and dates that look realistic.

That can create confusion because people may believe those assignments were actually agreed.

Use a simple rule:

If the notes do not provide an owner or a due date, write TBD.

TBD means "To Be Decided."

Arun's review-meeting notes are below.

Turn them into a table and make sure missing information appears as TBD.

Without that rule, the AI assigned the venue task to Priya and created a due date. With the rule, the gaps remain visible so the team can assign ownership properly.

One Step at a Time

Understand First, Then Act

Some jobs require two different skills:

1. Understanding information.
2. Doing something with that information.

If the AI misunderstands step one, everything built on top of it may be wrong.

Instead of combining both tasks, separate them.

First ask the AI to understand the information.

Then review the result.

Only after that should you ask it to act.

The box below contains three customer comments.

Start with:

"Step 1: List the customer's concerns."

Then follow up with:

"Step 2: Draft a reply to comment 3."

You can now confirm that the AI understood the comments before trusting it to write a response.

Practise a Hard Conversation

The AI as a Practice Partner

AI is not just useful for writing. It can also play a role in a conversation so you can practise before speaking with a customer, colleague, candidate, or manager.

Some useful workplace examples include:

• A customer challenging your pricing
• An upset customer after a service issue
• A job interview
• A difficult discussion with a manager
• A supplier negotiation

What Good Looks Like

In a challenging conversation, focus on three things:

1. Ask an open question.

   • Questions beginning with what, how, or why encourage fuller answers.

2. Reflect the concern back.

   • Example: "So your main concern is the delivery delay."

3. Agree on a next step.

   • A meeting, a phone call, or an action item.

And for Maria:

No discounts.

Try It With Daniel

The AI will play Daniel, the customer from earlier lessons.

You play Maria.

Send the setup message, then reply to Daniel at least three times.

When you are finished, type:

"end role-play"

to receive feedback.

You can repeat the exercise as many times as you want. Each conversation develops based on what you say.

Common mistakes

Exercise

Complete each task below and then press Check.

What to Remember

• Big task? Outline first, then one section at a time.
• Meeting notes become clearer in a table with named columns and a TBD rule.
• Understand first, then act.
• Use AI as a practice partner for difficult conversations.
• Review each piece before moving on.

Key Takeaway

Large tasks become easier when you break them into small, reviewable steps.

Instead of asking AI to do everything at once, guide it through the process. Review the outline, verify the facts, check each section, and move forward one step at a time.

Next: Prompt Templates