Augmented Actions are wxrks's AI-powered translation and editing tools — a set of actions you can run on a segment to translate it, suggest an alternative, proofread it, check it for issues, or fix its tags, all without leaving the Editor.
​

💡 Who is this for? This guide is for Account Admins and any user with Editor access who need to enable Augmented Actions and find the right one for the job within the wxrks platform. Each action has its own dedicated article linked below with the full walkthrough.

Augmented Actions accelerate and enhance the human translation process — they don't replace it. Every action produces a suggestion; a linguist or reviewer still decides whether to apply it. The core actions work by blending three sources wxrks already has about your content — your Translation Memories, Machine Translation output, and Glossaries — into a single, context-aware suggestion.
​

By default, all operations use wxrks's own default models with no extra setup, but any operation can be routed to a different model, which overrides the wxrks model for that operation. An Account Admin can route some or all of these actions to an AI provider account of their own — see Configure Custom AI Providers with Model Management for the supported providers and how each action is routed. To use your own post-editing or quality-evaluation engine, or to ask wxrks to connect another engine, see Connect your own AI engine for automatic post-editing (Context Sensitive) and quality evaluation (Smells). For machine translation engines, see Connect your own MT engine (bring-your-own MT).

wxrks is SOC 2 Type 2 certified, and data shared with an AI provider follows the same encryption-at-rest, encryption-in-transit, and access-logging standards as the rest of the platform. Nothing is sent to a provider until Augmented Actions is turned on for your account (see below). For what the AI providers do with shared data (wxrks's own models and any provider you connect), see wxrks AI Privacy: Data Shared with AI Providers; for the platform's broader security practices, see wxrks.com/security.

Augmented Actions can be switched on or off at three levels, and each one depends on the level above it: the Account, the Organizational Unit, and the individual user's Editor preferences.
​

An Account Admin turns Augmented Actions on for the whole account from Settings > Translation Settings > AI Actions. The Enable Augmented Translation Actions toggle controls every action on this page at once — when it's off, none of them are available anywhere in the account, in any Organizational Unit.

This same page lists every individual AI action available on the account, including a few that don't appear as a manual button in the Editor — Review Assessment, Learning Terms, Part of Speech Classifier, Sous-Chef, and Active SpellCheck — so you can see at a glance what's active. For the full walkthrough of this page, including connecting your own OpenAI or Azure OpenAI account, see Enabling AI Integration in wxrks through OpenAI or Microsoft Azure.

Within an Organizational Unit's page, the LLM tab has its own Enable Augmented Translation Actions toggle. This only takes effect if the account above it already has Augmented Actions enabled — the Organizational Unit can turn the feature off for itself, but can't turn it on if the account hasn't.

The same tab also has an Enable Sending Learned Terms To Terminology Board toggle. It depends on the matching account setting on the AI Actions page (Settings > Translation Settings > AI Actions): when that's off, the toggle is greyed out and the tab shows "This action is disabled on your account" above it, even if Augmented Translation Actions is on.

Each user can fine-tune which actions run automatically for them from the Editor's Settings panel, without affecting anyone else's preferences:

This panel has an on/off switch for each action covered below — Enable Context Sensitive Translate, Enable Review Assessment, Alternative Suggestion, Proofread, Check Smells, and Fix Tags — plus Smells Language (Source, Target, or English, for how Translation Smells findings are displayed). Turning one off only removes it from your own Augmented Actions menu — it doesn't disable it for other users.

Selecting any segment reveals an Augmented Actions button in its bottom-right corner. Its dropdown menu is where most of these actions live:

Selecting Help at the bottom of the menu opens an in-app reference with a short description of each action, without leaving the Editor.

Each action below has its own dedicated article with the full walkthrough, screenshots, and troubleshooting — this page gives you the short version of what each one does and when to reach for it.

Blends your Translation Memories, Machine Translation, and Glossaries into one AI translation suggestion for a segment. Runs automatically the moment you open an empty segment, and on demand any time via Refresh or the Augmented Actions menu. See Context Sensitive Translate for the full walkthrough, including Translation Reasoning — a switch on the same card that explains exactly why the AI made the choices it did.

A genuinely independent second opinion — deliberately ignores your TMs, glossaries, and MT output, and reacts only to the source text and your current draft. Useful when you want a fresh take rather than a match reinforced by what's already in your TM. See Alternative Suggestion.

Checks the current translation for grammar, spelling, and punctuation issues, using the source segment as context, and applies its correction immediately — no confirmation step, but every edit is reversible from Track Changes. See Proofread, including how it's different from Spellcheck in the Editor and the QA Check Tool.

Flags a translation that might have a problem worth a second look — inconsistent terminology, awkward phrasing, a mistranslated idiom — automatically as you work or on demand via Check Smells. It runs against the same context the translation was produced with — the project's Glossary, its Translation Memory (TM), the Sous Chef Instructions in force, and, when you haven't edited the target since, the reasoning behind the Context Sensitive translation. It only flags; the final call is always yours. See Translation Smells. In the QA panel the same AI review runs in batch and produces Confidence Scores, which can confirm pre-translated segments automatically; see QA Confidence Score and Smells and Auto-Confirming Segments by Confidence Score. Which engine runs Smells and QA Smells can be set in Model Management.

Reinserts a segment's tags to match the source's tag structure — missing, extra, wrong-order, or wrong-position, all in one AI pass, without you dragging each one into place by hand. See Fix Tags.

When you edit a segment during a review step, automatically suggests one of your account's translation change categories plus a short comment explaining the change — you can always override both. See Review Assessment.

A related but independent AI feature: recognizes inflected forms of glossary terms and surfaces suggested variants in the Editor's Suggestions tab, which you apply manually. Runs on its own, separate from every action above. See Augmented Term Lookup.

A video walkthrough of these actions in the Editor is also available, though it predates the current "Augmented Actions" naming.