Skip to content
Operators

The Price Is Only Half the Bet: How AI Is Changing Sportsbook Trading

Kambi's data suggest faster AI-driven bet acceptance can boost accepted volume and revenue without necessarily increasing operator risk exposure.

Industry Analyst & Commercial Partnerships

· 9 min read

The Price Is Only Half the Bet: How AI Is Changing Sportsbook Trading — by Martin Eriksen, Industry Analyst & Commercial Partnerships

In its 4 November 2026 results, Kambi reported that its AI was handling 85% of bets across its network, with 59% of AI-traded live wagers accepted without a built-in delay. The technology could change how sportsbooks compete, not just through better odds, but by accepting more bets without taking on excessive risk.

Imagine two bookmakers offering almost identical odds on a football match. Both use sophisticated trading systems and live data to manage their exposure. Then a customer places an in-play bet. One bookmaker accepts it almost immediately, while the other takes several seconds before offering revised odds.

From the customer’s perspective, the difference is considerable. Both bookmakers advertised a price, but only one accepted the bet at that price without the usual wait.

For years, bookmakers have competed on odds, promotions, market coverage and betting apps. Those things still matter, but AI is bringing another part of the business into focus: what happens between displaying a price and accepting a wager.

Kambi’s AI is already changing how bets are accepted

According to Kambi’s Q3 2026 report, approximately 85% of bets across its network are now traded through its AI system. That figure covers Kambi’s partner sportsbooks rather than the wider betting industry, but it shows how far the technology has already progressed.

The more interesting figure concerns live betting. Kambi says its AI now accepts 59% of live wagers without a built-in delay. This doesn’t mean every transaction is literally instantaneous, but the system can remove the deliberate waiting period traditionally associated with in-play betting.

Anyone familiar with live sports betting will recognise that processing screen. You select your odds, submit a wager and wait while the bookmaker checks whether the price is still valid. If a goal, injury or another important event occurs, the odds may change, the market may be suspended or the bet may be rejected.

Those delays serve a purpose. Bookmakers need protection against accepting bets at prices that no longer reflect what’s happening in a match. The problem is that fixed delays can treat relatively safe bets much like those carrying considerably more risk.

Kambi’s system takes a different approach. It assesses individual wagers using match information, the bet type, and its confidence in the risk assessment. When risk is sufficiently low, it can accept the bet without the usual delay.

During its 4 November earnings call, Kambi estimated that the improvement would lead to approximately 18 million additional accepted bets during the year. It also reported that NBA live-bet rejections had fallen by around 32% compared with the previous year after moving the league to its AI trading platform.

The NBA comparison covers only the first few weeks following the change, while the 18 million figure is a projection. Nevertheless, both suggest that faster processing can help bookmakers accept wagers they might otherwise lose.

A rejected bet is business the bookmaker never receives

Sportsbooks spend heavily on attracting customers, developing their products and offering thousands of betting markets. Yet when someone finally tries to place a bet, the transaction may fail because the system can’t quickly confirm whether the price is still appropriate.

Some rejections are essential. A bookmaker accepting outdated odds immediately after a goal or red card could face substantial losses. But unnecessary rejections also cost money, and customers have little reason to keep trying if they regularly face delays or changing prices.

This is where Kambi’s AI becomes commercially interesting. Instead of making customers wait for every live bet, the system can assess which transactions need additional checks and which it can safely accept immediately.

For operators, the potential benefit is straightforward: more accepted bets without a corresponding increase in exposure to outdated prices. That could mean higher turnover and additional revenue without acquiring more customers.

Kambi’s network recorded an operator trading margin of 11.9% in the third quarter of 2026, its second-highest quarterly figure on record. The company cites its network scale and AI capabilities as competitive advantages, though sporting results also influenced the margin.

We cannot establish how much of that performance came from AI. What matters is whether the technology can consistently increase accepted betting volume while keeping trading margins at acceptable levels. Unlike a promotional offer or redesigned betting app, most of this improvement happens out of sight.

The best advertised odds aren’t always the best available bet

Odds comparisons usually focus on the prices bookmakers display. That makes sense, since higher odds generally offer better financial value when everything else is equal. But there’s an important limitation: the advertised price is only useful if the customer can actually obtain it.

Consider two hypothetical sportsbooks offering odds on the same football outcome.

Sportsbook ASportsbook B
Initially displayed odds2.102.08
Requested stake£100£100
At acceptanceRepriced to 2.04Original odds accepted
Potential profit at accepted odds£104£108
Illustrative example, assuming both transactions can be completed under otherwise equivalent conditions.

Sportsbook A initially looks more attractive. Its odds of 2.10 are better than Sportsbook B’s 2.08, but by the time the wager is accepted, the price has fallen to 2.04. The customer would have been better off using Sportsbook B.

This doesn’t mean Sportsbook A necessarily has inferior technology. Prices move for legitimate reasons, and a bookmaker may be entirely justified in revising its odds. The example simply shows why advertised prices don’t always tell the whole story.

Accepted odds, permitted stake, and the likelihood of completing the transaction all matter. This also raises an interesting question for odds-comparison services and betting affiliates. Most comparisons use the odds bookmakers display, but those figures don’t show whether customers will consistently receive those prices.

As AI becomes more involved in trading, bet acceptance reliability could become another key measure of sportsbook quality.

