RSI Sports Group builds sports infrastructure for clubs and venues, from the village pitch to the top of the competitive pyramid, under its own brands such as Grass2Grass, Duol, VersaCourt and PolyPrint. For PolyPrint, their printed courts and panels, the Sales Assistant AI agent has been looking for matching customers every day since late August 2026. In two weeks it assessed 1,123 candidates and lined up 79 leads with a named contact, 68% of which RSI approved.
RSI Sports Group has been building sports infrastructure for twenty years, and does it its own way. Studying grass first, then developing LED technology that makes it grow faster. Artificial turf that is 100% reusable, an airdome with an A energy label, a padel court you can move, tiles made from recycled polypropylene. It earned them a FIFA Innovation membership, and a place with the clubs and venues that set the bar. That same instinct, trying something a little earlier than the rest, brought them to an AI agent for the search work that precedes selling. Because who buys their printed courts and panels differs per project: the buyer sits in a different country, a different language and a different link of the chain every time. The moment they become interesting is a signal you have to spot coming, such as an awarded tender or a new build project in its design stage. That search work came on top of the regular work of a small team, so it happened in bursts.
The Sales Assistant is built inside the NextPrompt platform and is operated through chat. In conversation, six markets were put together, each with its own search terms, exclusions, target roles and scoring criteria. The approach is signal-driven and location-driven: the agent searches for events such as awarded tenders, new builds and expansions, and puts the trade register and the map alongside that. Per market it runs a round every day, deduplicates against what is already there, enriches what remains with a contact person, email address and LinkedIn profile, and scores every candidate against that segment's criteria. Anything below the threshold never makes it in.
Between 26 August and 11 September 2026 the agent completed 107 full rounds, averaging 48 seconds each. In those rounds it ran 789 searches, assessed 1,123 companies, filtered out 908 duplicates and kept 79 leads: 71 with a LinkedIn profile for the contact person and 54 with an email address. RSI reviewed every one of them themselves, because nothing goes out that nobody has seen. 68% made it through; the rest were rejected with a mandatory reason, usually because the company did not fit the segment. Those reasons feed back into the agent, so the next round searches more sharply.
Inside the workspace
The screens where RSI steers the Sales Assistant and reviews its leads.
“Very satisfied with the collaboration, approach, planning and result.
We map your use case, build the first Skill or agent, and hand it over to your own people. Fixed price, fixed timeline.