Quickchat AI logo
Customer Service · Agents Orchestration Frameworks

Quickchat AI

Quickchat AI builds custom AI Agents for customer support, sales and ecommerce teams. Each agent answers from your own knowledge base, triggers real API actions, and deploys across your website, WhatsApp, Slack, Intercom, Zendesk and Shopify.

Active GDPR compliant Free plan Freemium API available 16+ Verified by Guidaio
Overview

What is Quickchat AI?

Quickchat AI is a platform for building custom AI Agents that handle customer support, sales and ecommerce conversations. It is published by Incentivai Inc., a company founded in 2018 and backed by Y Combinator, co-founded by Dominik Posmyk (CEO) and Piotr Grudzien (CTO). The positioning is stated plainly on the home page: agents that resolve conversations rather than merely reply to them.

The product is organised around six modules. AI Identity sets the agent's personality, tone, job and guardrails without code. Knowledge Base ingests plain text, websites, help centre articles, Intercom articles, YouTube videos, a Shopify catalogue, and files in PDF, DOCX, XLSX, CSV, EPUB, MD and TXT, with automatic URL refresh from the Starter plan upwards. AI Actions connect the agent to HTTP APIs and MCP tools, with JSONPath response mapping and a confidence threshold. Go Live publishes the same agent to a website widget, a hosted chat page, an iframe or an external channel. Inbox lets a human take over, assign, close, summarise and translate. Insights surfaces topics, sentiment, anomalies and a Content Gap Analyzer.

Answers are grounded by a proprietary RAG pipeline with reranking. A source-constrained mode can restrict the agent to connected sources only, so that it asks a clarifying question, escalates, or replies that no answer was found rather than improvising. The Why AI Said That view exposes the sources, reasoning steps, guidelines and API calls behind every response. GPT-5.1 is the standard model on all plans; more advanced models are available on paid plans at a cost of one, two or four credits per message. A Quickchat MCP server lets you drive your agent from Claude or ChatGPT.

Enterprises choose between standard cloud, dedicated private cloud, a self-hosted LLM, on-premises and hybrid deployment. The headline figures — more than 80% of conversations resolved, USD 0.50 per resolution — are published and measured by the vendor itself. On certifications the site is explicit: ISO and SOC 2 are listed as in progress, not obtained, while GDPR readiness is presented as EU-ready.

What it does

  • Build an AI Agent from your own documents, files, URLs and product catalogue
  • Resolve support, sales and ecommerce conversations automatically, in several languages
  • Trigger real actions through HTTP APIs, MCP tools, Shopify, Stripe or Zendesk
  • Deploy one agent across website, WhatsApp, Slack, Discord, Telegram, Messenger, Instagram, Intercom, HubSpot and Zendesk
  • Escalate to a human with the full conversation context attached
  • Audit any answer — sources, reasoning, guidelines and actions — through the "Why AI Said That" view
  • Analyse sentiment, topics, resolution rate and gaps in the knowledge base
Audience

When to use Quickchat AI / When not to

A quick filter to help you decide if Quickchat AI is the right fit.

When to use Quickchat AI

  • Support teams buried under repetitive conversations — order status, refunds, opening hours, password resets — that want deflection without leaving their existing helpdesk.
  • Shopify merchants who need product recommendations, order tracking and cart recovery handled in the storefront chat, with the catalogue detected straight from the store URL.
  • Sales teams that want inbound conversations qualified, routed and enriched with context before a human picks them up.
  • Internal IT and HR helpdesks answering the same questions about leave, onboarding and company policy all year round.
  • Compliance-minded organisations that need EU data residency, role-based access control, audit logging and a signable DPA before deploying an agent.

