AI Chat is the conversational AI workspace inside AIUNIFY PROS SUITE.
Unlike a Standard AI Generator, where you normally enter one Prompt and receive one Output, AI Chat is designed for an ongoing back-and-forth conversation.
The current interface identifies the workspace as:
AI Chat Assistant
with the description:
“Have natural conversations with AI to get answers, brainstorm ideas, or solve problems.”
AI Chat currently supports:
This chapter also explains important current-version limitations, including the absence of a universal saved Chat history, New Chat button, Copy button, Export function, message-editing controls, and persistent conversation storage in the current AI Chat interface.
AI Chat appears under:
Sidebar → Core Tools → AI Chat
The current route is:
AI Chat → /chat.
You can also locate it through:
when AI Chat is available to your account.
The Chat route requires an authenticated user and is wrapped in the PROS SUITE feature-access system.
If you are not authenticated, PROS SUITE redirects you toward Login.
If AI Chat is not included or has been restricted for your account, normal PROS SUITE feature-access rules can apply.
To open AI Chat:
At the top of the workspace, PROS SUITE displays:
AI Chat Assistant
and:
“Have natural conversations with AI to get answers, brainstorm ideas, or solve problems.”
This description accurately summarizes the role of the current Chat workspace.
The primary conversation card is titled:
Chat.
This area contains:
When no messages have been sent, AI Chat displays:
“Start a conversation with the AI assistant!”
and:
“Ask questions, get help, or just chat.”
There is no required setup form before beginning a normal Chat.
At the bottom of Chat is the message field with the placeholder:
Type your message here...
Use this field for your question, instruction, follow-up, or other conversational request.
The Chat system is configured as a general AI assistant.
The current system instructions describe it as a helpful, intelligent conversational assistant designed to provide clear, accurate, engaging responses and to analyze images when provided.
Common uses include:
The most important difference is conversational continuity.
Typically:
Prompt
↓
Generate
↓
Output
↓
Modify Prompt and generate a replacement Output.
Typically:
Message 1
↓
AI Response 1
↓
Message 2
↓
AI Response 2
↓
Message 3
↓
AI Response 3
The previous conversation is passed as history for subsequent Chat requests.
When you send a new message, AI Chat provides the prior messages collection as conversation history to the shared AI system.
This means follow-up questions can build on earlier exchanges within the current Chat session.
You could begin with:
Create three positioning ideas for a B2B software company serving sales organizations.
After the AI responds, you could say:
Make the second idea more disruptive.
Then:
Now rewrite it for a LinkedIn post.
Because earlier Chat messages are supplied as history, the AI can use the preceding conversation as context rather than requiring you to repeat the entire original request each time.
When a Chat request is made:
This is the basis of AI Chat's conversational behavior.
To start:
The user message immediately becomes part of the visible Chat conversation before the AI response is completed.
The circular button beside the message field contains the Send icon.
It submits the current message.
The Send button is disabled when:
or when:
AI Chat supports:
Enter
to send the current message when Shift is not being held.
This differs from the Standard Generator, which uses Ctrl + Enter.
The current behavior is:
Send Message
Allows you to create a new line in the message instead of submitting it.
The Chat handler only submits when Enter is pressed without Shift.
Use Shift + Enter when you need a structured message.
For example:
Create a marketing strategy with these requirements:
Shift + Enter
Audience: insurance agencies
Shift + Enter
Goal: recruit sales professionals
Shift + Enter
Tone: authoritative B2B
Then press Enter normally to send.
The current message field expands within a defined range rather than becoming an unlimited document editor.
For extremely long source material, consider whether a dedicated PROS SUITE Tool is more appropriate.
User messages are visually distinguished from assistant messages.
The current Chat layout aligns user messages toward the right side and uses the User icon.
Assistant messages are aligned toward the left and use the Bot icon.
This visual separation makes longer conversations easier to follow.
Each completed Chat message displays a time beneath the message content.
The timestamp is generated from the message's local Date value.
AI Chat uses the shared streaming system.
Instead of waiting until the entire response is complete, text is progressively added as it arrives.
This allows you to begin reading while the AI is still responding.
While the AI is responding, a temporary assistant message appears in the Chat.
As text arrives, it is rendered progressively through the Markdown Renderer.
Before text or tool calls begin appearing, the temporary response area displays animated dots.
This means AI Chat is still working on the response.
