The biggest mistake companies make with AI, with David Harris and Vlad Bondarenko of ReferOn

Can operators now build their own affiliate software with AI? ReferOn's David Harris and Vlad Bondarenko say not yet, and explain where AI helps. They also cover ReferOn's management buyout and why winning and migrating big legacy clients is slow.

In this episode

• Bondarenko says a large product built with AI still needs people who know the job: you have to "sit in this proper chair", use the proper prompt, and get separately built modules working in one infrastructure layer (12:53, 13:10, 13:36).

• He says ReferOn already uses AI for QA test cases and for prototyping, and wants a quicker way to create tracking links in bulk from a CSV file, by AI or plain automation (15:14, 16:18, 14:42).

• Harris says AI will work alongside people rather than take over, because iGaming is built on relationships and people may not deal with "a robot" (17:43, 18:05).

• After the management buyout, clients should see no change in day-to-day support, while ReferOn gains full control of its roadmap and resources (9:59, 10:16).

• Persuasion is the biggest challenge with large clients, whose sales cycle can run six months or more; once past that stage, a migration is "fairly straightforward" (24:29, 24:47, 25:02).

• Bondarenko's goal is an autonomous affiliate program built by agents handling routine tasks, so affiliate managers can focus on deeper relationships, but he says it must be introduced very carefully (31:46, 32:59).

Episode from iGaming Real Talk, published on YouTube.