You’ve heard “study to construct”, “ship your individual instruments”, “automate your busywork”, however no one explains the place to begin.
Canva made this painfully clear: they gave 5,000 staff a full week to discover AI, and most couldn’t figure out how to start. No shock there, most vibe-coding instruments count on you to handle information, wrangle a terminal, and deploy to a server identical to a developer would.
Excellent news, you don’t want any of this to construct your individual advertising and marketing instruments and workflows.
I’ll stroll you thru the precise setup I exploit. It runs through Letaido within the browser, so that you don’t want to put in something, handle information, or go away your laptop computer working in a single day.
Suggestion
🔮 Letaido is an AI workspace that lives in your browser. It writes the code, shops the info, hosts the instruments, and runs your automations on a schedule.
Vibe coding helps you to construct the reporting dashboards, competitor displays, keyword-clustering instruments, and content material workflows you’d usually wait on a dev staff (or a funds) for. You describe the workflow the way in which you’d clarify it to a brand new rent, and the AI builds it.
Letaido is the AI advertising and marketing platform made for the job. You level it at your instruments and information, describe the report or workflow you need, and it builds and runs it within the cloud for you.
You may have two methods to create your builds.
There’s a personal Console solely your staff can see (like inner instruments and dashboards):
There’s additionally a public site (for pages you wish to publicly share):
Just a few issues price realizing up entrance:
All fashions in a single place. You’ll be able to change between Claude, GPT, Gemini, and others from a dropdown, per activity. No want for separate subscriptions or API keys. 
Native connectors. 35+ advertising and marketing instruments like Slack, HubSpot, Notion, GitHub, Linear, Mailchimp, Stripe, WordPress, or Airtable already linked. 
Pre-built advertising and marketing expertise. An Ahrefs-maintained library of expertise, a set of steps to your AI agent to observe, primarily based on actual finest practices, so that you’re not ranging from a clean immediate. 
Constructed for groups. No matter one individual units up (expertise, connectors, reminiscence, apps) is there for everybody, so the staff compounds one another’s work as an alternative of rebuilding it. Seats are free (the $99/month is per workspace, not per individual), so you’ll be able to invite everybody, and roles maintain it secure: members can chat and construct, homeowners and admins approve new connectors or deal with secrets and techniques. 
Share outdoors the staff. For a one-off like a consumer report, ask the agent to create a visitor hyperlink to your app or report, and it creates a public website with a default Letaido area. It’s also possible to change the area to 1 you already personal, or purchase a brand new one inside Letaido, and it’ll type out the safety certificates for you as effectively. 
Internet hosting and secrets and techniques are dealt with for you. This removes the 2 greatest newbie dangers: uncovered API keys pushed to GitHub and unintentional public experiences. You solely want to think about the place to publish your workflows – inner information belongs within the Console, whereas the general public website is a URL anybody can discover and Google can index, so double-check earlier than it goes dwell.
Suggestion
⚠️ If you wish to observe together with me, you’ll want Ahrefs. The $99/month covers the Letaido platform and your AI credit, however the agent reads your Ahrefs undertaking information by your present Ahrefs plan, so an energetic account is required, and your information limits (tracked key phrases, initiatives, and so forth.) observe no matter plan you’re on.
- Join at letaido.com – hit “Get began” and create an account.
- Identify your group – the subdomain you choose turns into the URL the place your workspace lives.
- Invite your staff – seats are limitless and free
That’s it. No editor, no Git, no keys to paste.
Select an AI mannequin and mode
What’s nice about Letaido is that you simply’re not locked right into a single mannequin. Your subscription contains $50 in AI credit, so you’ll be able to change fashions freely to match regardless of the activity calls for.
On the mannequin itself, three elements matter, and so they shift always as suppliers ship new variations: high quality, pace, and value. Artificial Analysis plots all three throughout 500+ fashions if you’d like the dwell image.
In case you nonetheless fear about spending your $50 AI credit, let me provide you with an instance with several types of AI fashions.
Let’s say you ask Letaido to have a look at your Ahrefs undertaking, discover key phrases your prime 3 opponents rank for that you simply don’t, cluster them by matter, and hand you a prioritized listing of content material concepts with goal key phrase, quantity, issue, and a one-line angle for every, the spend can appear to be this: 
| Mannequin | Tier | Price per run | $50/month | While you’d choose it |
|---|---|---|---|---|
| GPT-5.4-mini | Low cost workhorse | 3¢ | ~1,600 runs | Quick, high-volume, easy reasoning |
| Claude Sonnet 4.6 | All-rounder | 24¢ | ~200 runs | Sharper clustering + extra logical reasoning |
| Claude Fable 5 | High-tier | 38¢ | ~130 runs | Highest-stakes strategic calls and proposals |
My default is Claude for something the place I describe intent loosely and wish it to determine the form. I change to a less expensive mannequin for high-volume, mechanical work (reformatting, batch variations) to maintain credit score burn down.