While the current response is streaming:
Allow the response to finish before continuing.
The current AI Chat interface does not expose a:
Stop Generating
button.
Once a Chat response begins, allow it to finish.
This is similar to the current Standard Generator behavior.
After streaming finishes, the final assistant message is added to the conversation history.
The completed response uses:
AI Chat responses are displayed through PROS SUITE's Markdown Renderer.
Depending on the response, formatting can include:
As:
AI Chat automatically attempts to scroll to the bottom of the conversation.
This helps keep the newest response visible.
After an AI response finishes:
Continue for as many exchanges as appropriate for the current session.
One of the main advantages of Chat is that you can refine the result conversationally.
For example:
You:
Create five webinar titles for AIUNIFY.
AI:
Provides five titles.
You:
Make number 3 more enterprise-focused.
AI:
Refines it.
You:
Now create a subtitle and three bullet points.
This is more natural than repeatedly re-entering the entire original request.
Unlike the shared Standard Generator, AI Chat contains an image attachment control.
The Paperclip button opens a file-selection dialog that currently accepts:
image files.
To attach an image:
The file selector currently uses:
accept="image/*".
Therefore, the verified attachment capability in this Chat interface is for images.
This chapter does not claim that the same button accepts general PDFs, Word documents, spreadsheets, or arbitrary files.
AI Chat currently limits an uploaded image to:
5 MB.
Images larger than that are rejected by the current user interface.
If the image exceeds the limit, PROS SUITE displays:
File too large
with:
“Please upload an image smaller than 5MB.”
Choose a smaller image and retry.
After a valid image is selected, a thumbnail preview appears above the message Input.
Use this preview to confirm you selected the intended image.
The image preview includes an X control.
Select the X to remove the attachment before sending.
The image will no longer be included with the message.
You can submit:
Image + Text
together.
For example:
Attach a screenshot and enter:
Explain what this chart is showing and identify the three most important trends.
The text and attached image are sent as part of the AI Chat request.
The current form also allows submission when:
because the submission check permits either meaningful text or an image.
For better results, however, include a clear instruction whenever possible.
Once sent, an attached image appears within the user's message bubble.
This helps preserve the visual context of the current conversation.
After your message is submitted:
The sent image remains visible in the conversation message itself.
The current Chat system instructions explicitly identify multimodal/image-analysis capability and instruct the assistant to analyze provided images.
Potential uses include:
Results still require appropriate review.
The Chat attachment feature should not be confused with a persistent Files workspace.
Current AI Chat:
Attach Image → Ask About Image → Continue Conversation
It does not establish:
AI Chat uses the shared selected Model from the PROS SUITE authentication/AI context.
The shared streaming request sends that Model with the Chat request.
Therefore, the Header AI Model Selector can affect AI Chat.
If you want to use another Model:
The next supported Chat request uses the currently selected Model.
Because the Model is selected globally at request time, you can change Models between Chat turns.
For example:
Messages 1–4
→ Model A
Then select:
Model B
Message 5
→ Model B processes the next supported request while prior conversation messages remain part of the history supplied by Chat.
Selecting another AI Model does not automatically clear the existing Chat messages.
The conversation remains visible in the Chat component's current state.
If you then submit another message, the existing prior-message collection continues to be supplied as history.
If you choose a Model requiring an API Key that is not configured, PROS SUITE can display:
Model Not Available
and direct you to Settings.
Follow Chapters 10 and 11 before continuing Chat with that Model.
The current AI Chat configuration provides the AI with two source-defined utility tools:
These are additional capabilities beyond ordinary conversational text generation.
The calculator utility is described as:
“Perform basic mathematical calculations when requested by the user.”
It accepts a mathematical expression.
For example:
Calculate (1250 × 0.18) + 350.
AI Chat can invoke its calculator utility where appropriate.
The current Chat implementation sanitizes calculator expressions to basic mathematical characters, including:
Therefore, it should be understood as a basic mathematical calculator utility, not a full spreadsheet or advanced symbolic mathematics system.
The second utility is:
get_current_time
described as:
“Get the current time in a specific timezone.”
For example:
What time is it in America/New_York?
The AI can request the current-time utility using the specified timezone.
The current time function uses the requested timezone with the browser/runtime's time formatting.
If the timezone cannot be interpreted, Chat can return an error relating to that timezone.
Use recognizable timezone identifiers when possible.
