June 2025 AI/ML Spotlight Roundup

It's always cool to see 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 May and June 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

Proactor

Website: https://proactor.ai/

Their tagline: Proactor AI listens in real time, proactively identifies needs, and acts before you ask.

  • What They Do: Proactor uses real-time conversational AI to monitor meetings and interactions, automatically identifying tasks, summarizing key points, creating action items, and initiating workflows proactively before explicit instructions are given.
  • Why They Stand Out:
    • Colin: I genuinely like the UI. I think all these notetaker-like apps need to evolve, and this is definitely the step in the right direction.
    • Jonny: Proactor’s proactive approach feels genuinely different from other AI tools/agents coming out. Kind of one of those things where it’s like, “yeah, that makes sense that AI should be able to do that”. It’s like having an assistant that reads minds and executes before you even realize you need something.
  • Concerns:
    • Colin: I’m at the point where I’m concerned about what skill becomes “unused” and therefore deteriorates over time with these things. What happens if a small throwaway line is important, but ignored by notetaker apps? This isn’t a critique of Proactor at all, but just a general observation about increasing reliance on AI tools.
    • Jonny: I do wonder about accuracy and how the tool will actually feel in day-to-day use. Could Proactor sometimes jump the gun based on casual mentions or misunderstood context? Will it seem weird/creepy when it starts suggesting things that are unexpected?
  • Future Outlook: If Proactor nails accuracy and context understanding, it could become the standard for AI agents.

Harvey.

Website: https://www.harvey.ai

Their tagline: Professional Class AI

  • What They Do: Harvey is an advanced AI platform specifically designed to assist legal professionals with tasks such as drafting documents, reviewing contracts, conducting deep legal research, automating due diligence, and managing complex legal workflows.
  • Why They Stand Out:
    • Colin: Now this is cool. I particularly like the bit where you can delegate complex tasks in natural language to a domain-specific personal assistant. This is a major win for UX.
    • Jonny: Harvey really nails it as a specialized AI built just for legal teams. Kind of like a super-efficient junior associate who actually loves paperwork. If it can consistently speed up tedious legal processes like contract drafting and deep research, it's a win that I think a lot of businesses would celebrate.
  • Concerns:
    • Colin: Who is accountable for mistakes? That’s a real question. If you hire a real law firm and they make major errors, that’s on them. If you use Harvey, will they assume responsibility? A quick glance at their legal page doesn’t give me an answer here, and that’s possibly by design?
    • Jonny: On the flip side, there's zero margin for error in law. Mistakes here aren't just inconvenient, they can be legally risky. Harvey needs flawless reliability or it could quickly become a liability. There also might be some push back from the legal industry so as to not disrupt it too greatly.
  • Future Outlook: Harvey is positioned to be the leader in legal AI and redefine productivity standards in the legal sector if it maintains high accuracy and reliability.

AlphaGenome from Google Deepmind

Website: https://deepmind.google/discover/blog/alphagenome-ai-for-better-understanding-the-genome/

Their tagline: AI for better understanding the genome

  • What They Do: AlphaGenome applies DeepMind’s AI expertise to genomics, accurately decoding genetic information, predicting disease risks, and unlocking personalized treatment options through advanced data analysis and modeling of DNA structures.
  • Why They Stand Out:
    • Colin: We’ve gone from AI for coding to AI for understanding DNA, the basic coding blocks of life.
    • Jonny: AlphaGenome is a classic DeepMind power move. Using AI to unlock the mysteries of our DNA has the potential to revolutionize personalized healthcare and genetic research. It’s some real sci-fi stuff. With all the buzz around AI agents, I think people sometimes forget that AI is enabling some major technological and scientific discoveries.
  • Concerns:
    • Colin: Please do not upload your own DNA sequence to Google’s API. Remember 23 and Me?
    • Jonny: Handling genetic information brings serious ethical responsibilities. DeepMind needs to tread carefully around privacy, ethics, and consent because one misstep could overshadow the enormous benefits.
  • Future Outlook: If ethical and privacy considerations are handled thoughtfully, AlphaGenome could be at the forefront of a radical transformation of personalized medicine and genetics research.

