Using AI in Actualog Social PIM

This topic explains how Actualog uses AI to help you manage product information: where you see AI in the interface, what it can do, and how you stay in full control of your data.

1. What AI does in Actualog

Actualog Social PIM uses AI to reduce routine manual work and help you keep product data clean, complete and consistent. In practice this means:

  • Fewer manual edits during imports
  • Faster product classification and mapping
  • Smarter attribute suggestions and content generation
  • Easier translations and localization
  • Clearer explanations when something goes wrong

AI never replaces your decisions. It suggests, explains and automates repetitive steps, but you always confirm important changes.

2. Where you see AI in Actualog

2.1. Chat assistant (“Ask Actualog” panel)

Actualog includes a built-in chat assistant that you can open from most pages.

You can use it to:

  • Ask “why?” about errors or unexpected behavior
  • Get step-by-step guidance for imports and mappings
  • Ask for explanations of concepts (attributes, categories, communities, etc.)
  • Request short texts: product descriptions, feature highlights, hints for end-users

Typical questions:

  • “Why did my last product import fail and how can I fix it?”
  • “Help me map this supplier category to our existing categories.”
  • “Suggest key attributes for this new product type.”
  • “Improve this product description for a B2B catalog.”

The assistant works in the context of your tenant, so answers take into account your actual catalogs, categories and configurations where needed.

2.2. Smart imports and data quality

When you upload product data files, AI helps you understand and fix issues faster.

You may see:

  • Clear error explanations – instead of just technical messages, the system groups issues (for example: “Missing mandatory attributes”, “Unknown units”, “Unmapped categories”) and shows suggestions.
  • Fix hints – for many problems, the assistant can propose what to change in your file or mapping rules.
  • Trial import summaries – after a test run, you can ask the assistant to summarize what went wrong and what to do next.

This reduces the number of trial-and-error cycles when integrating new suppliers or marketplaces.

2.3. Category and taxonomy assistance

Actualog often works with multiple classification systems (internal categories, standards, marketplaces). AI helps you:

  • Understand where a new product “belongs” based on its attributes and description
  • Suggest suitable categories when you import items from new sources
  • Review suggestions for new categories or templates based on external structures

In practice you may see:

  • Suggested category mappings during import setup
  • Lists of “Suggested categories” created from external trees and ready for your review
  • Suggested attribute sets when you define a new category or product template

You always confirm or correct the proposal before it becomes part of your taxonomy.

2.4. Attribute and content suggestions

On product and category screens, AI can assist with:

  • Attribute ideas – when you create or refine templates, the system can suggest attributes that are commonly used for similar products (for example, dimensions, materials, energy class, packing information).
  • Attribute extraction – for some data, AI can help extract values from semi-structured text (for example, pulling “Height: 120 cm” out of a description).
  • Marketing content – AI can generate or improve product descriptions, bullet points and short marketing blurbs based on existing attributes and your tone of voice.

You can insert suggestions as-is or edit them before saving.

2.5. Translations and localization

If you work in multiple languages, AI can speed up localization while keeping terminology consistent.

You can use it to:

  • Suggest translations for product names and descriptions
  • Translate category names, attribute labels and helper texts
  • Adjust text to specific markets (for example, US vs EU wording)

AI suggestions appear in translation fields; you review, adjust and approve them before publishing.

2.6. Proactive hints and “sparks”

In some situations, Actualog can proactively offer AI-powered suggestions (“sparks”) based on your recent activity. Examples:

  • “Your last three imports had similar errors. Do you want a quick summary and recommended fixes?”
  • “Many products in category X are missing attributes required by target marketplace Y.”

You can configure how often these suggestions appear or disable them entirely.

3. How AI works with your data

3.1. Data access

To answer questions and generate suggestions, the AI assistant may access:

  • Product records and attributes
  • Categories, templates and taxonomies
  • Import logs and validation results
  • Help content and configuration data

Access is limited to your tenant. AI features do not see data from other tenants.

3.2. Conversation history

For convenience and quality of answers, Actualog stores:

  • Chat sessions and messages, so you can scroll back and continue where you left off
  • In some cases, longer-term insights (for example, “this user mainly works with Catalog A”) to personalize guidance

Retention periods and privacy settings can be configured at tenant level. Administrators can request removal of specific conversations or, where required, broader AI data.

4. Control, safety and privacy

Actualog is designed so that AI supports your PIM process instead of silently changing data.

Key principles:

  • Human in control
  • AI never deletes products, catalogs or categories on its own.
  • AI-generated changes to important objects appear as suggestions or drafts that require your confirmation.
  • Configurable permissions
  • Administrators can limit what the assistant can do (for example, “read-only help”, “propose changes”, “assist with imports only”).
  • Proactive messages and “sparks” can be turned on or off.
  • Transparency
  • When AI is involved, the UI clearly indicates that a suggestion or text is AI-generated.
  • You can always override, edit or ignore AI proposals.
  • Privacy by design
  • AI operates only on data inside your tenant.
  • Personal data is minimized in training and logs where possible, and access is governed by the same security model as the rest of Actualog.

5. When to use AI in your daily work

Use AI in Actualog when you want to:

  • Import a new supplier file and understand errors quickly
  • Map new categories or external trees into your existing taxonomy
  • Improve or speed up creation of product content and translations
  • Get explanations and short “how to” answers without leaving the page
  • Detect gaps and inconsistencies in product data before publishing

You can start at any time by opening the assistant panel and asking in plain language what you need. If you ever feel the assistant is too active, adjust settings or speak with your tenant administrator to tune its behavior.