The other algorithm is looking for pricing mistakes

Kambi’s technology is also developing in response to increasingly sophisticated software used by professional bettors.

Statistical models, automated odds comparisons and computer-assisted betting are nothing new. Professional bettors have used technology to identify pricing mistakes for years. What’s changing is how quickly these tools can monitor markets, identify pricing mistakes and act on them.

In its quarterly report, Kambi identifies AI betting agents that exploit pricing inefficiencies as a challenge for its trading systems. The company’s response goes beyond simply accepting or rejecting a wager. Its AI can update underlying match parameters after bets are placed and automatically adjust prices across related markets.

Consider what happens when a sophisticated betting model identifies an attractive price. The bettor places a wager, but that transaction may also give the bookmaker information to reassess the market. Other customers could then encounter different odds.

Not every professional bet necessarily tells a bookmaker something new. Prices also change because of sporting developments, shifts in market exposure and other information. Kambi hasn’t disclosed enough detail to establish exactly how its system responds to individual customers.

Still, the direction is clear. A successful betting model needs more than accurate predictions. It also needs to place wagers before the bookmaker changes the price.

For operators, faster and more connected trading systems could reduce the time pricing mistakes remain available. For sophisticated bettors, that could make some strategies harder to execute.

Who actually benefits from faster acceptance?

Three parties have different interests in this development: the sportsbook operator, its technology supplier, and the customer.

For operators, the immediate attraction is additional accepted betting volume. Kambi’s projected 18 million extra bets represent transactions its systems expect to accommodate because fewer wagers will be lost through acceptance delays and related rejections.

If those wagers generate commercially acceptable margins, they could create additional revenue without requiring the operator to acquire an equivalent number of new customers.

For Kambi, the commercial interest is also direct. CEO Werner Becher stated during the November earnings call that the additional accepted wagers were expected to generate more revenue for both Kambi and its sportsbook partners.

The precise financial benefit depends on commercial arrangements, the types of bets accepted and the resulting trading performance. The projection does not establish how much incremental profit either party will ultimately earn.

Customers receive a different benefit. Faster acceptance reduces transaction uncertainty, while fewer unnecessary rejections can make the experience more reliable. That is a genuine product improvement, even if the bookmaker also benefits commercially.

But convenience is not the same as financial value. A bet accepted immediately at poor odds remains a poor-value bet, while a more accurate system for identifying pricing mistakes could make some favourable opportunities less common.

AI can therefore improve the betting experience without improving customers’ expected returns. Faster transactions, better operator margins, and better value for bettors are three different outcomes, and the industry should not treat them as interchangeable.

Should we start measuring execution quality?

If more bookmakers adopt this technology, comparing their advertised odds may no longer be enough.

A more useful assessment would consider four things:

  • Original-price acceptance rate: How often does the bookmaker accept a bet at the odds initially displayed?
  • Accepted stake: How much of the customer’s requested stake will the bookmaker actually take?
  • Repricing and rejection rates: How often do odds change or wagers fail before acceptance?
  • Acceptance time: How long does the bookmaker take to process a bet?

These figures would tell us more about the experience customers actually receive.

Speed alone isn’t enough. A sportsbook might process every request almost immediately but frequently offer lower odds or reduce the requested stake. That wouldn’t necessarily make it a better product.

Results would also need to be compared across different sports and markets. A routine pre-match football bet is not the same as an in-play wager submitted during a rapidly changing event.

Kambi has provided evidence of faster processing and fewer NBA rejections, but there isn’t yet enough public information to compare acceptance quality in this way across individual partner sportsbooks.

For operators and suppliers, some of these figures may be commercially sensitive. For customers, affiliates and independent analysts, they could provide a much more complete picture of sportsbook performance.

Faster acceptance doesn’t mean fewer customer restrictions

An important distinction also exists between accepting an individual wager and deciding how much business to accept from a particular customer. In July 2025, the UK Gambling Commission reported that 643,779 of approximately 14.9 million active betting accounts in its sample had experienced some form of commercial restriction during 2024, equivalent to 4.31%.

Those restrictions included reduced maximum stakes and account closures. The figures don’t establish that AI was responsible, nor do they relate specifically to Kambi’s technology.

However, they do illustrate a separate issue. Removing unnecessary delays from live betting doesn’t mean a bookmaker will become more willing to accept business from customers it considers commercially undesirable.

Faster processing also needs to operate alongside responsible gambling controls. Making repeated in-play bets easier should not come at the expense of monitoring potentially harmful behaviour.

The next sportsbook advantage may be harder to see

We’ve previously examined how bookmakers rely on specialist data providers, pricing models and trading technology in There Are Hundreds of Sportsbooks. Who Actually Makes the Odds?

Kambi’s latest figures show how that technology is changing. AI can help bookmakers keep markets available, assess individual bets and reduce unnecessary rejections, often without customers noticing what’s happening behind the scenes.

For operators, that could become a meaningful competitive advantage. A bookmaker that accepts more appropriate bets at competitive prices, without taking excessive risks, may have an edge over rivals using slower or less responsive systems.

For customers, the test is simpler. Does the bookmaker accept the wager at the odds and stake originally requested, or does the price change before the bet goes through?

Kambi has shown that AI can speed up acceptance and reduce some rejections. Less clear is if those gains will consistently produce better accepted prices or greater value for bettors.

The price is only half the bet. Increasingly, the technology deciding whether to accept it may be just as important.

Contribute