Frequently Asked Questions

Find answers to the most common questions about the Algho platform.

General

What is Algho?

Algho is an artificial intelligence platform for creating and managing conversational virtual assistants, designed to improve customer experience and optimize support processes.

Which channels does Algho support?

Algho can be integrated on websites, mobile apps, WhatsApp, Facebook Messenger, Telegram and other messaging channels. Configuration varies depending on the chosen channel.

How do I get started with Algho?

To get started, log into the Algho platform and create your first assistant from the "Assistant Settings" section. The User Guide will guide you step by step.

Creating an assistant

What are the available methods to create an assistant?

There are 3 methods available: Wizard (quick creation with basic settings), Expert (full configuration of all settings) and Import file (from a .zip file of a previously exported assistant). See Creating an assistant.

What is configured in Wizard mode?

In the first step you enter the assistant's name, language, gender and voice. In the second step the icon, in the third step you customize some chat interface elements, in the fourth and fifth steps (only for accounts with Digital Human) you select the 3D model and background. See Creating an assistant#Wizard.

Which settings cannot be changed after Wizard creation?

All settings applied during creation can be modified later (see Assistant Settings), with the exception of the Assistant Language, which is defined during creation and cannot be changed afterwards. See Creating an assistant#Wizard.

What does Expert mode consist of?

"Expert" mode allows you to create an assistant by selecting all available settings at once in a single screen, without guided steps. See Creating an assistant#Expert.

How do I import an assistant from a file?

In "Import file" mode, simply select the .zip file containing the previously exported assistant and assign a name to the new assistant. See Creating an assistant#Import-file.

What is the purpose of the fourth and fifth Wizard steps?

The fourth step requires selecting the Digital Human and the fifth step the background. These steps are only required for Algho accounts that include the use of Digital Human. See Creating an assistant#Wizard.

Knowledge Base

How do I create a new intent?

From the "Knowledge Base" section, click the "Create" button at the bottom right. Enter a required name and an optional category. You can also set the "Exclude if" and "Use if" fields to control when the intent should be considered. See the intent creation guide.

How do I link forms and suggestions to an answer?

When creating or editing an answer, you can link an existing form (same assistant language) and one or more suggestions. Suggestions can be automatic (chosen by the assistant among QAs with the appropriate option) or user-defined (selected manually from the list). See Forms and suggestions.

What are synonyms and how do I configure them?

Synonyms allow the assistant to recognize different words with the same meaning. They are configured in the "Synonyms" section of the knowledge base by creating groups of equivalent words (e.g. "send", "ship", "forward"). They can be imported and exported in Excel format. See the Synonyms section.

How does assistant testing work?

You can test the assistant at any time by clicking the assistant icon at the top right of the platform. A test chat will open where you can simulate conversations. Interactions in this chat do not affect assistant statistics. See Testing.

Can I merge multiple intents into one?

Yes, you can merge intents with the same answers in two ways: automatic merge (merges all questions with the same answers into a single intent) or manual merge (you select the intents to merge and the destination intent). The operation is not reversible. See Intent management.

What are courtesy replies and engagement messages?

Courtesy replies are shown when the assistant has no answer in the knowledge base. Engagement messages are sent after a period of user inactivity to reactivate the conversation. Both are customizable with targeting, multimedia elements and suggestions. See Courtesy replies and Engagement messages.

How does Generative AI for the knowledge base work?

Gen AI allows the assistant to use generative artificial intelligence to: select the best answer from the knowledge base, rephrase questions based on conversation context, generate completely new answers for questions not in the KB, and start forms or intents after a generated response. See Generative AI for the Knowledge Base.

How do I move questions between intents?

From the questions list of an intent, select one or more questions and click the "Move" button. On the next screen select the destination intent. If the moved question was the only one in the source intent, the latter will be automatically deleted. See Intent management.

How do filters and sorting work in the knowledge base?

You can filter intents by type (Created/Base), by features (with form, with suggestion, with expiration, with activation, multimedia) and by category. Sorting can be alphabetical, by insertion date or by modification date. See Knowledge Base#Filters.

Assistant Settings

What comprehension parameters can I adjust?

In the assistant settings you can adjust three parameters: Confidence (accuracy vs. ambiguity risk), Flexibility (generalization vs. ambiguity risk) and Matching (syntactic vs. semantic matching). These parameters affect how the assistant interprets user requests. Learn more in Parameters.

How does assistant self-learning work?

The assistant can automatically learn from positive and negative user ratings. You can enable self-learning in the settings, choosing whether to learn from positive ratings, negative ratings (with suggestions) or from operators. Disambiguation self-learning is also available, which automatically adds a QA to the knowledge base after N choices. See Reinforcement.

Can I customize welcome messages and other replies?

Yes, from the "General Assistant Settings" section you can customize: welcome message (first interaction), return message (returning users), disambiguation message and disengagement message (Kiosk only). You can also set an initial video for Kiosk mode. See General settings.

What are reply targeting conditions?

