Vinfotech Launches AI Market Creator and AI Market Resolver for Fully Automated Prediction Market Operations

One operator brief creates thousands of markets automatically, while evidence-led resolution settles each with an audit trail. Both engines improve every cycle.

Vinfotech launched AI Market Creator and AI Market Resolver, two products that automate prediction market operations from generation through to settlement. AI Market Creator turns a single operator brief into thousands of original, resolvable markets a month. AI Market Resolver settles them against trusted sources with a confidence gate, source cross-checking and an audit trail on every decision.

Vinfotech launched AI Market Creator and AI Market Resolver to automate prediction markets. Operators set rules once, while AI generates markets and resolves outcomes using trusted sources and audits.”

— Akshay Jain, CEO Vinfotech (Viscus Infotech Ltd)

Both are available as a managed subscription or as a full source code licence, and both integrate with prediction market platforms built on other technology.

What actually decides whether a prediction market works

Liquidity and trading technology get most of the attention. But the thing users actually interact with is the question on the card, and the thing that keeps them coming back is whether those questions are interesting, timely and settled correctly.
Until now operators have had two ways to get there, and both have a ceiling.

Marketing Technology News: MarTech Interview with Mark Listes, CEO @ Pendulum Intelligence

The first is an editorial desk: a genuinely smart team writing markets by hand. This produces excellent, distinctive questions and it does not scale. Hours of research go into each market, breaking news gets missed overnight, entire regions and languages stay uncovered, and the cost line grows with every market added. Good editorial judgment is also hard to hire and harder to keep consistent across a team.

The second is copying. Watch what the large platforms list, mirror it, ship it. This is fast and it is what most new platforms do. It also means a user has no reason to choose your platform over the one you copied, because you are offering the same catalogue slightly later.

Vinfotech built a third path. AI Market Creator generates original markets continuously, in any language, at a volume no editorial desk can match, while staying inside rules the operator defines. AI Market Resolver settles them with evidence rather than judgment calls. The combination lets a platform build a catalogue that is genuinely its own rather than a lagging copy of someone else’s.

“A prediction market platform is its questions,” said Akshay Jain, CEO of Vinfotech. “You can hire a brilliant editorial team and cap yourself at what they can write in a day, or you can copy the big platforms and give users no reason to pick you. We wanted a third answer, where the catalogue is original, the volume is unlimited and the operator still decides exactly what gets made.”

Marketing Technology News: How MarTech Is Enabling Autonomous Brand Engagement Across Channels?

AI Market Creator: one brief, thousands of original markets

Operators brief the engine once with a category, a language and a set of editorial instructions. From there an AI pipeline orchestrator runs multiple generation engines in parallel, using AI and crawlers to turn live signals into drafted, resolvable market questions around the clock.

Full autopilot is the default: discovery, drafting, quality checks and publishing run with no human touch. Operators in regulated or sensitive categories can switch on a review queue where candidates wait for approval before going live.

Generation runs natively in any language, with titles, options, deadlines and resolution rules produced in the target language rather than translated afterwards.

The rules operators set

Automation without constraint is the failure mode Vinfotech designed against. Every generation cycle obeys rules the operator defines.

Market structure. Whether an engine produces binary YES or NO markets, multi-option markets, or both.

Outcome logic. Within multi-option markets, whether a market resolves to a single winning option or supports multiple winning outcomes, so one engine can address genuinely different event structures.

Resolution date limits. Operators can require that an engine only generates markets resolving on or before a chosen date, keeping a catalogue aligned to a campaign window, a sporting calendar or a settlement cycle rather than filling up with long-dated questions.

Category, language and editorial scope. Sports, finance, politics, crypto, technology, entertainment, culture and regional topics, each configurable against an operator’s business model and compliance position.

AI Market Resolver: settled with evidence, not guesswork

When a market opens, the Resolver registers it and plans which sources will decide it and when to check them. Checks are scheduled around the expected close, sparse while nothing can happen and rapid as settlement approaches. When the event occurs it pulls the result from official pages, data feeds and trusted news, saving each with a timestamp and an extracted snippet. Only then does it propose an outcome, with the proof attached.

