AI implementation in business: where to start and how to launch a turnkey AI bot
Mikhail, MOZGupdated November 2, 2026 · 11 min readcontents11 min read, 8 sections
The best way to bring AI into a business is to start with one process where the company loses money and time every day. Most often that's customer messaging: the same questions over and over, leads coming in at night, customers disappearing once they hear the price.
That's where an AI bot goes. It's also called an AI agent or an AI chatbot. Below we explain how to find that process in your business, what else AI already does for us and our clients, how we launch a turnkey bot and what it costs.
- Start with the process where the losses show up in money and time, usually customer messaging. At a laundry in Bali, the bot raised conversion to orders from about 25% to 43%.
- We launch a turnkey AI bot for customer messaging in about a month: 2 weeks of development and about 2 weeks of testing.
- With real conversations, the bot gets really good within another 1-2 months, and the first month of support is included in the price.
- To get started, we need about 20 full conversations with your customers, your answers to the brief and materials such as your catalog and price list.
- Developing an AI bot costs from $800, most often up to $1,900, plus the platform from $60 a month. We start with a chat and funnel audit for $0.
Take the test: where to start in your business
Check what applies to you:
How to read your result:
- Any of items 1-4. Start with an AI bot for customer messaging. The first step here is a chat and funnel audit for $0.
- Item 5. You need a bot that pulls data from the CRM on its own. We connect Kommo, amoCRM, Bitrix24, RetailCRM and companies' own CRMs through an API.
- Item 6. That's work inside your team, separate from the bot. Start with a consultation: together we'll work out where AI can take routine off you and your team. If you need a ready-made internal tool, our team will build it turnkey.
- None of them. If you don't get many leads and a manager replies within minutes, you may not need a bot yet. We'll tell you that during the audit.
Which process to start with when implementing AI in your business
We start with one question: where are you losing money and time right now? A customer messages you at night, no one replies by morning, and they order from a competitor. A loss like that shows up in the numbers, so this process goes first.
It's usually customer messaging. An AI bot understands everyday language, answers from your knowledge base and takes the customer all the way to an order on its own. The result is easy to measure: conversion to orders and the night orders you used to lose.
We recommend implementing artificial intelligence in your business step by step: first messaging and leads, connected to your CRM from the start if you need that, then work inside the team, such as content, calls and data collection.
Take on the next process once the first one starts delivering results. There's no need to bring AI into every department at once.
Example: a laundry in Bali
25% → 43%conversion in 3 monthsThe laundry was getting more and more inquiries, the managers couldn't keep up, and at night no one was there to reply, so those orders slipped away. Customers wrote in Russian, English and Indonesian.
We set up an AI bot on WhatsApp and Instagram. It replies in the customer's language day and night and takes the order on its own. If someone goes quiet, the bot follows up after 10 minutes or after an hour. It now handles about 10,000 conversations a month, and conversion to orders has grown from about 25% to 43%.
AI agents in business processes: what we've already launched
| where AI works | what it does |
|---|---|
| Customer messaging | Replies day and night in the right language, understands voice messages and photos, takes orders, follows up with people who went quiet |
| Recommendations and sales | Picks a shoe model, a car for the right dates, a driving school in the right city or a product from a catalog of 5,200 items, then takes the customer to an order or payment |
| Lead qualification | Asks about the goal, budget, area and timing, and passes only hot leads to a manager |
| Social media | Replies to comments under posts in the brand's voice, takes requests from bloggers |
| CRM data | Checks stock and available dates, records leads and orders in amoCRM, RetailCRM and the company's own CRM |
| Content | Writes posts and scripts, edits videos, makes carousels and publishes to social media on its own |
| Calls and notes | Sorts a call recording into notes and tasks, and answers questions from those notes as a “second brain” |
| Website and research | Writes website code and copy, collects search volumes and article topics |
At Avto In Line, online course sales with the bot grew by 300%. Read on for how we launch a bot like this and what it costs. There are more examples in the Comfers and Carloson Club case studies and in the article “AI bots for business: examples”.
Work inside the team is what a consultation is for. If you need a ready-made internal tool, our team will build it turnkey. For example, for a laundry we built a system that tracks items from intake to final check.
How we launch a turnkey AI bot
Turnkey means we take care of the audit, the build, the launch and the support. All we need from you is your materials and feedback.
- Chat and funnel audit for $0. On a call, we go through your chats and sales funnel, tell you whether you need a bot and give you a price.
- Logic and knowledge base. We define the conversation flows, the tone of voice and what the bot needs to know.
- Setup and testing. We build the bot based on the brief and test the conversations ourselves, then with a focus group, for example people you know.
- Launch and the first month of support. For the first weeks, we read the conversations ourselves and keep training the bot.
Timeline
Development takes about 2 weeks if we get your materials on time. Testing takes about 2 more. That's the timeline for a standard bot.
Complex integrations, large catalogs or several bots in one project take longer, and we'll give you the exact timeline during the audit. It usually takes 1-2 months of real conversations to fully tune a bot.
What we need from you
- about 20 full conversations with your customers;
- your answers to the brief;
- materials on your business: catalog, price list, presentations;
- your CRM funnel, if you have one, and CRM access if you need an integration.
How much it costs
Developing and launching an AI bot costs from $800, most often up to $1,900. You pay in two stages: 50% before work begins and 50% after launch. The first month of support is included in the price.
On top of that comes the platform the bot runs on: from $60 a month for the subscription and tokens, and you pay for it directly. The cost depends on your traffic, the bot's settings and a few other things. We'll give you the exact development price and a rough figure for the platform during the audit.
From month 2, support is optional: the Basic plan for $280 a month (you send feedback, we make the changes) or the Proactive plan from $450 a month (we watch the conversations ourselves and add new conversation flows). We break down what goes into the price in the article “How much an AI bot costs”.
If you'd like to figure things out first: a consultation
Sometimes it's too early for a bot, or the real task lies somewhere else. Or you just want to understand where AI can save time for you and your team. That's what consultations are for, and I run them myself.
The chat and funnel audit for $0 only covers the bot. A consultation is paid and broader: it's about AI in general. For example, which tasks to give AI first, how to connect it to your CRM and messaging apps, how to make content faster, how to train your team, how to build your own service in Claude Code.
You can bring your team. We work through your task using a real example. If it's about your own service, we start building it right in the meeting. That's vibe coding: you explain the task in plain language and AI writes the code, so you don't need to know how to program.
After the meeting you'll have a first working result or a clear plan for what's next. From there, you can carry on by yourself or with me.
- Single meeting: $200 an hour, online or in person if we're in the same city.
- Meeting series: $150 an hour when you pay for 6+ hours. A good fit when the task is bigger than one conversation: bringing AI into a department, building your own service or training your team.
Need a ready-made solution? Our team will build it turnkey: a website, an internal tool, a CRM, automation. You'll get a price and timeline once we've looked at your task.
More on the Consultations page.
