As far as we know, Iran's first agentic platform

Ateamofagentsrunsyouronlinebusiness.Youstillmakethecall.

HodHodHub is not a chatbot. It is a team of AI agents wired into the store, marketplaces and messengers you already run. They sense what is happening, work out what matters, draft the next step — and ask you before anything consequential moves.

  • No card required
  • No code required
  • Nothing consequential without your approval
Business homeProduct demo sample
TodayWelcome backYour business today — the whole picture in one look.
Live
  • Messages answered

    128+12%

  • Orders and payments

    24+8%

  • New contacts

    62+5%

Channels4 connected1 needs review

AIHodHod's proposed stepA best-seller is running low on stock

Shall I pull it from tomorrow's campaign send so you don't lose the sale?

Press it — this is the same loop you get in the real product

Before we say a word about AI

These six jobs eat your day, every day.

We didn't invent this list. It's what Iranian online-store owners complain about most. Open any one of them and see which agent takes it off you.

Tap any one of them

What it costs you now

You retype the same product details across several panels, and something is still incomplete somewhere.

What changes with HodHodHub

Finds the incomplete records and drafts the fix for review; you just approve it.

Agent on the jobCatalogue agentLive

A tool waits to be told.An agent starts on its own.

What an agent is, and why now

Everyone has heard about AI. “Agentic” is a different thing.

A chatbot answers. An agent works. The difference isn't sentence quality — it's action and accountability. The quickest way to see it is side by side:

Getting started
Ordinary chatbotWaits for you to ask
AI agentReads the situation and starts on its own
Where answers come from
Ordinary chatbotThe model's general knowledge
AI agentYour own store and channel data
What you get
Ordinary chatbotIt hands back text
AI agentIt does the job — once you approve
Accountability
Ordinary chatbotLeaves no trail
AI agentEvery action files a reviewable receipt

And this isn't just a hype cycle:

  • 15%

    of day-to-day work decisions will be made autonomously by agentic AI by 2028 — up from 0% in 2024.

    Gartner · 2025
  • 60%

    of brands will use agentic AI for one-to-one customer interactions by 2028.

    Gartner · 2026
  • 40%

    of agentic AI projects will be cancelled by the end of 2027 — over cost, unclear business value and inadequate risk controls.

    Gartner · 2025

Why that third number matters to us

Most agentic projects fail on exactly that: nobody could show what value they created, and nothing stopped them doing the wrong thing. HodHodHub's approval gate and receipt ledger exist for precisely that reason — not as a side feature, but as our answer to the statistic we just quoted.

These three numbers are third-party analyst forecasts, not our own product metrics. Each carries its source and year beside it.

See the full loop

The agentic loop

“Agentic” is not a label. It is a closed loop.

Any platform can call itself agentic. The real question is how many steps of the loop it actually closes. HodHodHub closes all six — from sensing to receipt.

Execution loopSix steps, on repeat
01Senses
02Reasons
03Plans
04Asks
05Acts
06Proves

Then back to step one — this time with what it learned on this pass.

  1. 01

    SensesIt reads your real connections, not a model's general memory.

    Orders, stock, prices, customer messages and campaign results come from the systems your business actually runs on. What is not connected is not seen — and we do not pretend otherwise.

    Sources read just now

    • WooCommerce store
    • Messenger channels
    • SMS campaigns
    • Per-channel profit report
  2. 02

    ReasonsIt separates the one number that matters from the hundred that don't.

    Conversion drops, stock-out risk, stalled orders, customers about to churn. Every finding arrives with its reasoning — and when the data is thin, it says so instead of guessing.

    Pattern found

    Product views
    flat
    Conversion rate
    falling
    Stock
    below threshold
  3. 03

    PlansIt turns the next step into a specific plan.

    What changes, on which channel, in what order, and what could go wrong. The plan is fully on the table before anything is touched.

    Proposed plan

    1. Pull the low-stock item from tomorrow's campaign
    2. Send the supplier a restock reminder
    3. Re-enable the campaign once stock returns
  4. 04

    AsksThis is the step most agents skip. We stop here.

    You see the exact before and after, then decide. No consequential change — price, stock, a campaign send — executes without passing this gate.

    Ready for approval

    Before
    campaign ignores stock levels
    After
    send only for items in stock
  5. 05

    ActsIt finishes the job inside the boundary you drew.

    Price, stock, campaign sends, customer replies — exactly as far as you allowed and not one step further. If it hits the boundary mid-task, it stops and asks.