Clear cases settle automatically. Ambiguous ones wait for a one-click operator decision. Every market carries a visible operational state across nine lifecycle stages, so nothing settles invisibly and nothing fails silently.

Accuracy: it would rather not settle than settle wrong

Every safeguard in the Resolver follows one rule, which is to act only when the evidence is overwhelming.

A strict confidence gate. The engine settles only past a high confidence threshold. Anything borderline routes to review rather than to payout.

Sources must agree. Independent sources are cross-checked for the same answer. A single conflict holds the market for a human.

It can say it is not sure. Where a rushed analyst would make a call, the engine abstains. It holds, escalates or voids instead of forcing an outcome.

The operator holds the override. Any outcome can be reviewed, corrected and re-settled, and the trail shows exactly what the engine saw and why it decided as it did.

The consistency argument matters as much as the accuracy one. An automated resolver never gets tired, rushed or inconsistent. Market ten thousand receives the same rigour as market one, on a Sunday night as on a Tuesday morning, which is not true of any human settlement desk at volume.

Every settlement carries an audit trail. There are no black-box decisions, which means a disputed outcome gets an answer rather than an apology.

Every cycle makes the next one better

Separately, each product solves a bounded problem. Together they change what an operator’s week looks like.

The team briefs the engines once. Overnight, markets appear across every category and language the operator asked for, each written to be resolvable, each inside the rules the operator set. Events happen. The Resolver watches the sources it planned for, gathers proof and settles. Clear cases pay out on their own. The handful that are genuinely ambiguous arrive in a review queue with the evidence already assembled, so a decision takes seconds rather than an afternoon of research.

The team’s job changes shape. Instead of writing markets and chasing results, they steer: adjusting briefs, flagging questions that missed the mark, approving the edge cases.

And the steering compounds. Every flag a reviewer raises is written into the engine’s editorial memory, so the next cycle drafts closer to the operator’s standards and the cycle after that closer still. Every settlement adds to what the Resolver knows about which sources hold up for which categories. Three months in, the review queue is shorter than it was in week one, not because anyone lowered the bar but because the engines learned where it sits.

This runs against how software usually behaves. Most platforms are at their best on the day they are installed and drift from an operator’s needs from then on. These move the other way. The catalogue grows more distinctive, the operations get lighter, the settlements get cleaner, month after month, without the operator hiring a single additional person.

Meanwhile an operator mirroring markets from a larger platform is doing exactly the same work in month twelve as in month one, and offering exactly the same catalogue as everyone else running the same strategy.

“The interesting part is not what the engines do on day one,” said Jain. “It is what they do on day three hundred, after they have absorbed a year of your team’s judgment. By then you are not running a copy of somebody else’s catalogue. You are running something nobody can reproduce, because it was trained on decisions only your team made.”

Customisation and source code

Both products are available as a managed subscription, with Vinfotech running the engines, or as a full source code licence for operators who want to host, extend and own the system inside their own infrastructure. Both also connect to prediction market platforms built on other technology, so an operator running elsewhere can add automated creation, automated settlement or both without replacing their stack. For operators starting fresh, the two ship as part of Vinfotech’s white label prediction market platform.

Both products are on an active development path, with the training layer, algorithm improvement, settlement structures and operator controls expanding continuously. Operators on subscription receive that work as it ships.

“This is not a category where anyone gets to declare the problem solved,” said Jain. “Market structures get more complicated, sources change, edge cases arrive that nobody modelled. We would rather tell an operator that this is an ongoing journey and that they can either travel it with us on a subscription or take the code and travel it themselves.”

Write in to psen@itechseries.com to learn more about our exclusive editorial packages and programs.

source https://martechseries.com/predictive-ai/ai-platforms-machine-learning/vinfotech-launches-ai-market-creator-and-ai-market-resolver-for-fully-automated-prediction-market-operations/