When not to use Quickchat AI

  • Individuals looking for a personal chatbot: this is a B2B platform for deploying agents on a company's own channels.
  • Teams that need to work from a phone app — there is no iOS or Android application, everything runs in the browser.
  • Developers on a small budget who need programmatic access: API access and conversation export only unlock on the Business plan at USD 999 per month.
  • Buyers whose procurement requires a signed ISO or SOC 2 certificate today, since both are advertised as in progress rather than obtained.
  • High-volume or spiky workloads on a low tier: AI credits are capped per plan and the agent stops replying once the monthly quota is spent.
Get started

How to use Quickchat AI

A typical end-to-end flow, from setup to results.

  1. Try the free Shopify demo from the dedicated page first if you run a store — no sign-up and no credit card
  2. Create a free account on app.quickchat.ai/register, again without a credit card
  3. Describe the AI: personality, tone, guidelines, system prompt and model
  4. Add the knowledge: upload documents and files, or paste URLs that refresh automatically; for Shopify, the store URL alone is enough to detect the catalogue
  5. Configure the AI Actions: connect Stripe, Shopify, an MCP tool or your own API, and map the responses with JSONPath
  6. Go live: install the chat widget, publish the public chat page, embed an iframe, or connect an external channel
  7. Work the Inbox: take conversations over manually, assign them, close them, use the AI summary and translation
  8. Read the Insights: topics, sentiment, resolution rate and content gaps
  9. Expect a minimal deployment — knowledge base crawl plus a widget on one site — to be live the same day
  10. Allow up to two weeks for a production rollout with custom AI Actions and helpdesk handover; Enterprise customers are guided through a three-step Agent Design Framework: define the need, identify the data sources, choose the technical foundation
Quick read

Pros & Cons

Pros

  • Pricing is published in full, from a free plan up to USD 999 per month, plus a per-resolution Enterprise rate — no quote needed to know what it costs
  • The free plan is permanent rather than a countdown: no credit card, unlimited AI Actions and all integrations included
  • Answers are traceable: the "Why AI Said That" view shows the sources, the reasoning steps, the guidelines and the API actions used
  • The publisher states it never trains language models on customer knowledge bases, chat logs or proprietary data, with logical isolation per tenant
  • Customer data is stored on Google Cloud Platform in Belgium, inside the European Union
  • One agent configuration covers a wide range of channels, and a minimal deployment can be live the same day
  • Deployment options stretch to dedicated private cloud, self-hosted LLM and on-premises, and the legal material — terms, privacy, GDPR statement, DPA, legal and security FAQ — is public and unusually detailed

Cons

  • ISO and SOC 2 were still shown as in progress on 18 August 2026: neither certification is held
  • API access and conversation export are locked to the Business plan at USD 999 per month
  • AI credits are capped per plan and the agent stops replying once the monthly quota is spent
  • Only one user licence is included up to and including the Essential plan at USD 99 per month, and the Enterprise rate requires a floor of 2,000 resolutions per month
  • There is no mobile application, on iOS or on Android
  • There is no training opt-out to switch on, because the publisher states it never trains on customer data — the customer relies on that declaration; subscriptions are also non-refundable except where the law requires otherwise
  • No Article 27 EU representative is named even though the controller is established in Delaware, and the competitive comparisons are published by the vendor itself, measured as of June 2026
Pricing

Pricing & Plans

Quickchat AI offers a permanent free plan at USD 0 per month, which requires no credit card and includes one licence, 50 AI credits per month, 50 knowledge base articles, standard models, all integrations and unlimited AI Actions. The lowest paid entry point is the Starter plan at USD 9 per month. Billing is monthly or annual, an annual subscription being invoiced for ten months, which the publisher presents as two months free, or a 17% discount. Enterprise contracts are priced per resolution from USD 0.50, with a minimum of 2,000 resolutions per month; only conversations the agent resolves without a human handover are billed. Payment is processed by Stripe and accepts card, Apple Pay and Google Pay. Plans can be changed at any time, and subscriptions are non-refundable unless the law requires otherwise.