When the AI makes a utility call, the streaming Chat can display:
Using tools:
followed by the tool Name and its arguments.
This lets users see that a utility is being invoked.
After the response completes, Chat appends information describing the tool used and its result to the final assistant response.
For example, the completed content can identify:
Tool Used: calculator
and:
Result: [calculation]
The current source-defined Chat utility list contains:
It does not establish a general user-facing Web Search or live Internet-browsing tool for AI Chat.
Therefore:
Do not assume AI Chat can automatically retrieve current public web information merely because it is conversational.
Use an appropriate connected/current-information system where one is specifically provided.
The source-defined Chat utilities also do not establish general AI Chat actions for:
Those capabilities should only be used through dedicated PROS SUITE systems or Tools that explicitly provide them.
The current general Chat instruction describes the assistant as:
Use Chat for broad interactive assistance rather than assuming it is a specialized automation Agent for every external business system.
The PROS SUITE feature catalogue currently defines:
AI Chat Assistant
with a source-level Token cost of:
10.
However, there is an important distinction between this feature definition and the visible Chat interface.
Unlike the Standard AI Generator, the current AI Chat screen does not display:
X tokens per use
above the conversation.
The Chat UI also does not display:
Your balance: X tokens
inside the Chat panel itself.
Use Sidebar/Dashboard Token information to monitor your account.
The feature catalogue currently defines Chat with tokenCost: 10.
But the actual backend implementation of /api/ai/stream, where final usage treatment occurs, was intentionally excluded from the sanitized User Operations source bundle.
Therefore, this manual will not claim that every live Chat message always deducts exactly 10 Tokens.
Use the current account balance and any current user-facing pricing/usage information as the operational reference.
The shared streaming workflow calls for a user/account refresh when an AI request finishes.
This allows updated Token information to be reflected in other PROS SUITE areas following supported AI operations.
Because Chat can involve many messages, check:
Sidebar → Monthly Tokens
periodically during longer sessions.
A conversational workflow can involve repeated AI operations rather than one single generation.
The current Chat page initializes:
messages = []
as local component state.
The supplied ChatPage source does not establish loading prior AI Chat conversations from a persistent user history service.
This has an important user-facing consequence.
The current source does not show:
Therefore:
Do not assume AI Chat conversations are permanently saved for later retrieval in the current version.
Because the conversation is maintained in local page state and no persistence mechanism is established in the current Chat source, navigating away or refreshing/recreating the Chat page can cause the current conversation to be lost.
For important Chat results:
The current AI Chat page does not establish a user-facing:
New Chat
button.
Do not look for a ChatGPT-style conversation Sidebar or New Chat control in this current workspace.
The current Chat source also does not provide:
controls.
If you need a completely fresh conversation in the current version, reloading/reopening the page may recreate the empty Chat state—but remember that this also means the current unsaved conversation can be lost.
The current source does not provide a control to:
an AI Chat conversation.
Conversations do not currently appear as named persistent projects.
The current AI Chat interface does not establish a Search field for past Chat sessions.
Global Search searches PROS SUITE Tools, not AI Chat message contents.
The current Chat interface does not provide:
controls.
This differs from the Standard Generator, which provides .txt Download.
The current Chat message rendering does not include a dedicated Copy control for assistant responses.
If you need to preserve a response, use your device/browser's normal text selection and copy functions.
The current Chat interface does not expose a dedicated:
Regenerate Response
button.
If you want another version, send a follow-up such as:
Give me another version with a more concise tone.
or repeat/reframe your request.
After a user message is sent, the current interface does not provide a dedicated Edit button for that sent message.
To correct something, send a follow-up.
Example:
Correction: the target audience is commercial insurance agencies, not retail insurance customers.
The current Chat message display does not provide a per-message Delete action.
Sent messages remain visible for the life of the current page state.
The current source does not establish:
controls.
Use follow-up messages within the same current session or begin a fresh page session.
The current Chat interface does not establish:
controls for individual AI responses.
Ordinary users do not receive a Chat settings field for editing the underlying system instructions.
The general Chat behavior is supplied automatically by PROS SUITE.
Simply describe your desired task and response characteristics in your message.
Even without a system-prompt editor, you can write:
Answer as a professional B2B software strategist.
or:
Keep the response concise and executive-level.
or:
Explain this as if I am new to the subject.
These become part of the current conversational request.
You:
Give me ten webinar themes for enterprise sales organizations adopting agentic AI.
