The next best thing to show, decided in the moment
What Next picks the best content or upsell for each fan, based on their previous action and unique context. It’s made to provide contextual offers, content, interaction and to drive progressive profiling.
Private beta from Monterosa LabsLabs products are Early Access.
- Action
- Predicted a home win in Saturday's derby
- Context
- Follows two clubs, age 15, no favourite set
- Gaps
- 17 fields your clubs asked for
- "You backed them. Cup final between your two clubs: who gets your scarf?"Question · favourite clubs70%
- "Where will you watch Saturday's game?"Question · how they watch58%
- "Which players do you follow?"Question · followed players54%
- Family ticket offerOffer · ticketsHeld back: under 18
Understand the moment and the fan universe
What Next makes fast decisions based on fan preferences, behaviours and weighted probabilities that drive the next best thing to show.
Example: one fan, straight after Wednesday's game
What they just did
The prediction, vote or quiz answer that brought them here, sent with the request.
Voted for player of the match
What you already know
The fields in their profile, and the ones your clubs still need.
Season ticket holder. Goes with family. Buys every kit.
What's happening in the match
The result, how big it was and what comes next, so the tone fits the mood.
Lost 4-1. Next home game in 4 days.
What Next picks
One question that fills a gap and suits the mood. Offers wait.
Asks how they got into the game. Holds back the kit offer after a heavy defeat.
Profiles that get richer automatically over time
Nobody likes form filling. What Next turns each missing field into a poll, a quiz question or a this-or-that, and asks when the fan is most likely to answer.
Some answers fill a field outright. Others let you infer it, with a confidence score. The two are always stored apart, so a guess is never written as a fact.
Let's settle it: what do you call a bread roll?
Tap an answer to see what it fills.
Fan profile
Region
One of the fields clubs ask for most
Favourite clubs
Only from the fan. Can be more than one.
Club leaning
Inferred from predictions. Never written as a favourite.
Learn or upsell
Every eligible option to either learn or present upsell options is weighted so What Next can determine what is most likely to generate value.
score = likely to say yes² × fit × value to you
Value comes from what your business goals are. Set each slot to data capture, upsell or both, and What Next only ranks what that slot is for. Switch it to see the ranking change.
This slot is for
Questions and offers are ranked together.
| Option | Likely | Fit | Value | Score |
|---|---|---|---|---|
| Away kit pre-orderUpsell | 73% | 70% | 0.62 | 0.231 |
| Family ticketUpsell | 55% | 56% | 0.96 | 0.163 |
| After the final whistle, what do you watch?Learn | 59% | 51% | 0.81 | 0.144 |
| Players you followLearn | 54% | 48% | 0.81 | 0.113 |
| How did you get into the game?Learn | 50% | 46% | 0.46 | 0.053 |
Sample figures for a season ticket holder who has just finished a quiz.
Guardrails
AI is wrapped in rules and guardrails that can be set up for each instance. Avoid inappropriate or non-compliant offers, questions or situations.
Under 18s
No offers, and no questions about where they live, spending or family.
After a heavy defeat
No ticket offers or marketing until the mood has passed.
Season ticket holders
Never offered a single match ticket. A family ticket is still fair.
One question per visit
A fan gets at most one question per visit, and is never asked what you already know.
Consent first
Marketing offers only go to fans who have said yes to hearing from you.
Your own rules
Add must and never rules for your competition, sponsor or season.
Every decision has a reason
Each response says what was picked, why, which guards applied and how every other option scored. Your team can check any decision, and fans can see a plain "why am I seeing this?" written by your editors.
You can start with rules, then run the AI alongside them without showing its picks to fans, then switch it on when you have seen enough.
What your developers see: the same decision, as data.
{
"picked": "scarf_after_prediction",
"source": "ai",
"reasons": [
"just_predicted",
"favourite_clubs_unknown"
],
"guards": [
"under_18: 8 removed",
"already_known: 2 removed"
],
"scores": [
{ "id": "scarf_after_prediction",
"likely": 0.70, "gain": 0.234 },
{ "id": "where_watch",
"likely": 0.58, "gain": 0.111 },
… 15 more
]
}Why am I seeing this?
You predicted a win for one of your clubs, and we'd like to know which club you'd back if they met.
What makes What Next different?
| Topic | Typical next best action | What Next |
|---|---|---|
| What it is for | Choosing the next offer or message, to sell or retain | Choosing the next question or offer, to learn about the fan or to sell |
| Missing data | Works with the data you have already | Fills profile gaps through polls and quizzes fans want to play |
| Context | Customer history, channel and recent clicks | What the fan just did, what you know about them, and the live match: score, form and fixture |
| Starting out | Propensity models learn from responses, so early picks rest on little evidence | A language model judges how well each option fits from the first fan, then outcomes are counted |
| Built for | Banking, telecoms and retail customer journeys | Sport and entertainment, before, during and after live events |
Book a demo
We’ll run What Next on your own scenarios and show you how it will increase profile accuracy and provide contextual upsell.
