에피소드 109: Building Autonomous AI Agents with Omkar Pandharkame Chief Strategy Officer at Supervity

Sad GPT's entire fundamental prospect is to really democratize the use of AI. So why is it only generative AI and why can't we just democratize AI agents? That is our inspiration and ambition to democratize the way people can build AI agents. My grandfather should be able to build an AI agent for himself. My 10-year-old nephew should be able to build an AI agent for himself.

Hi, my name is Demetri Panici and I'm a content creator, agency owner, and AI enthusiast. You're listening to the AI agents podcast brought to you by Jform and featuring our very own CEO and founder Aytekin Tank. This is the show where artificial intelligence meets innovation, productivity, and the tools shaping the future of work. Enjoy the show.

Hello and welcome back to another episode of the AI agents podcast. In this episode, we have Amar, the chief strategy officer.

Hey, I am doing fantastic. Thanks so much for having me over here on your podcast and yeah, looking forward to our chat.

Yeah, me too. I'm definitely looking forward to it. I'm really excited to hear more about what you are doing individually and what everyone over at Supervity is doing. So just to kind of give us a baseline understanding of yourself and the company, tell us a little bit about you, how you got into the world of AI and what you would say Supervity does.

Sometimes I get into an imposter syndrome zone because I don't belong to the world of AI by education. By education, I'm an organizational psychologist. I have studied law and economics but now I'm managing product and strategy for an AI first company. It was not a very natural progression for me but before Supervity I was an academician, a professor of entrepreneurship. Gradually before that I launched a couple of tech startups which were fairly successful.

A very interesting insight is that everything happening pre-GenAI I thought this is it but once the GenAI boom started, it was a no-brainer to leave a cushy academic job and really challenge the frontier of what's possible with GenAI. That's how I came into the whole scene of AI. I still remember a few months ago when VIP coding started getting mainstream, the tech startup I built about 10 years ago took six months and a 20 people dev team. I could write code on a lazy Sunday afternoon with some pizzas and beers in about 3 hours. That was an awakening for me about a year and a half back.

I joined the founding team over here at Supervity where I'm currently managing product and strategy. Strategy is such a broad abused word. It means doing everything from hiring people, pitching to investors, to making sure the bills are paid.

Supervity identifies as an AI employees company. It initially started as a pioneer in digital workers before AI employees became the term. Digital workers evolved into eworkers, AI agents, and now everyone talks about AI employees or AI teammates. Supervity started about six years ago as a services company but we reversed the model by saying we are service as software. We give a platform but also service on top of it. It's not pure SaaS or PLG motion where you just download and start using it. The idea is to give digital employees that become part of your workforce, working alongside you. We help develop and configure them.

This has been the most amazing year so far with great tailwinds. Every customer and large enterprise we talk to wants an AI employee right now in accounting, finance, HR teams. It's been a beautiful journey and this is just the start.

What do you think are the primary areas companies find AI can help them? What roles are you seeing as most interesting?

One interesting example is the accounts payable AI employee. In 2025, accountants spend so much time doing three-way, four-way, five-way matching. You imagine an accountant's office with heaps of paper and red lines showing mismatches between purchase orders, invoices, and contracts. These processes should have been automated yesterday. RPA bots tried but were expensive and failed royally, like having an elephant as a pet—big but no utility and expensive to maintain.

Enterprises are excited about our APAI employee launched last month because unlike traditional AI software with SaaS baggage, our AI employee is conversational. Clicking is dead; it's all conversational. Every morning it sends a message about invoice processing stats, problems, and learns vendor issues to suggest actions like sending emails or having conversations. This is how an accountant would talk and that's what an AI employee should do. We see a lot of traction in accounts payable.

Another use case is the recruiter, the super recruiter. It's easy to shortlist CVs based on job descriptions but a recruiter's job is end-to-end: sourcing, shortlisting, scheduling interviews, sending screening questions, making offers, collecting documents, and answering candidate questions before onboarding. Enterprises want to hire thousands with limited recruiters and we offer one super recruiter AI employee to handle it all. These are use cases with fantastic traction.

Very cool. Speaking of cool use cases, how does it actually work? You mentioned removing SaaS baggage to make it more conversational. What are you trying to do?

We compete with big players like NA10, Crew AI, and Lindy from a platform perspective. They are developer-first, not business-first. If a business person tries to create a workflow on NA10 or Crew AI, they struggle with nodes and connections. Our platform is designed so you talk to AI like an AI coach. You describe what you want and workflows get created without moving anything. You can chat and even use whisper flow to talk to the workflow to approve or reject it. It's very user-friendly.

I like to describe Supervity as a kitchen with AI capabilities embedded from the ground up: OCR, RPA, integrations, workflows. You tell us what you want—super recruiter, sales enablement coach, marketing specialist, AI community manager—and we show you how to build it or you can order from a menu card. We configure and deploy it in your environment so the AI employee is ready to work. This puts the customer at the center and avoids complicated tools only developers can use.