Proper subsequent to it’s a mode selector, which defines the construct mode the agent adopts: 
- Chat — discuss it by. It solutions and discusses, however received’t begin constructing.
- Construct — make or change one thing. It plans, then ships.
- Analyze — crunch information for perception, to not construct new software program.
- Plan — determine what and how earlier than doing it.
- Brainstorm — generate choices and instructions, no committing but.
- Auto — it reads your intent and picks the suitable posture itself.
Auto is a advantageous default. As an illustration, in case you’re in Chat mode and also you need one thing constructed, it’ll give you to modify.
Suggestion
💡What burns credit quick is working the heaviest mannequin on gentle work, and dragging one big chat throughout unrelated duties. Begin a contemporary chat when a activity is finished (your memory persists, so that you don’t lose context), and let cheaper fashions, like GPT-5.4 Mini, deal with the grunt work.
Create a data base
The context you write your self beats any exterior supply, as a result of it holds particulars no mannequin was skilled on: your actual options, positioning, and voice. Letaido offers you a number of methods to retailer details about your organization, so that you’re not re-explaining it each chat:
- Reminiscence. The agent reads ~/workspace/.reminiscence.md at first of each chat utilizing a number of tokens. I like to recommend including solely important data the agent makes use of each run (e.g. recurring vocabulary or stylistic guidelines). An excessive amount of data saved = extra tokens spent, maintain it transient.
- A shared staff wiki. Letaido retains a workspace data base the place you’ll be able to retailer a great deal of information you’ll be able to refer the mannequin to (choices, competitor information, design information, particular initiatives, and so forth.).
| Class | Reminiscence (.reminiscence.md) | Workforce Wiki (wiki/) |
|---|---|---|
| What it is | Important notes file | Workspace data base |
| When Letaido reads it | Begin of each chat (mechanically) | Just for particular duties |
| Token value | Each session (scales with measurement) | Solely when a web page is learn |
| Maintain it | Brief | Could be massive |
| Finest for | Recurring vocabulary; Stylistic/voice guidelines; A pointer to the wiki | Technique docs; ICPs; Rivals; Product FAQ |
| Rule of thumb | Solely what Letaido wants each time | What Letaido wants often / in depth |
| Discover it | controls -> reminiscence | controls -> wiki |
Create expertise, apps, experiences, artifacts, or automations
5 phrases you’ll see all over the place in Letaido. It’s necessary to tell apart between them, as constructing the flawed kind can affect the agent’s conduct and reusability.
A ability
A ability is a reusable set of steps, an output format, and guidelines that the agent reads on demand each time a request you say in plain language matches. The quickest technique to begin is to clone one from the Ahrefs library and customise it. 
It’s also possible to create your individual through chat.
- Give the ability a reputation that may set off it
- Describe the principles it ought to observe
- Reference the examples if vital
An app
An app is an interactive software the agent builds. It could learn your databases, name your connectors, and present outcomes. 
A month-to-month report you configure and verify within the console is an app, so is a writing workflow that takes a key phrase and fingers again a draft. 
The second you create, edit, or set off issues inside it, it’s an app.
A report
A report is a dwell, read-only view of your information. The agent builds it as soon as, and each time you open it, it pulls contemporary numbers and lays them out for you. 
A sentiment overview that’s present everytime you look, a lead-quality breakdown pulled straight out of your CRM, these are experiences. You don’t configure or handle something inside them, you simply learn. 
An artifact
An artifact is a frozen output, one thing the agent produced as soon as that doesn’t change afterward. 
A chart from final month’s competitor G2 evaluations, a one-off visualization. They’re snapshots: good for “right here’s what we discovered”. 
Suggestion
🔮 Apps, experiences, and artifacts can dwell in your Console and be shared together with your staff solely, or be revealed to your public website for anybody with a hyperlink to use.
An automation
Automation is a activity that runs by itself, at a particular time, with out you triggering it. There are two methods to create it:
With an agent – you describe what you need completed and when, the agent writes it, and runs it from then on: 
Or manually – automate an agent or a easy script to run with a set of limitations (schedule, max tokens used, and so forth.): 
Suggestion
💡 Rule of thumb: bundle know-how you’d in any other case re-explain as a ability, construct the factor you click on into as an app, and put an automation on prime when it ought to run with out you. A single construct can use all three, reminiscent of an automation that runs a ability on a schedule and writes the consequence into an app.
Add integrations
If you wish to pull your HubSpot information, scrape social media with Apify, or use different exterior instruments, the agent wants to make use of the connectors to deal with the endpoints, parameters, and the authentication behind the scenes. As an illustration, you join HubSpot as soon as, and each time you say “pull this week’s offers from HubSpot”, the agent is aware of precisely what to do.