Workflow 86

Website: https://www.workflow86.com/

Their tagline: Automate at the Speed of Thought

  • What They Do: Workflow 86 is a no-code automation platform that enables users to quickly design, deploy, and manage business workflows. It provides intuitive visual tools to build sophisticated automations without extensive coding knowledge.
  • Why They Stand Out:
    • Colin: This is like Zapier meets LangChain. Using AI as an expert to help you set something up is one of the best uses of AI. I need more of these. The UI is quite funny to see in action and there are quite a few integrations out of the box.
    • Jonny: Workflow 86 feels like approachable and accessible vibe-coding for businesses, enabling literally anyone to turn their ideas into working automations within minutes. This will enable a lot of business folks without a technical background but with an operations/productivity mindset to automate workflows easily.
  • Concerns:
    • Colin: None at the moment.
    • Jonny: My only reservation is scalability: it’s awesome for quick-and-easy automations, but can it smoothly handle complex processes without hitting performance snags?
  • Future Outlook: With the growing popularity of no-code platforms, Workflow 86 could be a major player if it maintains ease of use and handles larger-scale workflows gracefully.

Artifacts from Anthropic

Website: https://support.anthropic.com/en/articles/9487310-what-are-artifacts-and-how-do-i-use-them

Their tagline: N/A

  • What they do: Artifacts from Anthropic allow users to capture, save, and easily revisit important conversations with AI models. By creating shareable snapshots ("Artifacts") of specific interactions, you can document reasoning, share AI-generated insights across teams, and maintain a clear, organized record of your AI-powered workflow.
  • Why They Stand Out:
    • Colin: I’ve been using AI so much that using the search function is no longer helpful given some conversations might span multiple windows or have multiple results.
    • Jonny: Artifacts is solving a real-world problem for anyone who regularly uses conversational AI, keeping track of valuable insights. It's essentially giving your AI interactions a "save button," helping you to easily share, organize, and refer back to important AI conversations.
  • Concerns:
    • Colin: None at the moment, but it does make me think about all the possible lost conversations on hosted solutions if these things ever experience a cyberattack.
    • Jonny: The biggest hurdle might be managing the growing collection of saved Artifacts effectively. Without solid organizational tools, users could quickly get overwhelmed by too many snapshots.
  • Future Outlook: Artifacts could become a critical standard for collaborative AI workflows if Anthropic develops intuitive organizational tools that scale smoothly alongside frequent usage.

Higgsfield Canvas

Website: https://higgsfield.ai/edit/canvas

Their tagline: Paint products directly onto your image with pixel-perfect control

  • What They Do: Higgsfield Canvas provides AI-powered visual editing tools specifically tailored for digitally integrating products into existing images. It ensures pixel-perfect realism ideal for marketers, e-commerce, and creative teams.
  • Why They Stand Out:
    • Colin: Quite a few AI products don’t give a semblance of direct control (looking at you, image generators). This is a great step in the right direction when you absolutely want the final product to have a human’s fine-tuned touch.
    • Jonny: Higgsfield Canvas might just end the exhausting daily grind marketers face with image editing. The ability to select and modify specific parts of an image right within the interface is a game changer especially when you have an image that's nearly perfect but just needs a slight color tweak or object adjustment. This granular level of control is something I haven't seen executed this well in other image-editing tools.
  • Concerns:
    • Colin: None right now! It’s great!
    • Jonny: My main worry is the ability to stay ahead on this. With so many other image gen services, I’d like to see Higgsfield continue to innovate with this Canvas platform.
  • Future Outlook: If Higgsfield Canvas consistently delivers realistic, reliable results, it could dominate image-editing workflows in digital marketing and e-commerce and even compete with the likes of Adobe and Canva.

Dia by The Browser Company

Website: https://www.diabrowser.com/

Their tagline: Chat with your tabs

  • What They Do: Dia is a browser extension using AI to manage multiple open tabs by enabling conversational interaction. Users can quickly summarize, organize, and extract insights across multiple tabs through simple chat-based commands.
  • Why They Stand Out:
    • Colin: Oh man I always wondered why this hasn’t been a thing. I am the type to have tens of tabs open at once, and sometimes I just want to know what’s going on!
    • Jonny: Dia tackles modern browsing chaos head-on. I’ve said it before, but I think agents that work directly in browser are going to be big. Being able to chat directly with my tabs instead of manually sorting through dozens of them sounds refreshingly intuitive exactly what my cluttered browsing sessions need.
  • Concerns:
    • Colin: The usual concerns with security, but overall this is a very good concept!
    • Jonny: Accuracy in summarizing and maintaining context across multiple tabs might be challenging. If Dia struggles with contextual nuance, it risks becoming just another underused browser extension.
  • Future Outlook: If Dia manages to nail context understanding reliably, it could significantly transform the everyday browsing experience.