Marketers are drowning in AI tools, with a third reporting too many similar products that don't connect, according to a HubSpot report. A Singapore-based startup launching its public beta this week, Soma AI, aims to fix that fragmentation by pulling ad performance data, brand intelligence, and competitor research into a single plain-language workspace - built for less than US$100,000.
The platform, co-founded by Ian Chapman-Banks and Cheryl Ng-Dharsan, consolidates martech functions that typically live in separate dashboards. Chapman-Banks, a former Apple, Microsoft, and Motorola Asia Pacific executive who has led five exits worth over US$4 billion, said the lean build was possible because of how frontier AI models have changed development economics. "It's absolutely astonishing how the frontier models have changed the way that entrepreneurs think," he said.
Four components under one roof
Soma AI structures its offering around four modules. Brand Hub captures a company's voice, competitors, and customer personas so AI outputs carry context rather than generic advice. Ads Intelligence ingests data from Meta, TikTok, Shopee, and Google Ads, reconciles metrics that each platform measures differently, and flags bottlenecks without requiring users to stitch together spreadsheets. Brand Intelligence handles competitor monitoring and trend tracking. A fourth layer, Vibe Marketing, is a conversational workspace where marketers ask an LLM for plain-language recommendations.
The team stresses that Soma isn't simply a wrapper around off-the-shelf models. Ng-Dharsan, who spent 22 years in marketing roles at companies like ViSenze and Alpha JWC Ventures, recounted a tester who argued her Claude subscription could analyze an ad report just as well. The model flagged her cost per mille as too low. "But the lower your CPM, the better. Claude wasn't thinking like a performance marketer," she said.
Built on marketing brains, not tech training data
Soma differentiates in two ways. The prompts and guardrails were written by the co-founders themselves, drawing on their combined marketing experience, so the AI reasons from domain knowledge rather than broad training data. Users also feed the platform their own brand data upfront, giving every subsequent output specific context. The platform runs on Claude Opus for heavier analytical tasks and Sonnet for lighter ones, with Gemini as a backup to manage token costs and output quality.
Ng-Dharsan said the bigger competitive threat isn't established players like Supermetrics or Funnel.io, which dominate ads reporting but stop short of strategy. It's marketers who believe the LLMs they already pay for are sufficient. "Soma was built by marketing brains, but Claude was built by tech people who are trying to teach you how to do marketing," she said.
Chapman-Banks was candid about the platform's central vulnerability. "The weakness is anybody can do it," he said. "The strength is: do you have the go-to-market capability?" He added that the point extends beyond Soma. "Anyone who understands a vertical deeply - logistics, maritime, marketing - can now do what we've done."
Pricing, scale, and what comes next
Soma opens its public beta after roughly 50 users tested the platform in a closed period over the past two weeks. The team targets 250 customers by the end of Q3 and considers 1,000 users a meaningful scale in the region's martech context. Fundraising isn't on the immediate agenda; Chapman-Banks said efforts will likely begin toward the end of 2026.
Pricing starts at US$59 per month for the Starter tier, scaling to US$179 for Pro and US$449 for Max. Unlike traditional SaaS subscriptions, plans are token-based. Users can top up when they run out rather than paying for unused capacity.
Why this matters for marketers
Marketers spend too much time on admin and not enough on strategy. Soma's approach signals that the barrier to building domain-specific AI platforms has collapsed, but the real differentiator is whether the builder understands the job. For marketing leaders evaluating their stack, the question isn't just which tool to buy - it's whether the tool thinks like a marketer or a general-purpose model trained on the internet. As AI becomes embedded in marketing workflows, AI for Marketing shifts from a vendor choice to a core capability decision. For those leading that transition, an AI Learning Path for CMOs can provide the strategic grounding to separate domain-aware tools from generic ones.
Your membership also unlocks:
