<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: 杨子豪</title>
    <description>The latest articles on DEV Community by 杨子豪 (@yzh317179958).</description>
    <link>https://dev.to/yzh317179958</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4253471%2F8ad4280e-ed3a-42d2-8193-eadc43fdfacc.jpg</url>
      <title>DEV Community: 杨子豪</title>
      <link>https://dev.to/yzh317179958</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/yzh317179958"/>
    <language>en</language>
    <item>
      <title>What Is an AI Agent? A Plain Answer With a Real Example</title>
      <dc:creator>杨子豪</dc:creator>
      <pubDate>Thu, 19 Nov 2026 09:27:48 +0000</pubDate>
      <link>https://dev.to/yzh317179958/what-is-an-ai-agent-a-plain-answer-with-a-real-example-2ka2</link>
      <guid>https://dev.to/yzh317179958/what-is-an-ai-agent-a-plain-answer-with-a-real-example-2ka2</guid>
      <description>&lt;p&gt;An AI agent is software that uses a language model to decide which steps to take, and which tools to use, in order to finish a task. You give it a goal, such as "answer this customer email". It works out what it needs to know, looks those things up, and produces a result. A text generator answers from what it was handed. An agent goes and gets what it needs.&lt;/p&gt;

&lt;p&gt;We run AI agents on live customer service for an online brand. The example below is what one of them does with a single email. The volumes are in our &lt;a href="https://ai.dingboai.com/en/case-study-ai-customer-service.html" rel="noopener noreferrer"&gt;case study&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  A working example
&lt;/h2&gt;

&lt;p&gt;A customer writes that their parcel is marked delivered and nothing has arrived. The agent:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;reads the email, plus a summary of the earlier messages in the same thread;&lt;/li&gt;
&lt;li&gt;finds the order from the sender's address and the order number in the text;&lt;/li&gt;
&lt;li&gt;asks the carrier's tracking service for the current status;&lt;/li&gt;
&lt;li&gt;searches the knowledge base for how the company handles this situation;&lt;/li&gt;
&lt;li&gt;looks up past mistakes recorded for this kind of case;&lt;/li&gt;
&lt;li&gt;writes the reply in the customer's language;&lt;/li&gt;
&lt;li&gt;decides whether this is a reply it may send alone or one a person must approve.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Nobody listed those steps for it in that order. A different email, such as a question about battery range, leads to different steps: no order lookup, one knowledge base search, an answer.&lt;/p&gt;

&lt;h2&gt;
  
  
  The parts of an AI agent
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;A language model.&lt;/strong&gt; It reads the input and decides the next step.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Tools.&lt;/strong&gt; Functions the model can call: look up an order, read tracking, search the knowledge base, create a ticket. Without tools, a model can only talk.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Instructions.&lt;/strong&gt; What the job is and what a good result looks like.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Context.&lt;/strong&gt; The facts of this case, prepared for the model: what the customer wants, what has been promised, what has been done.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Limits.&lt;/strong&gt; What the agent may do alone and what needs a person.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  AI agent, chatbot, workflow
&lt;/h2&gt;

&lt;p&gt;A chatbot is a conversation window. What sits behind it can be a script, a model, or an agent. A workflow is a set of steps drawn in advance: every case follows the same path. An agent chooses its path case by case. The three are compared in &lt;a href="https://ai.dingboai.com/en/ai-agent-vs-chatbot.html" rel="noopener noreferrer"&gt;AI agent vs chatbot&lt;/a&gt; and &lt;a href="https://ai.dingboai.com/en/ai-agent-vs-workflow.html" rel="noopener noreferrer"&gt;AI agent vs workflow&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Types of AI agents
&lt;/h2&gt;

&lt;p&gt;Textbooks sort agents by how they decide: simple reflex agents, model-based agents, goal-based agents, utility-based agents and learning agents. That list describes the inside.&lt;/p&gt;

&lt;p&gt;For a business, the useful split is by what the agent is allowed to do:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Answers.&lt;/strong&gt; It reads and replies. Nothing changes in any system.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Prepares.&lt;/strong&gt; It gathers the facts, proposes a solution and drafts the reply. A person approves.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Acts within limits.&lt;/strong&gt; It sends simple replies and creates tickets on its own, inside rules a person set.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;All three run side by side in the deployment we operate. In the 30 days to 19 November 2026, the chat agent handled 15,725 conversations with no message typed by a human agent. In email, 26,451 replies went out using an AI draft: the AI sent 3,331 on its own, agents sent 3,781 without changing a word, and agents edited 19,339 before sending.&lt;/p&gt;

