ChatBot
Also known as: ChatBot by Text, Text AI Agent
ChatBot is Text's customer support and sales agent offering. New customers configure AI Agent in Text, connecting company knowledge to editable skills that update tickets, customer records and external systems. Website and messaging conversations can trigger actions or transfer to staff with their history. The shared workspace combines AI, live chat and a help desk. Seat subscriptions include a small resolution allowance, with extra capacity sold in packages. Existing customers can continue using the older Visual Builder.
ChatBot is a first party product of Text, Inc., the company behind LiveChat and HelpDesk. The current offering directs new customers to AI Agent in Text, where support conversations, customer records, tickets and automation share a workspace. Existing ChatBot customers can still use the earlier Visual Builder. The current agent answers questions from company content and performs configured tasks during a conversation. Support teams and online stores use it for order inquiries, lead collection, product guidance and escalation. Multiple agents can have separate knowledge, skills and profiles, with the available number determined by the subscription.
Customers describe a skill in ordinary language and review the generated steps before activation. Those steps remain editable and can branch on conditions, collect missing details, update a ticket or send information to connected systems. Integrations reach Salesforce, HubSpot, Shopify, Slack, Klaviyo and Mailchimp, with HTTP requests for other services.
Agent skills start when the conversation matches their purpose; workspace skills also accept schedules, system events and webhooks. Website content refreshes on a chosen interval, while files and written articles supply additional knowledge. A human transfer carries the conversation context, and staff can take over replies. Skill run records connect each execution to its chat and expose the completed steps or failures.
Essential costs $25 per user per month on monthly billing, or $19 with annual billing, and includes one AI agent and ten monthly resolutions. Growth costs $99 monthly or $79 with annual billing and includes ten agents and 200 resolutions. Additional capacity comes in packages of 50 for $49.50. One conversation incurs at most one resolution charge when an answer or completed skill solves a customer request.
Packages expire with the billing cycle; purchased package capacity renews unless changed in subscription settings. The agent stops taking new chats when capacity runs out. A 14 day trial needs no credit card. Enterprise is an annual plan with custom AI built and run by Text's team, custom SAML identity integration and audit logs, alongside the platform's SOC 2 Type II controls and two factor authentication.
The product suits teams combining service automation with human support and commerce workflows. Its preview lets builders inspect replies and proposed skill triggers before publication, though ticket creation and transfers only run for real in production, where run records support troubleshooting. Customer profiles and conversation history provide context, with no separate long term memory managed by the agent. Text manages model operation and cloud hosting in the US and Europe, with no region choice for customers. The agent has no browser or desktop control, no customer model selection and no saved test suite for replaying and scoring support scenarios.
Vendor details
Canonical URL
https://www.chatbot.com
Category
Customer support agent
Subcategory
Text AI Agent with editable support and sales skills
Funding status
A product of Text, Inc., based in Boston.
Company status
first party product
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Skills connect Salesforce, HubSpot, Shopify, Slack, Klaviyo and Mailchimp, plus external HTTP endpoints. Text messaging and configuration APIs extend the same conversation platform; the external MCP tool set reads chats and tickets.
In practice
An order question triggers a connected lookup and an appropriate follow-up action.
A skill collects missing contact details before creating and tagging a support ticket.