AI Chat responds.
Then:
Rank the top three by executive appeal.
Then:
Develop the strongest one into a 45-minute webinar outline.
This demonstrates conversational refinement.
You:
I have a customer onboarding process with six manual handoffs. Help me identify where AI automation could reduce delays.
After the response:
Focus only on the handoffs between sales and implementation.
Then:
Turn that into a three-phase implementation plan.
Review this dashboard layout. Identify the three biggest usability problems and recommend improvements.You can ask:
Calculate 249 × 75 + 3,500.
When the AI invokes the calculator utility, Chat can display the tool call and resulting calculation.
You can ask:
What time is it in America/New_York?
AI Chat can use its current-time utility for a specified timezone.
You:
Write a mission statement for a commercial software company.
AI responds.
You:
Keep the same idea but make it bolder.
You do not necessarily need to paste the original response again because earlier messages remain part of the current Chat context.
For more complex conversations, include:
What are you trying to accomplish?
What does the AI need to know?
Who is this for?
What boundaries should be followed?
What format do you want?
Example:
I need a partnership outreach strategy for a B2B SaaS company. Target insurance agencies with 25–200 producers. Focus on executive outreach, revenue-sharing positioning, and a 30-day pilot. Give me a five-step strategy plus an introductory email.
AI Chat is especially useful when you expect the work to evolve through discussion.
Examples:
For a single predefined deliverable, a specialized Standard Generator may be more efficient.
Use:
Business Plan Generator
instead of AI Chat when your primary goal is a structured Business Plan.
Use:
Sales Pitch
when you specifically need a Sales Pitch.
Use:
Article Writer
for a formal article.
Use:
Competitor Analysis
for structured competitive analysis.
AI Chat is best when the task benefits from conversation rather than a fixed Tool workflow.
AI Chat and Research Agent are not the same.
AI Chat is the general conversational workspace.
Research Agent is a separate PROS SUITE Tool designated for more specialized research-oriented work.
Do not assume AI Chat has every capability of Research Agent.
Also distinguish:
A personal general-purpose AI assistant inside PROS SUITE.
A separate Core Tool with its own purpose and workflow.
The Sidebar lists both independently.
AI Chat is primarily a typed conversational workspace.
AI Voice Agent is a separate Core Tool.
Do not expect voice calling or voice-agent operations merely because you are using AI Chat.
Check:
The route requires authentication and feature access.
The button remains unavailable if:
Enter a message or attach an image, then wait until any current response finishes.
This is expected.
Use:
Shift + Enter
for a line break.
Use:
Enter
to send.
Check:
The attachment button is disabled while streaming.
If you receive:
File too large
“Please upload an image smaller than 5MB.”
reduce the image size and retry.
The current Chat file selector is configured for:
image/*.
General PDF/document attachment is not established for this current AI Chat composer.
Use another PROS SUITE workflow specifically intended for the file type where available.
While streaming:
Wait until the stream finishes.
If Chat encounters an error, it displays:
Error
with the returned message or:
“Failed to get response. Please try again.”
Read the actual error before retrying.
The shared AI streaming layer can return:
Failed to generate response
when the request fails and no more specific error is available.
Possible issues can involve:
If Chat is configured to use a Model requiring unavailable provider access:
See Chapters 10 and 11.
The current Chat page does not establish persistent Chat-history storage.
Its message collection starts as local page state.
If the page was reloaded or recreated, the prior Chat can be lost.
Preserve important responses outside the temporary Chat session.
There is no source-verified current:
area for this AI Chat workspace.
If prior content was not externally saved, this current source does not establish a recovery workflow.
The current Chat message component does not include Copy controls.
Use your browser/device's normal text-selection and copy method.
The current AI Chat interface does not provide a Chat download action.
Preserve important material by copying it into an appropriate document or other business system.
The current calculator is designed for basic arithmetic expressions.
If the expression contains unsupported content, it can return:
Could not calculate result.
Simplify the calculation or use an appropriate specialized Tool.
Use a valid timezone identifier such as:
UTC
or:
America/New_York
The current utility expects a timezone string.
Because persistent Chat history is not established:
Important Response
↓
Select Text
↓
Copy
↓
Save in the Appropriate Business Record
Do this before:
Conversation context becomes more useful when messages remain related.