&lt;h2&gt;
  
  
  What an AI agent is not
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;It is not a colleague with a memory.&lt;/strong&gt; It knows about yesterday only if the system hands it the facts.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;It is not always right.&lt;/strong&gt; A model can state things that are false. The cases we caught are in &lt;a href="https://ai.dingboai.com/en/ai-customer-service-hallucination.html" rel="noopener noreferrer"&gt;hallucination examples and fixes&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;It is not the answer to every job.&lt;/strong&gt; If the steps can be drawn in advance, a workflow is cheaper and behaves the same way every time.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  What makes an agent work on real cases
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Facts, not raw material.&lt;/strong&gt; A model handed a whole email thread or a whole order misses the point. On one email about spare parts, we passed only the order lines the email concerned. The input fell from 37,604 tokens to 4,722, and all four parts in the reply were then correct.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;One source for the rules.&lt;/strong&gt; How to speak to a customer is written once, and every part of the system refers to that one text.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;A person where money moves.&lt;/strong&gt; Refunds, compensation and replacements are approved by a person. The agent prepares them.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Measurement by what people change.&lt;/strong&gt; Every edit an agent's reviewer makes is a signal about what the agent got wrong.&lt;/p&gt;

&lt;h2&gt;
  
  
  Do you need one?
&lt;/h2&gt;

&lt;p&gt;An AI agent fits a job that repeats every day, in real volume, with a system behind it to look things up in, and where the steps differ from case to case. Customer service email is the plainest example. Nine more are in &lt;a href="https://ai.dingboai.com/en/ai-agent-use-cases.html" rel="noopener noreferrer"&gt;AI agent use cases&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Tell us the job in one line and the free plan describes, within two working days, what an agent would handle and what stays with your people.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published at &lt;a href="https://ai.dingboai.com/en/what-is-an-ai-agent.html" rel="noopener noreferrer"&gt;ai.dingboai.com&lt;/a&gt;. Dingbo AI builds and runs AI agents for customer service, paid per completed task.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>llm</category>
      <category>productivity</category>
    </item>
    <item>
      <title>How to Build an AI Agent That Survives Real Work</title>
      <dc:creator>杨子豪</dc:creator>
      <pubDate>Thu, 19 Nov 2026 09:27:12 +0000</pubDate>
      <link>https://dev.to/yzh317179958/how-to-build-an-ai-agent-that-survives-real-work-3gd7</link>
      <guid>https://dev.to/yzh317179958/how-to-build-an-ai-agent-that-survives-real-work-3gd7</guid>
      <description>&lt;p&gt;Building an AI agent that answers a demo question takes an afternoon: a model, a prompt, one tool. Building one that handles thousands of real cases a month is a different project, and most of the work is outside the model. These are the seven steps we follow, taken from running agents on live customer service. The volumes are in our &lt;a href="https://ai.dingboai.com/en/case-study-ai-customer-service.html" rel="noopener noreferrer"&gt;case study&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 1: pick one job and write down what "done" means
&lt;/h2&gt;

&lt;p&gt;Not "automate support". One job: "reply to emails asking where an order is". Then define a finished case: the customer has the latest tracking status and the delivery time the policy states. If you cannot say what done means, you cannot test the agent.&lt;/p&gt;

&lt;p&gt;Pick a job that repeats daily, has real volume, and has a system behind it to look things up in. How to choose is in &lt;a href="https://ai.dingboai.com/en/ai-agent-use-cases.html" rel="noopener noreferrer"&gt;AI agent use cases&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 2: connect tools, read-only first
&lt;/h2&gt;

&lt;p&gt;List what a person looks up to do this job: the order, the tracking page, the policy document. Each becomes a tool the model can call. Start with tools that only read. An agent that can read everything and change nothing cannot damage your data while you learn how it behaves.&lt;/p&gt;

