Three endpoints between
detection and conversation
ReMind plugs into the detection system you already run.
When you detect a player, your backend calls ReMind, and the conversation starts on-screen.
2s
TO ENGAGE
Reach players where they are.
98%
PASS ALL OUR CHECKS
No critical hallucination.
15yr
EXPERTISE
in RG research, AI and operations
The Intervention Flow
1
Trigger & Initiation
Your system detects a risky behavioral pattern.
Your backend calls the ReMind API with player information.
2
Intelligent Interaction
ReMind conducts a real-time, conversational, risk assessment including affordability questions. Variations can be preset.
3
Risk Assessment
As the player interacts, ReMind analyzes responses in real-time, categorising risk and raising alerts if needed.
4
Actionable Feedback
Once the intervention concludes, ReMind provides a summary, risk justification, and recommendations.
5
Quality Assessment
ReMind concludes with a thorough QA measuring player engagement, assessment accuracy, and more.
Every message is checked
before your player sees it.
ReMind is built for responsible gambling interactions. Not general-purpose chat.
That's fine for writing emails. It isn't fine for talking to someone who may be in trouble.
ReMind is built only for RG conversations and wrapped in four control layers.
The four control layers
Three architectural guardrails isolate and manage risk in real time
Tier 1: Input Shield.
Inbound player input is structured as secure JSON to proactively minimize prompt injection attempts.
Tier 2: Core Intelligence.
Strictly bounded, specialized conversational models focused entirely on responsible gambling interactions.
Tier 3: Output Guardrails.
Multi-layered semantic and deterministic sanity checks inspect every single message before it reaches your players.
A fail-safe corrective loop operating continuously behind the scenes
Self-Correction Retries.
Real-time automated adjustments triggered instantly if an output check flags an anomaly.
Instant Model Switch.
A resilient fallback strategy that shifts core language models to keep conversations running if a model fails.
98% of conversations pass all our 14 checks.
Based on over 2,000 simulated UK conversations with over 200 player profiles.
A conversation passes only if it passes every checks below.
1. Technical quality (5 checks)
Every reply complete, well-formed and timely.
2. Correct risk level (2)
ReMind's assessment matches the player's profile.
3. Correct alerts (2)
Addiction, financial harm, AML and suicide alerts fire when they should. And stay quiet when they shouldn't.
4. Assessment coverage (1)
All required topics covered, unless the risk level justified skipping them.
5. Intervention completion (4)
Key questions covered, relevant RG tools shared, valid summary sent to your team.
Explainable AI
ReMind provides more than a chat with a risk score. It provides the "why" behind every decision, ready for your next audit.
Risk Justification
We explain exactly what triggered a specific risk assessment, citing specific linguistic markers and behavioral patterns identified during the chat.
Intervention Summary
Receive a clear summary of what was discussed, the player’s receptiveness to change, and the specific tools recommended by ReMind.
Next Best Actions (Optional)
Each interaction concludes with a suggested operational path, helping your RG experts focus on the highest-priority tasks.
Illustrative example.
Live in weeks, not quarters.
Standard JSON outputs and a simple three-step flow keep integration straightforward.
Simple Authentication
Secured via server-to-server calls using dedicated API keys.
Clean 3-steps flow
Intervention initialisation
1
POST /chat
Initiate intervention with player metadata (language, activity metrics).
Receive unique intervention ID and first ReMind question.
Intervention progress
2
POST /message
Submit player input with the relevant intervention ID.
GET /message
Receive ReMind output, current risk level, real-time alert, and intervention status.
Intervention end
3
POST /summary
Request an intervention summary for a given intervention ID.
GET /summary
Receive intervention summary with justifications and recommendations.
Your detection already works.
Now act on every case.
Take Action Now.
Choose how you want to evaluate and deploy ReMind.
[ The Interactive Sandbox ]
Evaluate the Tech.
Test ReMind in our sandbox environment.
Free access.
Test your own risk scenarios.
View automated risk assessment and summary.
(Registration Required)
[ The Pilot Program ]
Validate the ROI.
Run ReMind alongside your existing outreach and measure the results on your own players.
Live in about 8 weeks.
A/B tested against your current process.
No long-term commitment.