    Executing

    • Campaign rule saved
    • Item removed from tomorrow's send
    • Supplier reminder queued
  6. 06

    ProvesEverything it did stays reviewable.

    What happened, when, on whose approval, and what came of it. The receipt exists for follow-up and audit — so you don't have to trust us, you can just look.

    Receipt filed

    Action
    campaign stock rule
    Approved by
    business owner
    State
    written to history

The product itself

That all sounds good. Now look at the actual platform.

The four places you genuinely open every day. Each one removes one of the problems above — and you can work with them right here.

It's usable — try it

Problem: chasing orders

Start the morning with one clear picture, not ten tabs.

Business home is your day in summary: what sold, what's stalled, which connection needs a look, and the step HodHod proposes next.

  • Stalled orders surface before the customer walks
  • Every proposal has an approve button, not just a report
  • One screen, no bouncing between panels
HomeProduct demo sample

Today in summary

What needs you

  • All main connections are healthy
  • One order has been stalled over two hoursOpen
  • Two actions are waiting on your decisionOpen

Authority boundary

How far is it allowed to go? You decide.

Take one real event and watch exactly how HodHodHub behaves at each level of authority. This is the only setting you have to think about — the rest runs itself.

Event

A best-selling item drops to three in stock — and it ships in tomorrow's SMS campaign.

DefaultWhat the agent does

I prepare the whole plan and wait for your approval.

I'll pull the item from tomorrow's campaign and send the supplier a reminder — the moment you press approve. Until then, nothing moves.

Execution trace

  1. Three-step plan drafted with reasoning
  2. Before and after prepared for review
  3. Awaiting owner approval

All three levels share one thing: a receipt. No action goes untraced.

Straight answers

We'll also tell you what isn't there yet.

In a market where “AI” is being stapled onto everything, the simplest way to earn trust is to draw our own limits ourselves.

Live today

  • Reading from a WooCommerce store and a WordPress site
  • Customer replies and bot building across messengers
  • SMS and messenger campaigns with audience segments
  • Revenue, cost, net-profit and per-channel reporting
  • The propose → approve → act → receipt loop

In progress

  • Full depth of Digikala seller operations
  • Automated competitor price watch on Torob
  • Immutable final receipts for external audit

Things we will never do

  • Change a price or stock level outside the boundary you set
  • Answer a customer by guessing when there is no real data
  • Put unsourced numbers, badges or testimonials on our own pages

Common questions

What people ask before handing over the work.

Short, clear answers. If something else is on your mind, HodHod is right there.

What does “AI agent” actually mean?

An agent is an AI system that doesn't just answer — it does the work: it reads the situation itself, decides what matters, drafts a plan, asks permission, executes and files a receipt. A chatbot waits for a question and hands back text; an agent starts on its own and takes responsibility for what it did.

How is this different from ChatGPT or an ordinary chatbot?

A general chatbot answers from general data and does nothing for you. HodHod is connected to your real store, marketplaces and messengers, answers from that data, and instead of text proposes a specific action with an approval, execution and receipt path.

Is it really Iran's first agentic platform?

As far as we know and have been able to check, no other Iranian platform has shipped this full loop — reading real connections through to approved execution and a receipt — for online businesses. If you know of one, tell us and we'll correct the claim.

Will the agent change things in my business without permission?

No. Consequential changes — price, stock, campaign sends, switching automations on — pass through the approval gate. The only exception is a bounded autonomy rule you defined yourself, and even that files a receipt.

What if the agent gets it wrong?

You have two guards. First, every consequential action reaches you before it executes, so mistakes are usually caught at the proposal stage. Second, anything that did execute is in the receipt ledger, and rules and automations can be switched off at any moment.

Do I need to build a new store or system to start?

No. HodHodHub doesn't replace WooCommerce, WordPress or your current channels; it connects to what you already run and lays an orchestration layer over it.

Where does my business data go?

Connection credentials live in a secure vault, sensitive values are masked and access-limited, and access is scoped to your own workspace. The agents read from that data — not from the model's general memory.

Do I need to write code to set this up?

No. The whole experience is conversational: you say what you want and the agents get moving. API and webhook access is there for technical teams.

Give it the first objective

Tell the agents what matters today.

You don't have to hand over everything on day one. Give them one objective, watch how they work, then widen the boundary as far as you want.

No card required · No code required · Assisted trial in a sandbox