FX and CFD brokers are using AI to predict client value, engagement, and churn as rising acquisition costs and regulatory pressure push the industry away from volume-based onboarding toward behavior-based client qualification. Prakash Bhudia, Chief Growth Officer at Deriv, and Ivan Kunyankin, Data Science Team Lead at Devexperts, told Finance Feeds that brokers now prioritize early intent signals—including deposit speed, demo account usage, and first trade completion—over traditional metrics like cost per lead and first-time deposits. The shift reflects growing recognition that a smaller pipeline of qualified clients outperforms high-volume funnels that generate operational strain, weak retention, and compliance risk. Deriv automated 97.4% of client withdrawals by June 2026 and uses a 90-day AI model that identifies 68% of future high-value clients, while Devexperts warns that static segmentation by geography or deposit tier cannot reliably separate serious traders from casual users. Industry executives argue that acquisition has become the opening move rather than the entire strategy, with post-onboarding behavior now shaping product and marketing decisions across retail trading platforms.
Deriv Shifts Focus from Volume to Post-Onboarding Behavior
Prakash Bhudia, Chief Growth Officer at Deriv, told Finance Feeds that brokers track cost per lead and first deposit metrics but consider post-onboarding activity more revealing of client relationships. Bhudia said the company analyzes whether clients return without prompting and whether trading patterns appear sustainable rather than single-deposit behaviors. Deriv groups clients into active, at-risk, dormant, or churned categories and responds to each group differently. Bhudia identified smooth first withdrawals as a major trust moment for new clients, noting that Deriv automated 97.4% of client withdrawals by June 2026. He said lifetime value is a named priority in internal growth planning rather than a retrospective metric. Bhudia described acquisition as the opening move, with post-signup activity shaping product and marketing decisions.
Early Deposit Speed and Demo Usage Signal Client Intent
Bhudia said the first few days after registration separate intent from curiosity. He identified deposit speed and size as the strongest early signal, stating that clients who move quickly from registration to depositing meaningful amounts are far more likely to become high-value clients. Demo account activity before going live is the next strongest signal, with clients who practice before depositing retaining better than those who skip demos. Completing a first trade is a critical indicator, as clients who trade at least once are significantly more likely to build lasting habits than those who deposit without trading. Deriv runs a model on a 90-day window that identifies 68% of future high-value clients using these signals. The company is adding richer behavioral data, including in-app activity, feature usage, and time on platform, to improve the model. Bhudia said speed is the biggest indicator of intent versus curiosity.
Behavioral Segmentation Replaces Static Tier Systems
Bhudia said segmentation has moved beyond geography and deposit size, which he described as demographics dressed up as segmentation. Deriv's approach examines engagement patterns, responses to education, and whether activity connects to promotions. The company's AI nurture engine and AI persona agent treat clients according to live behavioral profiles rather than static tiers based on deposit amounts. Bhudia warned against dismissing bonus-driven clients too quickly, noting that some become top traders eighteen months later. He said deposit size indicates capability but not intent, and the two are often treated as the same. Bhudia concluded that a large first deposit reveals what someone can do but not what they will do.
AI Implementation Requires Process Redesign
Bhudia acknowledged that AI helps some brokers improve funnel quality, but most of the industry relies on static segmentation by geography, deposit tier, and acquisition channel. He said brokers ahead are using AI to serve clients in real time rather than simply labeling them. Deriv's personalization layer produces AI-personalized emails that run at two to two and a half times the performance of generic campaigns. The company's support agent Amy handles a large share of client interactions globally after Deriv rebuilt her workflow from scratch rather than automating old scripts. Bhudia said getting there took significant effort because what works on paper might fail in practice. He stated that the technology exists, but the gap is whether businesses will rebuild processes around what AI can do instead of adding AI to pre-AI systems.
Devexperts Links Client Quality to Retention Over Acquisition Volume
Ivan Kunyankin, Data Science Team Lead at Devexperts, told Finance Feeds that brokers have always focused on attracting and retaining traders, but competition has intensified. He said the pandemic increased time spent at home and grew the retail trading segment, while advances in technology and AI made it harder for traditional brokers to compete with new offerings. Kunyankin said these factors produced a marked shift toward building longer-term relationships and retaining a strong, high-value client base. He said AI-powered tools like Devexperts' DXtrade user profiling use real data to determine client information relatively soon after joining. Kunyankin explained that timing depends on trading activity volume rather than calendar time, with systems able to start building profiles after a certain number of trades. He said an image can start to be shaped a few trades in, though longer observation periods improve prediction accuracy.
Sudden Activity Changes Indicate Churn Risk
Kunyankin said different brokers define predictive client behavior differently based on their offerings, goals, geographic location, and regulatory environment. Devexperts found that rapid changes in behavior are a strong indicator of churn. Kunyankin gave the example of a dormant trader suddenly becoming very active, logging in often, and selling positions as likely signals of intent to leave. He said steady, consistent trading or measured, balanced behavior from early on tends to predict longer-term value. Kunyankin stated that static filters and heuristics cannot separate high-intent prospects from casual or bonus-driven users, warning against overreading early deposit behavior. He said brokers need advanced solutions that use behavioral data to make accurate, nuanced assessments based on traders' actions and habits early in the user journey. Kunyankin noted that AI frameworks analyzing large volumes of trader data can be extremely efficient in predicting outcomes.
FAQ
What early signals do FX brokers use to predict client value?
FX brokers use deposit speed and size, demo account activity before going live, and completion of a first trade as the strongest early signals of client intent. Deriv's 90-day AI model identifies 68% of future high-value clients using these behavioral indicators, according to Chief Growth Officer Prakash Bhudia.
How has client segmentation changed in retail trading?
Client segmentation has shifted from static geography and deposit tier categorization to live behavioral profiling. Deriv groups clients into active, at-risk, dormant, or churned categories and uses AI to respond based on engagement patterns, education responses, and activity connection to promotions rather than fixed deposit-based tiers.
What behavioral changes indicate a client is likely to churn?
Rapid changes in behavior, such as a dormant trader suddenly becoming very active, logging in often, and selling positions, are strong indicators of churn. Devexperts Data Science Team Lead Ivan Kunyankin said steady, consistent trading from early on tends to predict longer-term value, while sudden activity spikes often signal intent to leave.
