Original Title: how non-developers automate work like engineers (without writing code)...
Original Author: Damian Player
Translation by: Peggy, BlockBeats
Editor's Note: While most people still view AI as a "more efficient search tool," Perplexity is beginning to perform work.
This article revolves around a repeatedly overlooked difference—why some people get a segment of an answer when using AI, while others directly obtain deliverable results. The key is not in the model's capabilities but in the usage: is the tool treated as a conversational window or as an executable system that can be commanded and scheduled?
A new class of tools represented by Perplexity Computer replaces "asking questions" with "defining tasks" as the core mode of interaction. From contract review and competitive analysis to data cleaning and report generation, users no longer describe problems but directly define final deliverables. Coupled with the connection of enterprise tools and the solidification of personal context and style examples, this capability further evolves from one-time outputs to reusable, automated workflows.
More importantly, the boundaries of automation are being redefined. It is no longer merely about assisting in completing a step but can continuously operate, execute across tools, and even proactively suggest supplementary tasks. This means the relationship between people and tools is shifting from "usage" to "management and delegation."
In this change, the true watershed is no longer whether to use AI, but whether one has begun to use it to "deliver results."
Here is the original text:
Those who understand this will gain an asymmetric advantage. Soon, everyone will learn how to do it. But before everything becomes obvious, here are ways you can start in advance.
In the past year, developers have been running autonomous AI agents (like Claude Code, OpenClaw, etc.) in the background, which can conduct research, build products, and deliver complete results without requiring constant supervision or back-and-forth prompts. But you typically have not been able to leverage this unless you can use the terminal and write code.
Perplexity Computer has changed that. For the first time, non-developers can utilize the same capabilities. All you need is a browser and a task you can hand over to it.
Most people open Perplexity, input a question, get an answer, and then close the page. They miss the point. Perplexity Computer is not designed to answer questions; it is designed to execute tasks.
Stop asking questions and start giving it real work.
Why Most People Fail
Finance directors, lawyers, consultants… they open the tool, type in a question, get a decent answer, and then think, "Oh, a more advanced Google." Then they go back to spending 90 minutes cleaning that same spreadsheet they just cleaned last Monday.
The problem isn't the tool; it's the usage. They treat it like a chatbot.
Questioning Method: "What risks are there in this contract?"
Task Method: "Review this contract. Check each statement for supporting public sources; mark vague wording, missing standard clauses, and potential legal liabilities; list the top 5 risks with specific clause references; output a Word document with revision marks."
Same contract. One method gives you a list for you to read; the other provides you with a ready-to-send product.
Set Up This System in Just 10 Minutes
First, connect the tools. Click on the connectors in the sidebar. Perplexity can connect to over 400 applications: Gmail, Google Drive, Slack, Salesforce, Notion, SharePoint… connect all the ones you actually use.
Then let it know who you are. A single input is enough: "I work in a certain position at a certain type of company. I regularly produce these contents: X, Y, Z. Please remember this background in every conversation." It will retain this information long-term.
Next, tell it "what is good." Find 2-3 of your most satisfactory outputs, upload them, and input: "These are my best work samples. Please learn their format and tone for reference when generating content in the future."
That way, it is not guessing your style, but rather reverse-engineering a successful path you have already validated.
In 10 minutes, do this first.
A Real Example: The Monday That No Longer Wastes 90 Minutes
A financial analyst receives a data export every Monday, 150 rows, messy formatting: duplicate data, three date formats, and ratings written in words rather than numbers. Every week, before analyzing, she spends 90 minutes cleaning the data. The same issue recurs weekly.
She input a single command: Clean this file, remove duplicates, standardize date formats, convert text ratings to numbers; analyze the cleaned data; generate an interactive dashboard with filtering capabilities and provide a share link; output a PDF report comparing before and after cleaning; save all files to the "Monday Reports" folder in Drive.
Four minutes later: a clean dataset, an interactive dashboard, a share link, and a PDF report—all appeared in her Drive.
Then she asked: "Are there any improvements I haven't asked about that could make this more useful?"
The system suggested two points: first, set this task to run automatically every Monday at 7 AM; second, add a task to generate a management brief for Tuesday based on poorly performing sections.
She set both and closed the page.