Free — USD 0 per month
  • 1 licence
  • 50 AI credits per month
  • 50 knowledge base articles
  • standard models
  • all integrations
  • unlimited AI Actions
Basic — USD 29 per month (USD 290 per year)
  • 500 AI credits per month
  • 300 knowledge base articles
Essential — USD 99 per month (USD 990 per year)
  • 3
  • 000 AI credits per month
  • 500 knowledge base articles
Professional — USD 299 per month (USD 2,990 per year)
  • 3 licences
  • 9
  • 000 AI credits per month
  • 1
  • 000 knowledge base articles
  • Quickchat AI branding removed
Business — USD 999 per month (USD 9,990 per year)
  • 10 licences
  • 30
  • 000 AI credits per month
  • 2
  • 000 knowledge base articles
  • API access
  • conversation export
Enterprise — from USD 0.50 per resolution, minimum 2,000 resolutions per month
  • custom licences
  • agent development
  • custom AI Actions and data sources
  • unlimited testing
  • custom reporting
  • negotiated DPA and SLA
  • dedicated account manager and priority support
Special offers — 17% discount on annual billing, presented as two months free, with advertised savings of USD 18, 58, 198, 598 and 1,998 depending on the plan · Permanent free plan, with no credit card required to start · Free Shopify demo available from the dedicated page, without signing up · Volume discounts mentioned for Enterprise customers
Prices and plans listed above may evolve. Always check the official pricing page before subscribing.
Trust & Privacy

Data, GDPR & hosting

A consolidated view of how Quickchat AI handles your data.

GDPR overview

GDPR compliance is claimed explicitly, in writing, in the legal and security FAQ and on the vendor's comparison page, and it rests on published documents: a Privacy Policy last modified on 14 August 2026, Terms last modified on 23 January 2026, a dedicated GDPR statement, and a signable DPA governed by Polish law. Data subject rights are listed in full — access, rectification, erasure, restriction, objection, portability, withdrawal of consent, and complaint to a supervisory authority — and are exercised by writing to contact@quickchat.ai; CCPA rights are covered for California residents. International transfers rely on the European Commission's standard contractual clauses (Article 46(2)(c)) together with the EU-US Data Privacy Framework, alignment with which the publisher describes as still in progress. Incidents are notified to the controller within 24 hours. No Article 27 EU representative and no DPO are named anywhere on the site.

Who owns the data?

The customer owns the data. Quickchat AI's legal FAQ states that customers retain full ownership and intellectual property rights over all uploaded content, data, documents and knowledge bases, and that its contracts explicitly say the company claims no ownership over customer data. Quickchat AI, operated by Incentivai Inc., acts only as a data processor where applicable; for the website's own personal data, Incentivai Inc. is the controller. Knowledge bases are logically isolated per tenant, so nothing is shared between accounts. Internal access is limited to what is necessary, governed by role-based access control and audit logging, and production data is reached only with the customer's explicit authorisation.

Reuse rights

Because customers keep ownership of everything they upload, they can reuse their own content, knowledge bases and conversation records freely, without asking Quickchat AI for permission. On the publisher's side, the stated rule is that no customer knowledge base content, chat log or proprietary data is used to train any Large Language Model. Third-party LLM APIs — OpenAI, Anthropic, Gemini, Groq — are called strictly for inference, and the publisher states that the data sent to them is not used to train or improve those models; enterprises may instead self-host an open-source model such as Llama, DeepSeek or gpt-oss on their own infrastructure. On request, conversations can be scrubbed of personal data before any LLM processing. Named subprocessors include Google Cloud Platform, Google Analytics, Datadog, PostHog, HubSpot, and Stripe for payments. Data is encrypted in transit with TLS and at rest. Personal data collected through the website is used to deliver the service, provide support, send notifications, analyse usage and, with consent, for marketing.

Data retention & training

Retention summary
Quickchat AI states that data is retained only for as long as needed to provide the service or to meet legal obligations. In practice, retention does not exceed the period covered by consent, the duration of the contract, or the applicable limitation periods, and longer storage is possible where tax, accounting or anti-money-laundering rules require it. Usage data is kept for a shorter time, unless it serves security or service improvement. Customers can ask support to delete their account and content at any time, which triggers secure erasure procedures. Data from Meta channels such as WhatsApp and Messenger is kept only as long as necessary, and disconnecting a Meta channel starts the deletion process. When a contract ends, the DPA requires the processor to delete its copies of personal data unless a regulation says otherwise.
Trains on customer data
No
Subprocessors disclosed
Yes
DPA available
Yes
GDPR contact

Hosting summary

All customer data is stored on Google Cloud Platform infrastructure located in Belgium, inside the European Union, according to the publisher's legal and security FAQ. Knowledge bases are logically separated per customer, so no tenant can reach another's content, and data is encrypted both in transit, over TLS, and at rest. Enterprise customers are not tied to that default: the GDPR statement allows them to choose the geographic location of a dedicated server, and the deployment models on offer run from standard cloud to dedicated private cloud, self-hosted LLM, on-premises and hybrid architectures. Infrastructure is monitored continuously through Datadog and Google Cloud, and a public status page is published at status.quickchat.ai. Because the controller, Incentivai Inc., is established in Delaware, transfers outside the European Union do take place; they are framed by the European Commission's standard contractual clauses and by the EU-US Data Privacy Framework, alignment with which the publisher describes as still under way.

Hosting countries
🇧🇪 Belgium
Hosting regions
EU
Watch-outs

Things to keep in mind

Risks and trade-offs to weigh before adopting Quickchat AI.

  • AI credits burn at different speeds depending on the model: one, two or four per message, so an advanced model empties the quota four times faster than the standard one — and the agent simply stops replying once the credits are gone. Size the plan for a spike, not for an average month.
  • Handing first-line conversations to an agent can dull a team's own feel for its customers. The Insights and the Content Gap Analyzer only help if someone actually reads them; otherwise the gaps the agent papers over are never fixed.
  • ISO and SOC 2 are announced as in progress, not obtained. They should never be presented internally, or to your own customers, as achieved certifications.
  • The cost and resolution-rate comparisons against Intercom Fin, Agentforce, Sierra and Zendesk are published by the vendor itself and measured as of June 2026. Treat them as marketing claims and test them on your own traffic.
  • Unless you opt for a self-hosted deployment, third-party LLMs (OpenAI, Anthropic, Gemini, Groq) receive your conversation data for inference. PII scrubbing exists but has to be requested.
  • The data controller is established in Delaware and no Article 27 EU representative is named, even though hosting is in Belgium. Worth putting in front of your own legal team before signing.
  • API access exists only from USD 999 per month, and subscriptions are non-refundable except where the law requires otherwise, so an annual commitment made on the strength of a demo is hard to undo.
Setup

Setup & Integrations

Technical difficulty

Low for the standard path. No development skills are needed: personality, guardrails and knowledge sources are configured in the interface, and going live means dropping a widget on your site, publishing a hosted chat page or embedding an iframe. Connecting Shopify takes only the store URL. AI Actions are built from a prompt, with JSONPath response mapping for advanced cases. A minimal deployment can be live the same day; a production rollout with custom actions and helpdesk handover usually takes up to two weeks. Advanced setups — in-house APIs, MCP, self-hosted LLM, on-premises — need a technical team.

Deployment

Web appAPIShopify appSlack app

Integrations

Shopify Slack WhatsApp Discord Telegram Facebook Messenger Instagram Intercom HubSpot Zendesk WordPress Drupal Freshdesk Stripe YouTube Claude ChatGPT

Supported languages

EnglishGermanSpanishFrenchJapanesePolishPortugueseSwedish
Company

Behind Quickchat AI

Company name
Incentivai Inc.
Founded
01/01/2018
Country of origin
🇺🇸 United States
Headquarters
251 Little Falls Drive, Wilmington, New Castle, DE 19808, United States
UBO
INFORMATION_NOT_FOUND
UBO country
INFORMATION_NOT_FOUND
Domain registrar country
🇺🇸 United States
Legal contact
Support contact

Fundraising

Backed by Y Combinator: the mention and logo appear in the footer of every page, next to a "Powered by" credit for OpenAI and Google Cloud.
No funding round, amount or date is published anywhere on the site.

