AI dealer platforms, assessed: Sentient Studios against Playgon's AI Dealer
Two named B2B AI-dealer propositions are now in front of operators. One ships a single table in twelve languages on pure revenue share; the other is a milestone-funded build targeting Q3. Neither has published certification.
The cost base of a live-casino product is a floor and a payroll. Studio space is capex, presenters are opex that never stops, and every added language is another hiring pipeline. A vendor promising to remove the presenter is aiming at the category’s largest recurring line — which is why “AI dealer” went from conference panel to two named products inside a single April. We assess both as any vendor: what ships, what is roadmap, what the contract says, what is unpublished.
The two files
Shipping versus roadmap
The useful line divides product from plan. Sentient Studios has shipped something: BetHog’s funding release describes an AI dealer running since October 2025 across twelve languages. Twelve matters: localisation is the cost that scales worst in human live casino — a Portuguese table means Portuguese presenters on Portuguese shifts — and exactly the cost a synthetic presenter should collapse toward zero.
Playgon’s is earlier. The definitive agreement with Digital Nation Entertainment commits US$1,000,000 in milestone tranches against a Q3 2026 target, the release stating deployments are subject to applicable regulatory approvals. Milestone funding is discipline, not weakness. But it is roadmap, and roadmap scores only as something reality has not yet contradicted.
Contract structure, not technology, is the differentiator
Strip the machine-learning vocabulary and the two differ mainly in risk allocation. Sentient Studios sells on pure revenue share with no setup fee, no monthly minimum and no fixed term, per the release. For an operator that is close to a free option: integration effort is the only sunk cost, and a tile that fails to hold switches off without argument. It is also the pricing of a vendor that needs deployments more than margin — correct posture for an unproven category, and not generosity.
Playgon’s structure points the other way. A five-year exclusive locks a small company’s flagship roadmap to one counterparty’s delivery in a fast-moving category. Retaining 100% of the IP is the offsetting term, and a good one: Playgon owns what it pays for, which is not always how these are drawn. Defensible if DNE delivers; expensive if it does not.
The claim that needs an asterisk
BetHog chief executive Nigel Eccles is quoted in the funding release saying the company’s basic AI dealer is “10X more popular than its live dealer equivalent”, a comparison he attributes to six months of the company’s own observation. Treat it as what it is: an unaudited vendor claim in a fundraising announcement, with no published definition of “popular” — not sessions, not hold, not retention, no stated comparison set. Novelty traffic in a new format runs hot and then does not. This desk does not restate the figure as a finding.
What nobody has published
One gap decides the category, and it is the same for both. Neither has published independent certification or regulatory approval evidence, and Playgon’s release is explicit that deployment waits on approvals. That is normal for a new product class: testing houses and authorities must first work out what a synthetic dealer is. Still decisive for a buyer. Nobody puts a table on a floor on a press release, and the procurement question — who certified the RNG, the game logic, the fairness of a dealer with no hands — has no published answer.
Underneath sits the question no funding round settles: whether removing the human removes what operators are buying. Live casino’s premium over machine play has always been a person running a room — the argument we made in our review of Adventures Beyond Wonderland. An AI dealer that plays cheaply may just be an animated table game with a bigger rendering budget: cheaper, but in a far more crowded aisle.
Scope note: this desk assesses vendors for the adult professionals who buy from them, not for players, and nothing here invites anyone to gamble.
The file, both columns
Verdict
Advice to an operator’s product team: take the meeting, take the free-option pricing if it is genuinely free, and make certification evidence the gate, not the follow-up. What re-rates this category will not be a funding round. It will be the first named operator running a synthetic dealer on a live floor, with a testing house’s paperwork behind it.