&lt;p&gt;Design each tool to return a small, fixed answer. Fulfilment status should come back as one of a few fixed values. We once passed it as a sentence of description, and the code read every description as "shipped". Over three days, 25 address-change replies were held back, and 12 of those orders had not shipped.&lt;/p&gt;

&lt;p&gt;Separate "there is nothing" from "the lookup failed". An empty result and an error need different handling.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 3: prepare the facts before the model sees them
&lt;/h2&gt;

&lt;p&gt;This step decides the quality more than the choice of model. Do not hand the model the whole email thread, the whole order or the whole document. Turn them into the few facts this case needs: what the customer wants now, what was promised before, what has been done, what is missing.&lt;/p&gt;

&lt;p&gt;We learned this from two cases. In a thread of 25 messages, the model did not notice that the company had already promised a full refund, and contradicted it. On an email about spare parts, sending only the order lines the email concerned cut the input from 37,604 tokens to 4,722, and all four parts in the reply were then correct.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 4: write the rules once
&lt;/h2&gt;

&lt;p&gt;How the agent should think and speak is written in one text, and every part of the system refers to it. When the same rule exists in five places, the copies drift apart and the model follows whichever it read last. We found a reimbursement rule in five places that did not agree. Agents kept correcting the drafts until all five said the same thing.&lt;/p&gt;

&lt;p&gt;Write principles, not ready-made sentences. A model copies an example sentence word for word into every reply.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 5: set the approval lines
&lt;/h2&gt;

&lt;p&gt;Decide, action by action, what the agent does alone and what a person approves. Our starting lines: the agent answers questions of fact and published policy alone. Anything that costs money or sends goods is prepared by the agent and approved by a person. The detail is in &lt;a href="https://ai.dingboai.com/en/human-in-the-loop-ai-agents.html" rel="noopener noreferrer"&gt;who approves what&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Record every approved action from the version the person approved, not from the agent's first proposal.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 6: replay real cases before launch
&lt;/h2&gt;

&lt;p&gt;Take real past cases and run the agent on them without sending anything. Compare its output with what your people did. Three details matter:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Run each case twice. Model output varies between runs.&lt;/li&gt;
&lt;li&gt;Freeze anything that changes over time. If the agent can search internal messages, a replay today finds corrections posted after the original case.&lt;/li&gt;
&lt;li&gt;Make the replay unable to write or send. Check that by trying a write and seeing it fail.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;A small set is enough to see the pattern. We look at five to ten cases before running more.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 7: launch with review, then measure what people change
&lt;/h2&gt;

&lt;p&gt;At launch, a person reviews every output. Store the agent's draft and the version that was sent in separate fields. The difference between them is your measurement: in a review of 400 edited drafts, 53 percent of the edits removed a promise, 22 percent added stock status, and 10 percent replaced "we have no information" with an answer the knowledge base already held.&lt;/p&gt;

&lt;p&gt;Each number named a fix. Only when a type of reply goes out unchanged, case after case, do you let the agent send that type alone. In the 30 days to 19 November 2026, our email agent sent 3,331 replies alone and people sent 23,120 after review.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do when it goes wrong
&lt;/h2&gt;

&lt;p&gt;Remove before you add. Look for the repeated explanation, the two rules that contradict each other, the step that adds nothing. A new rule written for one case makes the next case worse. This is described in &lt;a href="https://ai.dingboai.com/en/ai-learns-from-agent-edits.html" rel="noopener noreferrer"&gt;how an AI agent learns from agent edits&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Build it yourself or hire it out
&lt;/h2&gt;

&lt;p&gt;The seven steps need someone who knows the job, someone who can connect the systems, and someone who watches the output every week. If you have all three, build. If not, the options are in &lt;a href="https://ai.dingboai.com/en/ai-customer-service-system-build-vs-buy.html" rel="noopener noreferrer"&gt;build, open source or buy&lt;/a&gt;. Our free plan lays out the seven steps for your job within two working days.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published at &lt;a href="https://ai.dingboai.com/en/how-to-build-an-ai-agent.html" rel="noopener noreferrer"&gt;ai.dingboai.com&lt;/a&gt;. Dingbo AI builds and runs AI agents for customer service, paid per completed task.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>llm</category>
      <category>productivity</category>
    </item>
  </channel>
</rss>
