Interactive AI content company Flam is trying to turn formats such as responsive video, streamed 3D experiences and visual AI agents into repeatable enterprise tools rather than one-off campaigns. The company raised a $40 million Series B in September and is expanding in Southeast Asia through a partnership with Indonesia’s Emtek.
TNGlobal previously covered Flam’s Series B and its regional expansion plans. In this TNGlobal Q&A, Shourya Agarwal, Founder and CEO of Flam, discusses how the company expects enterprises to reuse interactive media, where agentic systems should stop and hand control back to people, and what the Emtek partnership is intended to deliver in Indonesia.
Flam’s Series B announcement talks about moving interactive and agentic content into wider enterprise use. What does a repeatable deployment look like beyond a one-off marketing campaign, and which workflows are customers actually using more than once?
Almost every deployment is repeatable. Brands treat Flam content the way they treat a product or brand video: create it once and deploy it everywhere, across websites, ads, apps, retail and physical media.
The difference is that a Flam experience keeps evolving. Brands update products, offers, languages or creative directly from the backend, and every live placement reflects the change instantly, with nothing to rebuild or redeploy on the front end. The same experience improves over time instead of being replaced by the next campaign.
The term interactive AI content covers a wide range of experiences. What is happening technically when a user engages with Flam content, and what makes the interaction materially different from a conventional video, chatbot or personalized ad?
At Flam, interactive AI content means media that responds live to user input, rather than being fixed in advance like a conventional video or personalized only once like a typical ad. The user is not just watching content. They are continuously sending signals, such as taps, voice, gestures, choices or device inputs, into a real-time system that changes what is rendered, streamed or generated on screen.
A conventional video is linear and identical for every viewer. A chatbot can respond conversationally, but usually through text or voice without changing the visual experience in a synchronized way. A personalized ad is typically tailored before playback based on audience data or targeting rules, but it does not adapt meaningfully during the interaction itself.
Flam sits in a different layer: live, responsive media. The experience changes while the user is engaging with it, often within milliseconds, and those changes happen inside the content itself, visually, conversationally or spatially.
Flicks. Flicks are interactive videos. Instead of watching a fixed clip, the viewer can tap to change what is on screen, such as the product, person or setting. Flicks use Flam’s compression technology to store differences between frames and package multiple versions into a single video, which the company says keeps file sizes smaller and allows instant switching without reloading.
Airboards. Airboards bring interactive 3D to phones without requiring an app download. Flam says its 3D streaming technology packages models, textures and animations together and streams them in real time. Users can interact through touch, voice and haptic feedback.
Visual Agents. Visual Agents are AI-generated characters designed for video conversation with synchronized facial movement and speech. Flam says an agent can be created from a single image, while brands control the source material, responses and permitted behavior. The company uses its own speech, language and video models for these experiences.
Which metrics do enterprise customers care about once the novelty of an interactive format wears off? Are you seeing measurable improvements in conversion, completion, lead quality, sales or another business outcome that customers are willing to use when deciding whether to renew?
Brands using Flam have seen measurable gains across the funnel, from engagement to recall to conversion. According to our internal results, Flam formats deliver three times the dwell time of video ads, a 70 percent lift in brand recall and 25 percent more conversions.
Brands also gain a new kind of data. When someone picks a product color or asks an agent a question, the brand learns what that person actually wants, something a view count cannot show. These signals can feed media planning and product decisions, and they grow more valuable as they build up over time.
Southeast Asia is highly fragmented by language, culture, platform behavior and commerce habits. What has Flam learned about localization that goes beyond translation, particularly as it expands in Indonesia and considers markets such as Singapore?
Translation is the easy part. Our AI models were built multilingual from the start, designed to handle regional languages and mixed-language conversations natively.
The bigger lesson is that brands everywhere want more of their customers’ attention. Every Flam product is interactive and designed to hold that attention and move brand KPIs. We are changing the format of content itself, and that shift works the same way in Jakarta as it does in Tokyo or New York. Brands already know which messaging resonates in their markets. They bring that knowledge to Flam, and our local teams and partners help close any gaps.
How will the partnership with Emtek work in practice? Which parts of Emtek’s media, commerce or digital ecosystem are the most natural starting points, and what would a successful first year in Indonesia look like?
Emtek is our partner in Indonesia, in the same role our partners play in Japan and Nigeria: local distribution and access to brands, agencies and publishers.
The most natural starting points are Emtek’s streaming and digital publishing properties, including Vidio, where interactive formats can run at scale for advertisers who already buy there. A successful first year means 20 or more enterprise brands running always-on deployments, and local creative and sales capability that does not depend on our teams outside Indonesia.
As content becomes more agentic, what actions can these systems safely take on behalf of a brand or user today? Where do you deliberately keep a human or deterministic workflow in the loop?
Brands decide what their agents can do. Today, our customers most often configure Visual Agents to answer product questions, guide comparisons, capture leads, book appointments and hand users to checkout or a human team.
Each brand defines the boundaries: which actions the agent can take, which systems it connects to and when it hands off. For sensitive areas such as pricing commitments, payments, contracts, eligibility decisions and regulated claims, brands typically keep their own deterministic workflows or human review in place. An agent can explain a financial product; the bank decides who gets approved.
Every interaction is logged, so brands can audit what happened and why, and expand what their agents do at their own pace.
Generative systems can introduce hallucination, off-brand responses and regulatory risk. How do you constrain outputs when an interactive experience is representing a company publicly, especially in regulated sectors or markets with different advertising rules?
Brands stay in control of what their agents say. Each brand uploads its own material, such as product data, approved claims and policies, and sets the guardrails for what the agent can and cannot discuss. The agent answers from that material. If a question falls outside it, the agent declines or connects the user to a human.
We have designed our models to prioritize truthfulness and numerical integrity so prices, specifications and figures are stated as the brand provided them. In regulated categories and markets with different advertising rules, brands can tighten guardrails further, and every interaction is logged for compliance review.
What are the biggest barriers to enterprise adoption now: integration with existing systems, content production costs, measurement, brand safety, procurement, or simply convincing teams that the format produces more than short-term engagement?
We are not seeing major barriers at this stage. The technology is already being deployed across enterprises and demand is strong across industries.
If anything, interactive content opens up many possible use cases, so brands sometimes need help deciding where to start. That is where our teams come in, helping each brand pick the use cases that are most relevant to its business.
Flam has talked about Southeast Asia as an expansion market. What role do you expect Singapore to play, whether as a customer market, regional base, partnership hub or source of enterprise demand?
Southeast Asia gives us both reach and scale. Singapore is where many agencies and multinational brands make regional decisions, making it a natural hub for partnerships and accounts. Indonesia, with our partner Emtek, gives us scale.
Looking 18 to 24 months ahead, what evidence would convince you that interactive AI content has become a durable enterprise category rather than another short-lived format trend?
We believe the category is already forming, with more than 100 enterprises live on Flam. The signal we are watching now is permanence: interactive content becoming a standing part of every brand’s content mix, the way video did.
Shourya Agarwal is Founder and CEO of Flam, an interactive AI content company developing responsive video, streamed 3D experiences and visual AI agents.
Editor’s note: This Q&A has been lightly edited for clarity and TNGlobal house style. The substance of the interviewee’s responses has been preserved.
Share your perspective: TNGlobal welcomes contributed insights and expert commentary from across Asia’s technology and innovation ecosystem. Submit a contribution for editorial consideration, or explore more conversations in our TNGlobal INSIDER and TNGlobal Q&A and Interviews archive.