A manager opens a failed skill run to inspect its steps and associated chat.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
9.5 / 14 capabilities · 68%
| Integrations & Tool Calling | Full |
|---|---|
|
Connected skills can create Salesforce leads, send Slack messages, update help desk tickets and write customer details, with further built in connections to HubSpot, Shopify, Klaviyo and Mailchimp. Each service connection uses its own credentials or API token, and HTTP request steps reach services outside the catalog. Customers choose which actions an AI agent may invoke during a support conversation, so an order lookup, a discount code or a CRM update all run from the same skill builder. Sourcetext.com/features/skillsread 2026-10-20 |
|
| Workflow Orchestration | Full |
|
A skill starts as a plain language description that the builder turns into editable steps, which the customer reviews before activation. Those workflows support conditional paths, parallel branches, delays and variables passed between actions. An information collection step pauses the conversation until the required details arrive, then the run continues into a ticket or customer record update. Builders can change a generated workflow after activation and republish it, and one AI agent can hold several skills with separate triggers. Sourcetext.com/help/advanced-workflows-usageread 2026-10-20 |
|
| Knowledge Grounding & RAG | Full |
|
The Knowledge Hub accepts public website pages, uploaded files and manually written articles. Website sources refresh on an interval the customer sets anywhere from five to 365 days, and separate agents can draw on different knowledge collections. A knowledge gap view groups the customer questions the agent could not answer, so staff can write the missing answer or article straight from that list and the next conversation on the topic picks it up. Sourcetext.com/help/configuring-your-knowledge-hubread 2026-10-20 |
|
| Human Oversight & Guardrails | Partial |
|
Customers define when the AI agent transfers a conversation to an available person or a particular team. Staff receive the history, can take over replies and can supervise selected chats. When nobody accepts a transfer, a configured fallback can create a ticket after 30 seconds. The operating controls focus on escalation and human participation; administrators do not configure a general approval inbox for reviewing every external action before it executes. Sourcetext.com/help/ai-agentsread 2026-10-20 |
|
| Security, Identity & Governance | Full |
|
Enterprise accounts can connect a custom identity provider through SAML 2.0, and two factor authentication is available on every plan. Workspace permissions limit what each teammate can open or change, and Enterprise audit logs record administrative activity on request. Text maintains SOC 2 Type II controls covering security and availability. Its data processing agreement is built into the main agreement, with European standard contractual clauses for transfers, so no separate signature is needed. Sourcechatbot.com/chatbot-for-enterpriseread 2026-10-20 |
|
| Observability & Auditability | Full |
|
The Runs view lists each skill execution with its date, time and chat identifier, and opening a run shows the steps it completed. Inside the conversation, system events mark when a named skill starts, completes, fails, waits for human feedback or is rejected, and failure events carry the error and its reason. An operator investigating a complaint can open the chat, see which skill fired and find the step where it stopped, alongside the aggregate resolution and satisfaction reports. Sourceplatform.text.com/docs/messaging/references/system-activitiesread 2026-10-20 |
|
| Memory & State Persistence | Partial |
|
The AI agent replies with the conversation history and customer context in view. Skills can inspect the current chat, the visitor's recent page visits and customer attributes, and contact details collected in a chat are saved to the customer profile, so a returning visitor arrives with name and history attached. That continuity lives in the platform's customer, chat and ticket records; the agent has no store of its own for what it learned, with separate retention or deletion settings. Sourcetext.com/help/ai-agentsread 2026-10-20 |
|
| Deployment & Data Residency | Not documented |
|
Text operates the service in its own managed cloud on infrastructure in the United States and Europe, with regional backups and traffic protection. Customers install a widget or connect their messaging channels to that hosted workspace, and no setting lets a customer choose which region holds its data. The current AI Agent offering has no package to install in a private cloud or on premises, and the Enterprise plan's custom AI is built and run by Text's own team. Sourcetext.com/partners/solution/it-system-integratorsread 2026-10-20 |
|
| Prebuilt Agents / Templates / Packs | Full |
|
Ready made skills include Checkout Coupon Issuer, Bundle Discount, Product Insurance Upsell and Consultation Booking, plus several transfer patterns for handing a chat to a person or team. Selecting one inserts an editable prompt into the skill builder; the customer reviews the generated steps and activates the result. Consultation Booking, for example, collects the visitor's details and creates a follow up ticket so a person can arrange the meeting, while the commerce skills issue codes or offers inside the chat. Sourcetext.com/help/custom-skillsread 2026-10-20 |
|
| Triggers & Channel Coverage | Full |
|
AI Agent skills start when a live conversation matches the task they handle, such as an order status request. Workspace skills can also start from chat or ticket events, schedules and webhooks. Website chat and connected messaging channels bring customer requests into the same service workspace. Context such as the visitor's page, location or time can narrow when a skill runs, while routing rules decide which agent or team receives the conversation. Sourcetext.com/features/skillsread 2026-10-20 |
|
| Model Flexibility & Routing | Not documented |
|
Agent settings control instructions, tone, answer length, languages, knowledge and available skills. Text runs the model service behind those settings and does not name the provider; customers cannot choose between model makers or supply their own inference endpoint for the agent. An OpenAI action inside a workflow handles a separate text processing step and leaves the support agent's model unchanged, and an outside assistant connected to Text's MCP server runs its own model. Sourcetext.com/help/ai-agentsread 2026-10-20 |
|
| APIs / SDKs / MCP Extensibility | Full |
|
Text exposes authenticated messaging, customer data, configuration and reporting APIs, with SDKs for custom chat interfaces and backend integrations. AI and human replies share the same conversation format, so integrations can read or send messages and manage the surrounding customer experience. OAuth scopes constrain access to platform resources. A separate hosted MCP endpoint offers scoped chat and ticket retrieval; the broader write operations are available through the platform APIs. Sourceplatform.text.com/docs/getting-startedread 2026-10-20 |
|
| Testing, Debugging & Optimization | Partial |
|
Builders try questions in a preview panel against an agent's configured knowledge and instructions. The panel shows when a skill would trigger, but ticket creation and transfers do not actually run there, so connected actions are checked by running the workflow manually with sample data or by reading live runs afterward. Each test is typed by hand; the product keeps no saved test cases and runs no scored comparison between agent versions. Sourcetext.com/help/ai-agentsread 2026-10-20 |
|
| Browser / Computer-use | Not documented |
|
The agent can use information about the visitor's current page and recent browsing activity to choose a skill or personalize a conversation. Its actions run through Text operations and connected service integrations. The website widget is a messaging interface, and website ingestion supplies knowledge content. Neither gives the agent control of a general browser session for clicking through unrelated applications, entering data into arbitrary screens or operating a desktop. Sourcetext.com/help/custom-skillsread 2026-10-20 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
Text's AI Agent, the product at chatbot.com, can now turn a plain language description of what should happen into a skill, an automation the agent triggers during a conversation or that runs on its own when an event fires or someone hits run. Workflows and custom skills are merged into one Skills feature, shown in a redesigned visual builder where every step stays editable even after activation. Skills can use Salesforce, HubSpot, Klaviyo, Mailchimp, Shopify, Slack and OpenAI, with HTTP requests or webhooks for other tools.
Bears on: Agent capability
View sourcePricing
$25 per user per month; $19 with annual billing
Paid seats plus AI resolution packages
Included quota
Essential includes one AI agent and ten monthly resolutions; Growth includes ten agents and 200 resolutions.
What is public
Essential $25/user/month, Growth $99; annual equivalents $19 and $79. Extra capacity is $49.50 per 50 resolutions.
Billing mechanics
One charge per conversation that solves a question or completes the requested skill. Failed skills and unanswered handoffs are excluded.
Cost watchouts
Essential includes only ten monthly AI resolutions. Extra packages renew unless changed; removing seats does not credit the current term.
Variable cost rationale
Seat count and successfully resolved conversations determine cost; additional channel providers may bill separately.
Additional watchouts
At the limit the AI agent stops taking new chats. Package capacity purchased during a cycle is included in the next renewal unless changed.
Overage / add-ons
50 additional resolutions cost $49.50. Unused units expire at cycle end; package capacity renews unless adjusted. The agent stops taking new chats at its limit.
Sales call required
Mixed (some tiers require a call)
Free / trial
14 day free trial
Lowest paid plan
Essential
Commercial notes
14 day trial without a credit card; no permanent free plan. Enterprise adds SAML and audit logs.
Key ambiguities
Enterprise pricing and service commitments require a quote.
Cancellation / refund
Plan, seat and package reductions take effect at renewal. Removing users during a cycle creates no credit.
Support SLA / resale
Enterprise adds dedicated onboarding, an account manager and negotiated service commitments.
Missing data
Enterprise contract price and workload-specific external channel fees.
Related vendors
- Decagon — AI agent platform purpose-built for customer support, with deep…
- 5.Y — 5
- Ada — Agentic customer experience platform whose AI agents autonomously…
- Aisera — Enterprise agentic AI platform that orchestrates specialized agents…
- ASAPP — Generative AI for enterprise contact centers that plugs into…
- Assembled — AI customer support orchestration platform that unifies autonomous…