That's really cool. What big competitors are you playing against?

We play the David and Goliath story. Competitors include NA10, Crew AI, Lindy, Gumloop. NA10 is the big daddy with a great community but is developer-first. We are business-first, speaking the language of business users who want to build AI agents without going to dev teams. Wipe coding levels the field but hasn't reached the AI agent space yet.

Why do you think it's taken more time for wipe coding to reach AI agents?

It's about choosing your niche and market. Competitors focus on developers because of their mindset. The business world is bigger than the developer world. I come from a business background and don't want to spend hours learning to build simple AI agents. I want to talk to the platform and say build an agent for me.

Yesterday, Google launched anti-gravity, their wipe coding platform powered by Gemini 3. I was excited to try it but it’s still an IDE and not quite there yet. Gemini 3 is impressive but the experience hasn't fully translated to AI agents. OpenAI’s agent kit is complicated. So our inspiration is to democratize AI agents so anyone, even a 10-year-old, can build them.

Your nephew might be tech-savvy though. I hope so. I think accessibility to AI agents hasn’t reached the level it should. Many people interested in AI lack the high competency needed to build good agents. Your goal is to make it as easy as possible for most people to build AI agents that can do full-time work, even without high technical skills.

Absolutely. We build for enterprise customers and partners like PWC, Accenture, HPE, Deloitte who train on our platform and build AI agent solutions. We want to simplify it so anyone can talk to a super coworker AI agent who understands business context and has long-term memory. Current AI agents have short-term memory and lose context when switching. We want AI employees that know expectations, KPIs, tools, and wait for commands.

That is a very good vision. What are some biggest misconceptions about AI agent tools you encounter?

A common misconception is about agent orchestration or multi-agent systems. People expect an orchestrator agent that manages operator agents. Operator agents do single tasks; orchestrator agents manage them. Building this is complex but we have built it. Crew AI has a good system too. People want to see an orchestrator agent but it responds to commands and is made by stitching operator agents together.

Our unique approach is an AI command center where all operator and orchestrator agents reside. The command center is the interface where you talk to the agents. This approach will set an industry standard for AI agent systems in organizations.

Orchestration is key moving forward. Are you familiar with tools like Relevance AI? I like their 16-bit design. It’s smart and nostalgic. I use it for orchestration, like managing emails by routing to specific agents and generating tailored responses. It’s very useful.

You have orchestration down and are removing full-time work with deterministic roles while also helping delegate work. How do you see this impacting job growth and change?

It's tricky. Automating work means automating full-time employment. Amazon recently cut jobs and TCS slashed 20,000 jobs despite being profitable. They are cutting fat, not costs. Repetitive work is fat in organizations. Losing it makes companies run faster. We work with governments and enterprises to repurpose jobs as work is automated. Human minds aren’t designed for repetitive work. That era is ending and a new era of human in command, not human in the loop, is beginning.

Human in the loop informs humans about AI actions. Human in command means humans control AI employees who take more work at the same capacity. We work with growing companies and tell them to empower current staff with AI agents instead of hiring more. A CFO told me in three months they won’t hire anyone new because AI agents will empower existing staff and grow business without increasing headcount. This improves average revenue per employee.

I agree. People don’t quit jobs because they’re overworked but because they feel they have no meaning or aren’t learning. AI agents will enable hyperfocused businesses where small teams grow big by focusing deeply on one thing. Administrative nonsense will disappear and unfulfilling jobs will be replaced by meaningful work.

What is your favorite part of working where you are and the most exciting thing you are working on or releasing soon?

I love sitting customer-facing and understanding real problems. When customers realize AI agents can solve their issues, it feels like magic. It’s like being a magician entertaining a well-invested audience. We have three exciting things coming: launching version 3 of our platform, a conversational AI agent creator called Supernova on January 2nd, our accounts payable AI employee launched recently, and our super recruiter launching December 1st that handles everything from sourcing to onboarding employees.

Do you have a favorite personal AI tool you use daily?

I love meeting recording tools that send transcripts to AI agents to save time post-call. I use Granola, which transcribes audio well even when other note takers are blocked. I also like Whisper Flow for voice commands and Cluso, an Indian startup that creates product demo videos from natural language instructions. Of course, I also use Supervity.

Where can everyone go to check you guys out?

Supervity is at supervity.ai. People often confuse the spelling but it’s supervity.ai. We just launched a Thanksgiving version of our website with Tim the turkey and free gifts like subscriptions and tarot card readings. Visit supervity.ai for surprises.

Thank you so much for your time. Everyone, please check out supervity.ai. Thanks for watching and we'll see you in the next one.