For example:
Conversation A
→ Website positioning
rather than mixing:
Website positioning
Recipe ideas
SQL debugging
Hiring interview questions
Contract wording
into one long Chat.
For unrelated tasks, starting with a fresh page state can help keep context cleaner.
If the assistant misunderstands something, say:
Correction: the client is an insurance agency, not an insurance carrier.
Then continue.
This is more effective than assuming the AI will independently detect the misunderstanding.
You can say:
Because the prior messages remain in current-session history, Chat can use that context.
Instead of:
Explain this.
use:
Explain this in five bullet points and end with three recommended next actions.
Clear formatting instructions make conversational output more immediately usable.
Attaching an image without text is supported, but a clear instruction usually improves the result.
For example:
Analyze this screenshot and tell me why the primary call-to-action is difficult to notice.
is more targeted than simply sending the image.
Before attaching an image, check whether it contains:
Crop or redact unnecessary sensitive information before using the image in an AI workflow when appropriate.
The current AI Chat utility list does not establish general web browsing.
If your task requires verified live information, use a PROS SUITE capability specifically designed for that information source where available.
Do not assume conversational confidence means the information was live-retrieved.
AI Chat can help analyze and explain, but consequential outputs should be appropriately reviewed.
Examples include:
AI-generated content should not substitute for appropriate professional or authoritative verification.
Each conversational turn can initiate another AI request.
During extensive Chat sessions:
If the current Model performs well, there is no requirement to change it.
Change Models when:
Remember that Model availability and Token behavior can differ.
The complete user workflow is:
Core Tools
↓
AI Chat
↓
AI Chat Assistant
↓
Optional:
Select AI Model
↓
Optional:
Attach Image
↓
Type your message here...
↓
Enter
or:
Send
↓
User Message Appears
↓
AI Streams Response
↓
Optional:
Calculator / Current-Time Tool Call
↓
Completed Assistant Message
↓
Review
↓
Ask Follow-Up
↓
Continue Conversation
↓
Preserve Important Results Before Leaving
A powerful Chat pattern is:
ASK
Initial question
↓
CLARIFY
Provide missing context
↓
NARROW
Focus on strongest direction
↓
REFINE
Adjust tone, structure, audience, constraints
↓
FINALIZE
Request final usable version
↓
PRESERVE
Copy final content into your business workflow
You should now be able to:
The source verifies the following user-facing Chat capabilities:
Conversation
Input
Multimodal
AI Response
Utilities
The current source does not establish user-facing controls for:
These capabilities should not be included in the user manual as though they currently exist.
The current workspace can be understood as:
AI CHAT ASSISTANT
↓
OPTIONAL MODEL SELECTION
↓
MESSAGE
Text
and/or:
Image under 5 MB
↓
SEND
Enter or Send button
↓
AI STREAMING RESPONSE
↓
Optional:
CALCULATOR
or:
CURRENT TIME
↓
COMPLETED RESPONSE
↓
FOLLOW-UP MESSAGE
↓
CONVERSATION CONTEXT CONTINUES
↓
Before leaving:
PRESERVE IMPORTANT CONTENT
AI Chat Assistant is AIUNIFY PROS SUITE's general conversational AI workspace.
It is described in the current interface as:
“Have natural conversations with AI to get answers, brainstorm ideas, or solve problems.”
Users enter messages through:
Type your message here...
and can use:
Enter → Send
or:
Shift + Enter → New Line.
AI Chat carries the previous message collection into subsequent AI requests, allowing users to build a conversation through follow-up questions.
Responses stream progressively into the Chat and are rendered with Markdown formatting.
The current workspace also supports image attachments. Images must be under 5 MB, and the verified selector accepts image/*.
AI Chat currently exposes two utility capabilities:
The current feature catalogue defines AI Chat Assistant with a source-level Token cost of 10, but the Chat workspace itself does not display a per-message cost or Token Balance, and the backend streaming implementation responsible for final usage treatment is outside the sanitized User Operations source bundle. For that reason, this manual does not claim that every live Chat message will always cost exactly 10 Tokens.
Most importantly, the current Chat messages are maintained as page-level conversation state, and the supplied source does not establish a persistent Saved Chat system. There is currently no verified New Chat, Clear Chat, Saved Chats, Copy Response, Download Conversation, Edit Message, Delete Message, or Regenerate Response control.
The recommended operating principle is:
Ask → Review → Follow Up → Refine → Finalize → Preserve Important Content Before Leaving