Building Sunny
The World's First AI-Powered Blackjack Dealer Crypto casino BetHog recently introduced Sunny, the world's first AI powered blackjack
dealer. Nigel Eccles, Co-Founder and CEO, underlines why Sunny represents proof that AI can deliver enhanced entertainment experiences, not just efficiency improvements.
Te idea for Sunny came earlier in the year when I grew frustrated with the experience of playing live-dealer games. Te promise of live- dealer has always been compelling: you're playing with a real human who brings fun and interactivity to the experience. Every time I sit down to play, I imagine the energy of playing with a great dealer in a real casino. But too often, the actual experience falls short of that vision.
Our players tell us that over 80 per cent of the time, the dealer feels robotic and unengaged. Tis isn't surprising when you understand the reality of the job. Live dealers typically work late-night shifts, often in a language that isn't their first. Tey're managing the mechanical aspects of dealing while trying to entertain customers they can't see or hear.
Te economics of live-dealer operations make this even harder since studios run multiple games simultaneously, spreading dealer attention thin. Dealers work long shifts under constant camera surveillance, having to maintain game integrity while somehow keeping their energy up for invisible players on the other side of the screen. Tere has to be a better way!
BUILDING SOMETHING BETTER
We set ourselves an ambitious challenge: create an AI dealer that felt like the best real-world dealer you've ever experienced. Tis dealer would be available 24/7, always engaging, always fun, and always making the game better. We quickly discovered that building the first
98
version of our AI dealer, Sunny, was surprisingly easy. But through repeated internal play, we hit our first major problem. Te early version became incredibly repetitive. Sunny would say the same things over and over, commenting in situations where silence would have been a better option. Our testers started tuning her out completely.
Sunny is built on Large Language Models (LLMs), and getting her personality right required understanding not just what to say, but how and when to prompt the model. Te repetition problem turned out to be one of our biggest technical hurdles. LLMs naturally tend toward repetitive output, especially in constrained conversational settings. Tis happens because of how the models prioritise recently generated text and how their probability calculations can get stuck in patterns.
We tackled this from multiple angles. We tuned the temperature settings to control randomness, added penalties to discourage reusing phrases, and built a contextual memory system that tracks what Sunny has said across multiple sessions. Te real breakthrough came when we built a sophisticated state management layer that sits between player actions and the LLM itself.
Tis layer tracks everything: game state, recent conversation history, player patterns, and emotional context. When Sunny speaks, she's not just reacting to the immediate moment. She's considering dozens of variables such as your win streak, how long it's been since her last comment, whether you've seen similar cards recently, and critically, what she's talked about in the last fifty interactions.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132 |
Page 133 |
Page 134 |
Page 135 |
Page 136 |
Page 137 |
Page 138 |
Page 139 |
Page 140 |
Page 141 |
Page 142 |
Page 143 |
Page 144 |
Page 145 |
Page 146 |
Page 147 |
Page 148 |
Page 149 |
Page 150 |
Page 151 |
Page 152 |
Page 153 |
Page 154 |
Page 155 |
Page 156 |
Page 157 |
Page 158 |
Page 159 |
Page 160 |
Page 161 |
Page 162 |
Page 163 |
Page 164 |
Page 165 |
Page 166 |
Page 167 |
Page 168 |
Page 169 |
Page 170 |
Page 171 |
Page 172 |
Page 173 |
Page 174 |
Page 175 |
Page 176 |
Page 177 |
Page 178 |
Page 179 |
Page 180 |
Page 181 |
Page 182 |
Page 183 |
Page 184 |
Page 185 |
Page 186 |
Page 187 |
Page 188 |
Page 189 |
Page 190 |
Page 191 |
Page 192 |
Page 193 |
Page 194
