It's always cool to see the AI products enabled by advances. This AI and Machine Learning Startups roundup list is to highlight disruptive startups, emerging technologies, and niche breakthroughs in our space. The list contains companies we talked to or explored for our own usage throughout the month of April and May 2025.

Disclaimer: None of these entries are paid-for-advertisement or sponsored —this roundup is simply GMI Cloud's commentary on what excites us. The companies or projects named were not involved in creating this listicle, and this may be the first time they have heard about it.

Know a company that should make our list next month? Let us know @ [email protected] !

AL/ML Spotlights

Merlin AI

Website: https://www.getmerlin.in

Their tagline: Ask AI to Research, Write, Summarize in 1-click

  • What They Do: Merlin AI is a versatile Chrome extension designed to streamline your digital tasks. It assists with researching, summarizing content, drafting emails, and creating social media posts all within your browser. Whether you're working with articles, videos, or PDFs, Merlin AI helps you process and generate content efficiently.
  • Why They Stand Out
    • Colin: The AI extensions are starting to make their way into life. Companies have created AIs and now their biggest problem is making it seamless to use the AI. Merlin gets the job done and is taking the right step to make sure their AI is easy to adopt.
    • Jonny: The Chrome extension is awesome for this type of use case. Instead of having to paste text or url into another AI tool, Merlin can just work off of what you’re already seeing in browser whether that be an article or YouTube video. This reduction of friction makes the tool much more usable, in my opinion.
  • Concerns
    • Colin: Be aware that for Merlin AI to work correctly, it necessarily needs to track you. Their privacy policy makes it obvious that you have to opt out of having your data being used for retargeting purposes. I do wish it was opted out by default. 
    • Jonny: As of now, there’s a limit to how many queries you can submit on the free tier. 
  • Future Outlook: As AI assistants become more embedded in daily workflows, Merlin is well-positioned to grow alongside browser-based productivity tools. Its broad utility across research, summarization, and content creation makes it a likely fixture for knowledge workers, students, and digital marketers.

Balloonary

Website: https://balloonary.com/

Their tagline: Easy AI ads, serious growth.

  • What They Do: Balloonary helps small businesses and solo entrepreneurs quickly launch ad campaigns across major platforms like Google, Facebook, and Instagram using AI. You just describe your business, and it builds the campaign for you.
  • Why They Stand Out
    • Colin: Not every business owner or team has the skillset or in-house talent to make good ad campaigns. It's interesting to see AI with niche use-cases trying to solve these talent gaps. 
    • Jonny: Balloonary stands out by narrowing AI’s broad text-geneartion capabilities into a focused, high-impact use case: ad creation. This kind of purpose-driven tool is exactly where generative AI can shine and bridges the gap between content generation and business outcomes.
  • Concerns
    • Colin: Tools that simplify the process of creation are good for creators, but give a false sense of advancement for those who aren't in the space. While "good enough" is often a desirable benchmark, the actual expertise and experience required to run good ad campaigns comes with intangible metrics like "taste" that I doubt the person using the AI can adequately mimic. 
    • Jonny: While Balloonary simplifies ad creation, its success hinges on real campaign performance. Since its focus is on smaller businesses, users may outgrow it if they need deeper optimization or more functionality.
  • Future Outlook: With more small businesses shifting online, demand for easy-to-use ad platforms is rising. Balloonary could become a key player in democratizing digital marketing, especially if it expands support to new ad platforms and refines AI-driven performance optimization.

Higgsfield

Website: https://higgsfield.ai/

Their tagline: Create ready-to-share product videos from a single photo

  • What They Do: Higgsfield is an AI video tool that lets users generate expressive motion and facial animations from a single photo or short video. It's aimed at creators looking to make fun, stylized content quickly.
  • Why They Stand Out
    • Colin: The speed of generation feels markedly faster than other ones I've played with. I haven't actually tested with benchmarking in mind, but this is great in an era of consumer demand for immediacy.
    • Jonny: Higgsfield brings a fresh spin to AI video creation by focusing on short-form, ad-ready content with standout features like motion controls and stylized templates. This makes it incredibly appealing for brands looking to produce eye-catching social media videos quickly and without heavy production resources.
  • Concerns
    • Colin: I don't have any concern with Higgsfield itself, but just wanted to voice a generic concern: how do users differentiate these platforms? Price? Quality? Speed of generation? Ease-of-use? It bears further thinking for what we'll see in a few years down the line.
    • Jonny: The space for AI-generated video is heating up fast, with big players like Runway and Pika pushing the envelope on quality and control. Higgsfield will need to differentiate itself through either superior results, ease of use, or niche focus to stay competitive. 