From then on, every Monday, it would run automatically—regardless of whether her computer was on.
This is precisely the capability that developers have been using for the past year. Now, you can use it in your browser.
What People Are Already Using It For
@gregisenberg did a live test on the @startupideaspod podcast.
He provided just one task: identify companies advertising on competing podcasts, find the real sponsors, and write a personalized email for each.
The system found Ramp's VP of Growth, pulled content from a podcast episode he participated in two weeks ago, drafted a cold email referencing his specific statements in the show, and then sent it directly. Greg did not say "send;" the system determined the task was complete and executed it on its own.
It then proactively suggested: monitor competing podcasts, and as soon as new brands start advertising, immediately notify and include corresponding contacts—"contact them right when the budget is initiated."
Ultimately, this process concurrently researched 96 potential clients and scheduled follow-up emails for Day 3 and Day 7.
On the Marketing Against the Grain podcast, the team used it to audit the entire HubSpot product page: automatically crawl the entire site, score according to custom criteria, sort issues, and generate a shareable site report. What would have taken the team a week was completed while recording the episode.
These were all accomplished live, not demonstrations or scripted presets.
Usage for Specific Tasks
InFinance, a portfolio analyst gave just one task before NVIDIA's earnings report was released.
The result was: a real-time interactive dashboard containing $130.5 billion in revenue, a 75% gross margin, a 114.2% growth rate, a complete profit and loss statement, and projected profit margin trends from fiscal year 2021 to 2028, all supporting filtering and sharing links.
No Excel, no manual data search, completed in 5 minutes.
Perplexity can directly access data sources like SEC disclosures, FactSet, S&P Global, PitchBook, and more—no API key or additional authorization needed; the functionality is built into the system.
Legal Scenario:
"Review this contract. Check each statement for supporting public sources; mark vague wording, missing standard clauses, and content that could result in legal liability under [specific state] contract law; list the top 5 risks with specific clause references; output a Word document with revision marks."
A reviewer once uploaded a proposal claiming a certain market grew 43% year-over-year. Perplexity Computer found that the real data was only 4%, intercepting the issue before signing.
Marketing Scenario:
"Analyze the top-performing content over the past 30 days from [Competitor 1], [Competitor 2], and [Competitor 3]; identify the content form and themes with the highest interactions; detect content gaps; generate a 30-day content calendar based on these gaps, and save it as a Google Doc."
Set it as a scheduled task. Weekly, it automatically generates the latest competitive analysis without manual research.
Operations Scenario:
"This is our Q1 CSV data. Please clean the data; analyze revenue by region and product line; identify the three biggest issues; generate a one-page action recommendation; create a one-page report PowerPoint; save all files to the project folder."
Five deliverables, one command. While you are in a meeting, it's already done.
Model Council: 60 Seconds to Three Judgments
When faced with a decision that has real consequences, simply input the question once. Perplexity will simultaneously call Claude, ChatGPT, and Gemini and summarize their consensus and divergences via a "combinator."
· Areas of agreement among all three: high-confidence conclusions.
· Areas of divergence: require further judgment.
Someone asked whether to choose a price of $297 or $497 for a product. The three models provided different answers, but the combinator found that their only agreement was: do not set it below $297. The decision was thus made.
Many companies spend money hiring consulting firms to lock analysts in a conference room to reach conclusions.
Here, you only need one command.
The True Core Capability
To gain actual value from Perplexity Computer, 80% depends on one thing: can you clearly describe the "final output."
It's not about technical configuration. It's about whether you are clear enough about what you want to deliver. Do not describe steps; describe the outcome.
After each task is completed, remember to ask one more question: "Is there anything I haven't asked that could make this result more useful?"
It will almost always point out blind spots. Use it every time.
Start Here
Open Perplexity (pro version $20/month). Go to the Computer page, click connectors, and first connect Gmail and Google Drive.
Input your three background sentences (only once). Upload 2-3 of your best work samples to teach it your style. Then choose a task you spent over 2 hours on last week that is similar every time: describe it in terms of "final deliverables," and send it. Observe the execution process. If it is a recurring task, set it to run automatically before closing the page.
Developers have been using this for a year. The gap in output between them and others is real.
This is the way to close that gap.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