Targeting allows showing a reply only under certain conditions: by user emotion or attention (requires emotional recognition package), by source (interaction channel), by form (conditions based on completed fields) or by context (navigation conditions). See Targeting.

How does the Digital Human Editor work?

The Digital Human Editor allows you to configure the appearance, positioning, lighting and animations of the virtual avatar. It includes settings for: camera position and zoom (Avatar, Camera, FOV), rotation presets, HDR tone mapping, lighting configuration, shadows, LUT for color correction, and breathing, smile, speech and idle animations. See Digital Human Editor.

How do I configure advertising banners in Kiosk mode?

In the "Banner" section of the settings (available only for accounts with the Kiosk feature) you can enable and set the display of advertising banners in the Kiosk interface, defining the dimensions in pixels and the transition time between banners. See Banners.

How do automatic commands after N interactions work?

In the "Commands" section you can enable automatic command activation after a certain number of messages, identical answers or courtesy replies. Available commands are: OPERATOR REQUEST, STOP CONVERSATION and NARRATION. See Commands.

How do I customize the chat interface?

From the "Interface" section you can customize the company logo, the assistant icon, chat colors (header text, background, interactive elements), the "powered by" text and response speed (slow, medium, fast, dynamic). See Interface.

Live Chat

How do I create operator groups?

Operator groups are created from the "Operators" section on the left. After naming the group, you can enable "Priority management" to set the notification order of operators, select the group operators, choose which assistants to manage and set a conversation filter via regular expression. See Live Chat#Operator-groups.

How does operator priority management work?

With priority management enabled, Live Chat is signaled first to the operator with the highest priority. After a customizable number of seconds (default 10s), if the operator does not respond, the chat is also signaled to subsequent operators. Without priority, all operators are in a single list. See Live Chat#Operator-groups.

How can an operator take over a conversation?

The operator can view active conversations (connected users) and open conversations (taken over). By clicking the play icon, the operator can enter the conversation even if not expressly requested. Conversations can be locked with the padlock icon to prevent other operators from intervening. See Live Chat#Operator-Interface.

How does operator handover work during a Live Chat?

The operator can pass the conversation to a colleague by clicking the operator handover icon, selecting the colleague and sending an optional request message. The colleague receives a notification and can accept or decline. The operator who requested the handover receives a notification with the outcome. See Live Chat#Operator-handover.

How can an operator add questions to the knowledge base?

During a Live Chat, the operator can check a user's question and the related answer provided, and click the "+" at the bottom right to add them to the assistant's knowledge base. See Live Chat#Managing-a-conversation.

How does automatic operator call work?

In the assistant settings you can enable automatic operator call after a certain consecutive number of courtesy replies. In this case, the Live Chat request notification takes a warning form, and the user is asked whether they want to accept or decline the transfer to an operator. See Assistant settings#Live-Chat.

Which Live Chat messages can be customized?

You can customize: operator connection message, operator disconnection message, operators unavailable message, operator connection request message and operator connection refusal message. See Assistant settings#Live-Chat.

Conversational Forms

How do I create a conversational form in Algho?

From the "Forms" section in the main menu, click "Add". Configure general settings (name, category, language, Gen AI prompt, welcome, failure and thank you messages), privacy consent options, Gen AI and auto-fill settings, knowledge base search, skip/interrupt texts, email/web service settings and the final command. See Conversational Forms#General-settings.

How does Generative AI work in forms?

Gen AI operates on two levels: 1st level — analyzes the user's initial question extracting information to automatically fill form fields; 2nd level — analyzes the user's answers to individual questions to validate them. Both levels can be active simultaneously. See Conversational Forms#Generative-AI-in-forms.

What are nested forms and how do I configure them?

Nested forms allow you to create different question flows based on user answers. You create a "parent" form with admissible values, then create separate ("nested") forms and link them to individual admissible values. At the end of each nested form, the flow always returns to the parent form. See Conversational Forms#Nested-forms.

How do I trigger a form through the knowledge base?

Enter the knowledge base and edit the desired intent. Scroll to the "Select which form to link" option, click the arrow and select the form. Save: next to the intent a green chat bubble will appear indicating the form has been linked. See Conversational Forms#Via-Knowledge-Base.

How does automatic form field filling work?

The assistant can automatically extract information from the user's initial question to fill form fields. Auto-filling after each field is also possible (extracts information from the answer to one field to fill subsequent ones) as well as auto-filling fields already completed in the past in different forms. See Conversational Forms#AI-Gen-and-auto-fill-options.

How do I configure admissible values for a field?

After creating a field, from the summary screen click the "Admissible values" icon. Add values with a name (visible to the user) and optional aliases (synonyms to catch different terms). You can configure display mode (HIDE, SCROLLABLE, FULL SCREEN) and matching options (semantic, with synonyms, ignore semantics). See Conversational Forms#Admissible-values.

Which commands can be executed after a form is completed?

Available commands are: OPERATOR REQUEST (starts Live Chat), START NEW CONVERSATION (reloads the assistant), RESTART NEW CONVERSATION (reloads keeping the same user ID) and STANDBY (interrupts the conversation). See Conversational Forms#Command.