Future Outlook: As short-form video content dominates online engagement, Higgsfield’s intuitive motion tools may attract content creators, marketers, and educators. Its creative-first approach aligns well with AI video’s shift from novelty to mainstream production tool.

Suno

Website: https://suno.com/home

Their tagline: Make any song you can imagine

  • What They Do: Suno lets you generate original music tracks with AI. Just type in a vibe or prompt, and it creates full songs. It’s built for creators, streamers, and anyone needing custom music without the licensing headache.
  • Why They Stand Out
    • Colin: So I've actually messed around with Suno since they started, but in the beginning I could still hear where the AI is being weird with the music. It was still very much at the level of being close to good, but being weird. Now, it's actually getting extremely hard to figure out if it's AI-generated or real, and this is really cool.
    • Jonny: As a self-proclaimed music snob, it's rare for AI-generated music to impress me but Suno genuinely does. It's leading the pack in AI music generation, producing full songs with vocals and structure that sound surprisingly polished. For creators and hobbyists alike, it unlocks a new way to produce music without needing instruments or a studio.
  • Concerns
    • Colin: I have the same refrain. Creator tools will empower those who know how to create, and give a false sense of competence to those who don't know how. I have no worries about skilled experts using these tools, but I caution those who want to get into music-making (or any artistic endeavor) to lean too much on AI as they set out in their journey.
    • Jonny: As the space grows, Suno will face increasing scrutiny around copyright, originality, and the potential for misuse (e.g., deepfake vocals). Also, the novelty could wear off unless the platform continues to evolve with more creative control and personalization for users.
  • Future Outlook: AI-generated music is entering the creator economy in a big way, and Suno is at the forefront. With continued improvements in style fidelity and licensing clarity, it could become a go-to tool for influencers, game developers, and indie filmmakers looking for custom soundtracks.

Figma AI

Website: https://www.figma.com/ai/

Their tagline: Your creativity, unblocked with Figma AI

  • What They Do: Figma AI brings generative and assistive AI into the design workflow. It helps you speed up ideation, layout, and iteration so designers can move from concept to prototype faster than ever.
  • Why They Stand Out
    • Colin: Figma has been a useful and popular tool for many years, so it's cool to see how they've folded AI into their feature-set. This is useful for any prototyping workflow, pretty intuitive, and doesn't confuse me at all (a relevant, personal benchmark). 
    • Jonny: Figma AI stands out because it's already proving its value in real workflows. Designers are actively using its tools to accelerate wireframing, ideation, and visual polish. They’ve shown they can ship good products and hopefully that can convince users to continue using them and their new AI features.
  • Concerns
    • Colin: Cost. Figma AI is currently free during beta testing, but all AI has a computational cost attached to it. Enjoy the free beta while it lasts and pay attention to whatever costs Figma ends up attaching to your bill for any AI features.
    • Jonny: The big question is whether Figma can keep innovating at this pace. Now that they've set a high bar, staying hungry and continuing to push the boundaries of design + AI will be essential to avoid stagnation.
  • Future Outlook: Figma’s AI is poised to become a design copilot, reducing friction in brainstorming and iteration. As design teams face mounting pressure to deliver fast, context-aware AI features could become standard, making Figma an even more dominant collaborative design platform.

Lovable

Website: https://lovable.dev/

Their tagline: Idea to app in seconds, with your personal full stack engineer

  • What They Do: Lovable is a text-to-app tool that turns simple prompts into working full-stack web apps. You describe what you want, and it handles the frontend, backend, and deployment all through AI.
  • Why They Stand Out
    • Colin: This is… kind of like what non-coders really wanted. It's almost there too. You describe what app or project you want to create, and the Lovable app creates it all and then hosts it for you to test it out. It's actually incredible to watch a simple description I type get turned into MVP that I can actually interact with in the nearby window.
    • Jonny: Lovable is tapping into the vibe-driven coding trend where aesthetics, speed, and shareability matter. You can see what others have made, remix templates, and build real full-stack apps quickly. It’s fun, community-driven, and lowers the barrier to getting from idea to execution
  • Concerns
    • Colin: To get good at Lovable you absolutely need to be very prescriptive with how you describe the project in mind. This is one of the hilarious catch-22s: part of good software engineering is understanding client requests.
    • Jonny: Full-stack generation is ambitious, and there's a risk of overpromising. 
  • Future Outlook: As demand grows for no-code and low-code development, Lovable’s promise of turning text into full-stack apps is compelling. Its success will depend on scaling integrations and managing app complexity, but its potential to reshape startup prototyping and MVP launches is strong.