AI customer service agent: buy one or have it built
Buy an AI customer service agent as a product when the answers sit in your website, PDFs and FAQ. Have one built when the agent must read or write live data in your ticketing system, ERP or CRM.
Customer service AI mostly answers the same twenty questions; the first decision is whether you buy the agent as a product or have it built.
Buy the product when the answers are already in your content; have an agent built when they live in your systems. SimplyBoost is our own product: an AI support agent that answers from your website, PDFs, FAQ and catalogue on website chat, WhatsApp and Instagram, from €39 a month. It is live in days and it fits most businesses. A custom agent, which Crux Digits builds, starts at a 2,500 EUR audit and makes sense when the answer has to come out of your own systems rather than your own content: an order in an ERP, a ticket in Zendesk or TOPdesk, a policy in a claims system. This page is written so you can tell which one you are.
Three fixed facts
What is settled before the agent answers its first question
- 2 Aug 2026AI Act transparency in forceArticle 50 of the EU AI Act: a customer must be told they are talking to an AI system. One line in the opening message.
- €39 a monthSimplyBoost Starter, 500 AI repliesOur own product, answering from your website, PDFs and FAQ on website chat, WhatsApp and Instagram. Live in days.
- 4 to 6 weeksProduction-ready MVP, €20,000 fixedA custom agent on one channel, reading live data from your ticketing system, ERP or CRM, measured on your own tickets.
Sources: Regulation (EU) 2024/1689, Article 50; simplyboost.io/pricing; our pricing.
What it looks like in practice
Four moments, not four features
The same agent, in the places your customers already are.
The question that arrives when you are closed
Most support questions land outside the hours you staff. That is the gap the agent covers first, and the one nobody is paid to sit through.
One agent, trained once
Website chat, WhatsApp and Instagram DMs get the same answers, because it learns your content once rather than per channel.
What it does when it does not know
It says so and passes the conversation to a person, with the transcript attached, into Slack, email or your helpdesk.
In the language the customer wrote in
It answers in the language it was asked in. Deciding which languages you staff for the handover is the part that needs a decision.
What is AI customer service?
AI customer service is software that answers customer questions in writing, on your website, in WhatsApp or by email, from your own content, and hands the conversation to a person when it does not know or when the question needs a decision. The useful version is grounded: it quotes your knowledge base, your product documentation and your policies, and it refuses rather than guesses.
Two kinds sit under the same name. A chatbot answers from content. An AI agent also reads or writes in your systems: it looks up an order, opens a ticket, checks a warranty. Which of the two you need decides whether you buy a product or have something built, and that is what the rest of this page is about. The longer comparison: AI agent or chatbot.
Buy the product, or have one built
| SimplyBoost, from €39 a month | A custom agent, from a 2,500 EUR audit | |
|---|---|---|
| Where the answers come from | Your own content: website, PDFs, FAQ, catalogue | Your own systems: ERP, WMS, CRM, claims or patient records, through their interfaces |
| Live in | Days. One line on the site, three steps on WhatsApp | Weeks. The integration is the work, not the model |
| Channels | Website chat, WhatsApp, Instagram DMs | Whatever you already run, including the ones a product cannot reach |
| Who runs it after | You do, in the dashboard | Your team or your ICT partner, with the code and the evaluation set handed over |
| Best when | The twenty repeat questions are answerable from what you have already published | An answer is wrong unless it reads live data, or the process is regulated |
| Worst when | Every answer needs a live lookup in a system | The questions were never that complicated and a product would have done |
What you get
Three things a support agent has to do before it is worth switching on
A knowledge base ready for AIAn answer from your own content
Website, PDFs, FAQ, catalogue, knowledge articles. It does not know anything else, and it says so when the answer is not there.
Connecting your ticketing systemA handover with the transcript attached
Into Slack, email or your helpdesk. Zendesk, Freshdesk, TOPdesk or Salesforce Service Cloud stays the record; the agent creates or updates the ticket.
The number to check firstA measured share before you buy anything
A sample of your own closed tickets, marked for what a knowledge article could already have answered. That share is the ceiling for the product; the rest is what a built agent adds.
AI customer service software: what you buy, and what you have built
Most searches for AI customer service software end at one of two things. The first is a product that answers from your own content: SimplyBoost, our own product, is trained on your website, PDFs and FAQ, answers on website chat, WhatsApp and Instagram, books appointments in Google Calendar, reads Shopify order status and hands a conversation to Zendesk, Freshdesk, Gorgias or Slack when it should. It installs through a WordPress plugin, Wix, Webflow, Shopify or Google Tag Manager, and it says so when the answer is not in what you added. The second is an agent built on your systems, which is what you need when the answer sits in the ERP, the ticketing system or a customer record rather than in a document.
Automating customer service with AI starts with the twenty questions your team answers most. If they are answered by your content, the product covers them within days. If answering them means reading or writing live data, it is an integration project: a 2,500 EUR audit that measures which share of your tickets AI can take, a 20,000 EUR Production-ready MVP on one channel, production from 50,000 EUR, and about 500 EUR a month of managed AI support afterwards. The cost of AI customer service is therefore either a monthly plan from 39 EUR or a build from 2,500 EUR, and this page's table shows which one your questions point to.
What the agent actually does, and what it refuses to do
The useful part of a support agent is not the answering. It is the not-answering. An agent that guesses confidently is worse than no agent, because a wrong answer in writing is a commitment your customer will hold you to.
- It answers from what you gave it. Website, PDFs, FAQ, catalogue. It does not know anything else, which is the point.
- It says when it does not know, and hands over. To Slack, to email, to your helpdesk, with the conversation attached so nobody starts from the beginning.
- It asks what you would ask. Name, email, phone, company, captured in the conversation rather than in a form nobody fills in, scored hot, warm or cold, exported as CSV.
- It books. Into Google Calendar, at the end of the conversation, while the person is still interested.
On Instagram it does one more thing worth naming, because it is the feature people ask for by name: someone comments your keyword and the link lands in their DMs in seconds, on one post or on all of them, with up to six follow-ups. If you already run ManyChat, the keywords you use today come with you.
It answers in writing, not on the phone; for calls, see AI voice agents for customer service. If you only need the product, start at SimplyBoost. If the chatbot has to run on your own systems and data, see having a custom AI chatbot built.
Connecting it to your ticketing system
In an organisation of a few hundred people or more, the AI is not the support system. Zendesk, Freshdesk, TOPdesk, Salesforce Service Cloud or Dynamics 365 Customer Service is, and it stays the record. The agent answers what it can and, for everything else, creates or updates a ticket with the transcript, the verified customer, the category and the language attached, so routing and reporting keep running where your team already works.
- Your ICT partner or IT department keeps running the ticketing system, the identity provider and the change process. The agent is one more integration in that landscape, released through your own test and acceptance environment.
- Crux Digits builds the agent and the connection to the ticketing system's API, writes the evaluation set and hands over the code and the runbook. We do not run your helpdesk and we do not replace your ICT partner.
- Your support lead owns the knowledge articles the agent quotes, because an agent is only as current as those articles.
The handover field by field, per system: AI customer service and your ticketing system. Preparing the articles: a customer service knowledge base ready for AI.
Automating customer service with AI: what to do first
How many questions can AI handle? Nobody can tell you that from the outside, and a percentage on a vendor's homepage is about their customers, not yours. The number that matters is measurable before you buy anything: take a sample of your own recently closed tickets and mark which ones a knowledge article could already have answered. That share is the ceiling for what an agent answering from content can deflect. The share that needed a lookup in a system is what a built agent could add.
The method and the arithmetic: AI service desk: the number to check first. The audit does exactly this on your own tickets before anyone builds anything.