Social

Official links

Resources

All the official URLs gathered for verification and reference.

Compare

Alternatives

Tools that compete with or complement Quickchat AI.

I Intercom FinZ Zendesk AI AgentsA AgentforceS SierraH HubSpot Breeze
FAQ

Frequently asked questions

What is Quickchat AI?
It is a platform for building custom AI Agents for customer support, sales and ecommerce. The agents answer from your own documents, files and product catalogue, can trigger actions through APIs, and are deployed on your website, helpdesk and messaging channels.
How much does Quickchat AI cost?
There is a permanent free plan, then paid tiers at USD 9, 29, 99, 299 and 999 per month. Enterprise contracts start at USD 0.50 per resolution with a minimum of 2,000 resolutions per month; a resolution combines confirmed ones, where the customer validates the answer, and assumed ones, where the customer leaves without asking for a human. A simple greeting does not count.
Which channels can a Quickchat AI agent be deployed on?
The web — chat widget, iframe or hosted chat page — plus WhatsApp, Slack, Discord, Telegram, Facebook Messenger, Instagram, Intercom, HubSpot, Zendesk, WordPress, Drupal and Shopify.
Is my data used to train AI models?
No. The publisher states that customer knowledge base content, chat logs and proprietary data are never used to train Large Language Models. Third-party LLM APIs are called for inference only, and the data sent to them is not used to improve those models.
Where is customer data hosted?
On Google Cloud Platform infrastructure located in Belgium, in Europe. Knowledge bases are logically separated per customer, and data is encrypted in transit and at rest.
Is a Data Processing Agreement available?
Yes. A DPA is published at quickchat.ai/dpa and is required whenever personal data may be processed on the customer's behalf. It designates Incentivai Inc. as the processor and is governed by Polish law.
Is there an API?
Yes, and it is documented on docs.quickchat.ai. API access is reserved for the Business plan at USD 999 per month, as is conversation export.
How long does it take to deploy an agent?
A simple widget backed by a knowledge base crawl can go live the same day. A production deployment with custom AI Actions and helpdesk handover typically takes up to two weeks.
What happens when my AI credits run out?
The agent stops replying and shows an unavailability message until the quota resets or the plan is upgraded. Email alerts are sent at 75%, 90% and 100% of the quota. Credit consumption depends on the model chosen: one, two or four credits per message.
Is Quickchat AI SOC 2 certified?
No. As of 18 August 2026 the site shows SOC 2 as in progress, along with ISO. GDPR readiness is presented as achieved, but neither security certification has been obtained.
Conclusion

Should you pick Quickchat AI?

Quickchat AI is not a newcomer. Incentivai Inc. has been building conversational AI since 2018, with Y Combinator backing, in a market now crowded with helpdesk vendors shipping agents of their own. What the platform puts forward to stand apart is less spectacular than most claims in this category, and on paper more verifiable: every answer can be audited down to its sources, reasoning steps and API calls; pricing is published in full, from a permanent free plan to USD 999 per month and a per-resolution Enterprise rate; customer data sits on Google Cloud in Belgium; and the publisher states it never trains language models on customer content.

The reservations are just as concrete. ISO and SOC 2 appear as in progress, so no certificate is held today, and a procurement team that needs one will have to wait. API access and conversation export unlock only on the most expensive published plan. The performance figures that carry the marketing — more than 80% resolution, USD 0.50 per resolution, 60% fewer new tickets — are the vendor's own measurements rather than audited third-party results, and are best read as claims to test, not as benchmarks.

The natural fit is a support or ecommerce team that wants to automate repetitive conversations without migrating away from its existing helpdesk, that cares about EU data residency, and that can start on the free plan to judge grounding quality against its own documents before committing to anything. Teams that need a mobile app, an audited certification or programmatic access on a modest budget will find the fit tighter.