How do I configure attempts and error messages for a field?

For each field you can set: success text (after a valid answer), error text (if the answer is not validated) and form interrupt attempts (from 0 to 99, default 3). After the maximum number of attempts, the form stops and the failure message is shown. See Conversational Forms#Messages-and-attempts.

Statistics

What metrics are available in the Statistics section?

The Statistics section offers: number of conversations and messages, unanswered questions, disambiguations, average conversation depth, average/total duration, response speed, trend over time, response type (correct, courtesy, disambiguation), interaction type (form, engagement, questions), unique users per channel, completed/failed forms, input type and openings per domain. See Statistics.

How can I filter statistical data?

You can filter statistics by date range, by specific assistant and by status (e.g. active conversations). Conversation trend charts can be viewed by hour or by day, and response type can be filtered by hour, day or totals. See Statistics#Conversations.

What does "response speed" indicate in statistics?

Response speed is a metric that measures the time taken by the assistant to respond to user requests. It is calculated at the conversation level and represents the overall efficiency of the assistant in processing questions. See Statistics#Conversations.

Which in-depth sections are available for each assistant?

For each assistant you can access: Conversations (performance in user interactions), Emotions (statistics on recognized emotions), Attention level (statistics on attention level) and Questions (ranking of most frequent requests). The Emotions and Attention sections are only available for enabled accounts. See Statistics#Assistants.

What do the Interaction type and Input type sections show?

Interaction type describes the different forms of interaction (form, engagement, questions). Input type catalogs the types of input received (keyboard, click, engagement, voice input). Unknown input refers to Facebook, Whatsapp and Telegram channels. See Statistics#Interaction-type and Statistics#Input-type.

What metrics are available for Live Chat in statistics?

The LiveChat Messages section contains statistics on messages exchanged in Live Chat mode (sent and received). The LiveChat Actions section reports actions taken during Live Chat sessions. See Statistics#LiveChat-Messages and Statistics#LiveChat-Actions.

What are Openings by domain and by type?

Openings by domain lists new conversation openings categorized by source domain (e.g. company website, app). Openings by type breaks down openings by type: text conversation, voice and text conversation, digital human conversation. See Statistics#Openings-by-domain and Statistics#Openings-by-type.

What data is available for Web Services calls?

The "Web Services Calls" section allows you to view statistical data related to web service calls made by the assistant, with the ability to set the time range and filter by assistant type and status. See Statistics#Web-Services-Calls.

Smart Document Access

What document formats are supported by Smart Document Access?

Smart Document Access supports uploading documents in the following formats: PDF, DOC, DOCX and XLSX. Uploaded documents are indexed to be used by the assistant in responses. See Document management.

How does the Smart Document Access crawler work?

The crawler analyzes websites in two phases: navigation (maps the site structure by following links) and extraction (generates indexed PDFs from page content). You can configure exclusions via regex, maximum depth, maximum pages and schedule weekly or monthly scans. See Smart Document Crawler.

What happens when I delete or replace an indexed document?

When you replace a document, the old file is deleted only after the new one has completed indexing. When you delete a document, it will no longer be available for assistant responses. You can delete multiple documents at once via multi-select. See Document management.

How do I link an assistant to SDA documentation?

Go to "Assistant Settings" and then to the "SDA" section. You can link one or more subfolders at a time, provided they belong to the same folder. The assistant will query the documentation to generate responses. See Linking an assistant to documentation.

What search types are available in Smart Document Access?

There are 3 search modes available: Vectorial (converts questions into mathematical vectors and calculates similarity), Semantic (analyzes word meaning and linguistic relationships) and Hybrid (combines both, recommended). See Smart Document Access#Search-and-answer.

How does answer generation with Gen AI work in SDA?

By enabling "Use Gen AI to generate responses", the response prompt editor is enabled. You can enter instructions that the AI will use to generate the response based on the found paragraphs. You can adjust the maximum number of input tokens and the length of the generated response. See Smart Document Access#Response-generation.

How do document display options work after a response?

You can choose whether to: not show the "Show documents" option, show only the most relevant documentation, or configure advanced settings to exclude documents from the crawler or manually uploaded ones. The "Do not show the 'Show documents' option" and "Advanced preview settings" options are mutually exclusive. See Smart Document Access#Documentation-options.

What are acronyms in Smart Document Access?

In the SDA search settings you can enter word associations (acronyms) to help the search. For example, entering "book" and "schedule" as acronyms of "reserve", the question "How can I schedule an appointment?" will be searched as "How can I schedule or reserve an appointment?". See Smart Document Access#Search-settings.

How do I configure the crawler for website indexing?

Within an SDA subfolder, enable "Enable on subfolder", add a configuration, enter the starting URL, configure navigation options (exclusions, maximum depth, maximum pages), extraction rules (filters, selectors to remove) and scheduling (weekly or monthly). See Smart Document Crawler#Setting-up-the-Crawler.