Where it goes wrong for the customer
The public complaint about AI in customer service is rarely the answer itself. It is the loop: a bot that cannot help and will not let go. Three design rules prevent most of it.
- Say it is AI, at the start. Customers forgive a machine that says what it is, and the AI Act requires it (see the next section).
- A person is one step away. Every conversation offers a handover, and the handover carries the transcript so the customer does not repeat themselves.
- It refuses rather than guesses. A wrong answer in writing is a commitment; “I do not know, a colleague will take this over” is not.
What the AI Act and the GDPR require of a customer service chatbot, including transcript retention: customer service chatbot: what the AI Act and GDPR require.
What changes when your customers are across Europe
Three things, and none of them is the model.
- The AI Act tells them they are talking to a machine. Article 50 transparency applies from 2 August 2026: a system that interacts with people has to make that obvious. In practice this is one line in the opening message, and it costs nothing. Pretending otherwise is the part that costs.
- Language is a support decision, not a translation one. An agent that answers Dutch in Dutch and German in German is straightforward; deciding which languages you will then handle when it hands over to a human is the real commitment.
- GDPR applies to the transcript, not just the answer. A support conversation collects a name, an email and often an order. Decide the retention period before you switch it on, because the honest answer to a data request is the one you can produce.
Where the questions touch a regulated decision rather than an order status, an agent should route to a human by design. See the EU AI Act risk checker for where a given use case lands.
Examples by sector
The clearest case for an agent that reads your systems is after-sales for technical products, where the right answer depends on the model and the serial number: an owner asking about a fault code, an installer needing a setting during a job, the parts counter looking for the right drawing. These are question types, not client cases.
- AI customer service for dealer networks, with pages per trade such as heat pump brands, forklift dealers, machinery dealers and solar and battery brands.
- After-sales AI customer service for technical products: four example exchanges and what the documentation needs.
Belgium: Dutch and French in one service desk
A Flemish organisation with customers in Brussels and Wallonia needs its customer service in Dutch and French, and often in German and English too. One agent can answer each customer in their own language from one knowledge base; the decision that costs something is which languages your team staffs when the agent hands over. Belgian language law leaves the language of customer service free and binds invoices, labels and staff documents. More in multilingual customer service in Belgium with AI.
What each route costs
SimplyBoost publishes three plans at a flat monthly price in euros: Starter at €39 a month (500 AI replies), Growth at €89 (2,000 replies, plus Shopify actions, advanced analytics and the public API) and Pro at €169 (6,000 replies, analytics with export, dedicated support). The website widget, WhatsApp, Instagram, calendar booking, lead capture and handover to your team are in every plan, and yearly billing lowers the monthly price. Current plans are on simplyboost.io/pricing.
A custom agent follows our three fixed steps: an AI audit and strategy at 2,500 EUR that decides whether you need one at all, a Production-ready MVP at 20,000 EUR working against your own systems, and production from 50,000 EUR, integrated, monitored and handed over with the evaluation set and the code. Prices exclude VAT and the audit is credited in full against the MVP if you continue within 60 days.
The gap between €39 a month and 20,000 EUR is the whole reason the first section of this page exists. Most businesses asking for a custom support agent want the product and have not been told it exists.
Frequently asked questions
What is AI customer service?
Is AI customer care the same as AI customer service?
Should we buy SimplyBoost or have an agent built?
What does the agent do when it does not know the answer?
Which channels does it answer on?
Which AI customer service software should a mid-sized company buy?
How do we automate customer service with AI without losing customers?
Chatbot or AI agent: which do we need?
What does AI customer service cost?
Will it hallucinate or invent answers?
How long does it take to go live?
Does it work in more than one language?
Do we have to tell customers they are talking to AI?
Who owns the conversations and the data?
Tell us the twenty questions you answer most
A free 30 minute call. We tell you whether the product covers them or whether it needs building, and what each costs.
Book a free consultation →