Connectors can learn, but additionally take motion through your instruments – submit a abstract to Slack, publish a submit to WordPress, open a Linear ticket, replace a HubSpot deal, or ship an e-mail.
There are tons of widespread instruments like Stripe, Fathom, or Mailchimp obtainable – a full listing of connectors may be discovered here.
So as to add entry to your connector:
- Inform the agent through chat you wish to add the precise connector
- Get the API keys (click on “ get” within the card in case you need assistance finding it)
- Paste the keys and click on “Submit”
- Your connector is added underneath “Controls” → “Permitted”, you’ll be able to revoke it at any time
Each good construct comes all the way down to the identical 4 strikes: write the immediate, level it at your information, share it with the staff, and begin with one workflow you already know.
Writing the immediate
Not like a one-off chat message, that is the instruction the agent will execute time and again, which suggests something obscure or lacking will get repeated on each run. Three issues do a lot of the work:
- What output you need. Describe the completed factor. A Slack message, a desk, a abstract as concretely as you can.
- The place the info comes from. Identify the precise supply the agent ought to pull from, whether or not that’s a connector, a file, or a web page to scrape, so it by no means has to guess.
- What the directions and limiters are. Spell out the steps it ought to observe and, simply as importantly, the boundaries it shouldn’t cross. What it ought to examine, ignore, and go away untouched.
Just a few examples: 
- Output = Slack updates about new and misplaced offers each week.
- Knowledge = HubSpot offers (native connector).
- Instruction = examine this week to final, flag new offers vs offers misplaced, submit to #sales-updates.
- The place it lives = runs each Friday at 8am by itself -> a scheduled job in Letaido.
The perfect practices boil down to:
- Be particular. Imprecise requests set off a clarifying questionnaire. Not “assist me with onboarding emails” however “draft a 4-email welcome sequence for brand spanking new signups, one CTA every, underneath 120 phrases.”
- Specify the format. “Reply in a desk,” “JSON solely,” “markdown with headings.” Default outputs are conversational, so saying the form up entrance saves edits.
- Say what to not do. “Don’t summarize what stayed the identical.”, “Reply precisely: No modifications detected”.
- Ask for a plan when it’s fuzzy. If a construct touches a number of methods otherwise you’re undecided it’s possible, ask for a plan.
- Feed info, by no means recall them. Level it on the connector or file. It ought to motive over what you give it, by no means provide your metrics from reminiscence.
Map the place your information exists
Earlier than you’ll be able to construct something, it’s essential to know what information you’re working with and the place it lives. Good builds run on actual information, and that information is already scattered throughout the instruments you utilize each day.
1. Record each software, app, and website you contact for work
Your analytics platform, your CRM, your undertaking tracker, the spreadsheets, the dashboards. For each, write down the particular information factors it holds. Don’t cease at “website positioning information”; get all the way down to “key phrase positions, clicks by web page, each day impressions.”
2. Join it to Letaido
As soon as you realize the place your information lives, there are 3 ways to push it to Letaido:
Connectors – Examine whether or not your software is already in Letaido’s connectors listing, whether it is, ask to attach it through + New chat.
API keys – in case your software isn’t within the connectors listing. Generate a normal API key from the software, then ask the agent within the chat to attach it.
Net scrape – for something on a public web page, no login required. Suppose competitor pricing, evaluate websites, or SERP. 
TLDR
- Join Letaido and join the one software your work lives in most.
- Decide the workflow you do most frequently. For me, reporting was the best win as a result of the steps by no means modified.
- Write it as plain steps, 1 by 5, the way in which you’d clarify it to a brand new rent.
- Describe it to Letaido one step at a time, and let it construct with you.
- Examine the output, then put an automation on it.
What to construct first
In case your head feels fuzzy, get impressed by the builds my colleagues shared for automating SEO, product marketing, or content marketing.
You can run an end-to-end keyword research process – take your major key phrases, type them into clusters, and create a content material transient 
Create a paid adverts marketing campaign primarily based on a competitor’s spend – discover key phrases competitor bids on, fetch and analyze their touchdown pages, generate creatives and a number of copy variations per advert group. 
Replace previous articles and combat site visitors decline – the pipeline fetches articles, extracts the web page content material, then runs diagnostics. 
You do not want a dozen builds. You want one, completed, from your individual context.
Actuality verify
The hype is loud proper now. These builds is not going to exchange your technique, your style, or your relationships. They won’t magically rank you primary. The primary model of something you construct shall be uncooked.
Nevertheless, automating your work will take the repetitive 70% off your plate so you’ll be able to spend your hours on the enjoyable stuff. The leverage comes from wiring steps collectively so a workflow you used to do by hand now runs from one command – or, with Letaido, by itself each Monday at 9am.
You recognize your job higher than any software does. Begin there, construct one small factor, and iterate it to your high quality bar.
While you give it a go, tag/DM me! I’d like to see your builds in motion 🙂
