The Arwen platform
Comment risk, insight and engagement can’t be solved with separate tools or delayed analysis. Arwen is the underlying platform that makes it possible to collect, understand, and act on every comment across languages, platforms, and use cases, without overwhelming teams.
It's the system that powers our products: Moderate, Engage, Protect, and Signals.
Why a platform approach matters
Most social tools stop at collection or reporting. They show you what happened after the moment has passed.
Arwen is designed to operate live. It runs continuously in the background, making decisions, enforcing policy, and surfacing what matters while conversations are still unfolding. That requires a different kind of system: one that combines automation, context, and human control by design.
Step-by-step
Arwen operates as a continuous loop with four stages. Each stage builds on the last, turning raw comments into safe outcomes and clear decisions.
Arwen connects directly to social platforms via secure APIs and collects all your comments across paid and organic content as they appear.
The platform understands text and emoji by default, supports over 30 languages, and can extend to images, GIFs, and video when needed. This creates a complete, real-time comment layer: the foundation for everything Arwen does next.
Every comment is analysed in context. Arwen looks beyond keywords to understand intent, sentiment, topic and risk.
Under the hood, this classification layer combines proprietary classifiers, fine-tuned large language models, open-source techniques, and rule-based logic where detection precision matters. The goal isn’t novelty, it’s accuracy, consistency, and explainability.
This is how Arwen distinguishes genuine questions from noise, early warning signs from isolated incidents, and coordinated behaviour from one-off comments.
Once content is understood, Arwen acts, based on rules you control. Clear violations are handled automatically and consistently through Moderate, reducing reviewer load without drifting into over-moderation. Repeat harmful behaviour can be managed quietly through configurable thresholds.
When engagement matters, Arwen can prioritise valuable comments and suggest on-brand, factually accurate replies through Engage, while keeping humans firmly in the loop.
And when risk escalates beyond routine moderation, Protect supports investigation and escalation with context and evidence, so teams don’t have to guess what’s serious. Automation handles scale. Humans retain judgement.
Patterns, shifts in tone, emerging narratives, and recurring issues are surfaced as Signals, tailored to your brand, your strategy and the needs of different roles in your organisation. These are bespoke prompts for action, designed to reduce uncertainty and speed up decision-making.
We also provide standard reports, providing a record and overview of what Arwen has done and the impact delivered.
This is how teams move from reacting to comments to understanding what’s happening while it’s happening.
Platform principles and capabilities
Arwen combines multiple AI techniques, real-time infrastructure, and human oversight into a single operating system for social.
Arwen uses AI and large language models where context and interpretation matter, and avoids them where rules and precision are more reliable. This hybrid approach delivers higher accuracy, fewer false positives, and decisions teams can explain and defend.
Arwen operates on live comment streams, not batches or delayed reports. Collection, analysis, and action happen as conversations unfold, so teams can intervene while outcomes are still changeable.
Every client operates with different standards, risks, and thresholds. Arwen is configured around your policies, tone, and escalation rules, ensuring the system enforces what matters to you, consistently and at scale.
Arwen doesn’t overwhelm teams with charts or reports. It surfaces role-specific signals and recommended next actions, set out in narrative form, helping people understand what’s changing and what to do, without slow manual analysis.
Routine, high-confidence actions are automated to maintain speed and consistency. Ambiguous, sensitive, or high-risk situations are surfaced for human review. Automation supports judgement, it never replaces it.
Accuracy isn’t just about catching more harm, it’s about not taking the wrong action. Arwen continuously monitors outcomes, incorporates human feedback, and tunes models to minimise over-moderation and unnecessary escalation.
From a handful of profiles to global operations, Arwen scales horizontally without changing how teams work. The same platform supports organic engagement, paid performance, Trust & Safety, and high-risk escalation.
Every decision Arwen makes is visible. Teams can review actions, understand why they happened, and adjust behaviour over time. There are no black boxes, only systems teams can trust.
Human review isn’t an exception case, it’s part of the feedback loop. When teams correct or intervene, Arwen learns, improving accuracy and confidence over time.
Arwen is used daily by teams managing performance, reputation, and risk at scale, across industries where mistakes are costly and judgement matters. These aren’t just customers. They’re partners who help shape where Arwen goes next.
Discover how our platform combines AI, automation, and human oversight to power every decision.