Spot signs of gambling harm in support conversations

Fields

• Channel (email, web chat, chat app, SMS, call)
• The whole conversation, oldest message first

Questions

1. Does the customer describe chasing losses? Yes / No

   Yes: The customer says they kept playing or deposited more to win back money they lost, or asks for help to win it back. No: The customer says nothing of the kind.

2. Does the customer describe gambling with money they cannot afford to lose? Yes / No

   Yes: The customer mentions borrowing, loans or credit to gamble, losing money meant for rent, bills, food or wages, debts from gambling, or being unable to pay for essentials. No: No such statement appears. Wanting a withdrawal quickly is not by itself a sign.

3. Does the customer describe losing control of their gambling? Yes / No

   Yes: The customer says they cannot stop, play for longer or spend more than they meant to, feel addicted, or regret how much they gamble. No: No such statement appears.

4. Does the customer ask to stop or limit their gambling, and how? Choice

   Choose the strongest request the customer makes anywhere in the conversation.

   • Self-exclusion The customer asks to be excluded or blocked from gambling for a period or permanently.
   • Close the account because of gambling The customer asks to close the account and gives their gambling as the reason.
   • Limit or break The customer asks for a deposit or loss limit, a time-out, a cooling-off period or a reality check.
   • No request The customer asks for none of these.

5. Does the customer express emotional distress about their gambling or its effects? Yes / No

   Yes: The customer expresses despair, panic, shame or hopelessness about their gambling, their losses or their situation. No: No such distress appears. Frustration with the service alone does not count.

6. Does the customer mention harming themselves or not wanting to live? Yes / No

   Yes: The customer mentions self-harm, suicide or not wanting to go on, in any words. No: No such mention appears.

7. Is there a sign the account holder or the person playing may be under 18? Yes / No

   Yes: The conversation mentions an age under 18, school or a school year, a parent paying or finding out, or someone else using the account who may be a child. No: There is no such sign.

8. Does someone other than the account holder raise a concern about the account holder's gambling? Yes / No

   Yes: A partner, relative or other person contacts support, or is quoted, with a concern about the account holder gambling. No: No such concern from another person appears.

9. How did the agent respond to the signs of harm in the conversation? Choice

   Judge the agent messages against our policy for customers who show a sign of harm.

   • Supportive The agent acknowledged it and offered the safer gambling tools or support, with nothing that promotes gambling.
   • Missed The agent replied without acknowledging the sign or offering any tool or support.
   • Harmful The agent promoted gambling, offered a bonus, or discouraged the customer from stopping.
   • Not applicable The customer showed no sign of harm, or no agent replied.

10. How strong are the signs of gambling harm in this conversation? Scale

    Rate the strongest evidence in what the customer or another person says.

    • None Nothing in the conversation suggests gambling harm.
    • Low One mild or ambiguous sign, such as a passing remark about losing more than planned.
    • Moderate One clear sign, such as loss chasing, a request to limit or stop, or gambling with money meant for essentials.
    • High Two or more clear signs together, or any sign of a possible minor or of self-harm.

Lens columns

loss_chasing, loss_chasing_probability, financial_strain, financial_strain_probability, loss_of_control, loss_of_control_probability, stop_request, stop_request_probability, distress, distress_probability, self_harm_mention, self_harm_mention_probability, possible_minor, possible_minor_probability, third_party_concern, third_party_concern_probability, agent_response, agent_response_probability, concern_level, concern_level_average