Episode 113: Building the Future of AI Recruiting with Jeff Fish Intelo Founder | EP113

When you're at this stage of the company, everyone you hire is a critical hire, whether they're in product, engineering, sales, or customer success, whoever we hire is critical.

Within Intello, we eat our own dog food, so everyone is using AI tools all day long, and if you're not using Gemini or Chat GPT every hour on the hour and you're in sales, you're probably not going to fit here.

If you're in marketing and you're not using a bunch of marketing tools that are AI, you're not going to fit here.

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 Jotform and featuring our 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 Jeff Fish, the co-CEO and co-founder of Intello.ai. How are you doing today, Jeff?

I'm good, Demetri. Living the dream. Always have to live the dream.

Just to kick things off, tell me a little bit about your background. How did you get into AI and what got you excited enough to be running Intello?

I've been in tech for 20 plus years. I started out in e-commerce, then technology in talent acquisition, and for the last 10 plus years I focused on China where AI has been growing exponentially.

I started a company in 2015 called Chatley, a WeChat management platform, which we ran for about 5 years until we were acquired by Salesforce. Then I ran Salesforce China for about 5 years.

During that time, my co-founder and co-CEO Rupesh had started Intello in stealth mode and came to me asking if I wanted to leave my job at Salesforce and run this with him. He runs product and engineering, I run the business.

He convinced me over about a year to join this adventure. We had worked together before and it was an opportunity I couldn't say no to, especially to get into agents purpose-built for very specific problems to solve.

We're focusing specifically on merchandising and planning and building agents just for retail.

That makes a lot of sense. What's it like having a co-founder and co-CEO? I'm an owner of a company without co-founders, so I'm curious about your perspective.

It's the second time I've done it. The first time was similar: I ran the business and my co-founder ran product and engineering. As long as the two leaders understand their roles and collaborate on decisions, it works well.

Some friends have had challenges with two people making decisions, but you have to find the right person with similar beliefs and values to ground the company.

We follow a V2MOM structure similar to what I did at Salesforce, and Rupesh and I had almost line-for-line aligned values and vision for the company and team.

If you have that alignment, you can be really successful. Differing values and perspectives can be very challenging.

I think it's great you can set boundaries and have different skill sets. What are your respective backgrounds?

Rupesh is a retail technology specialist from consulting and led retail for Sapiant. I came from e-commerce originally and worked for him in a product role while he was CEO.

We have similar capabilities but this product is his baby and I have skills in scaling. Together, he focuses on product and I focus on getting it to as many customers as possible.

You recently closed a $2 million funding round from Illuminate Ventures, known for finding game-changing companies. How did that partnership come about and what was pitching like?

We weren't planning to raise for another six months, but when I announced leaving Salesforce, Cindy Padnos and Jennifer Savage from Illuminate reached out. Cindy was a mentor from my Chatley days.

We gave them an overview of the product and pipeline, and they wanted to be involved immediately. They help us with big decisions as board members and align with our values and long-term vision.

As we scale our teams and product, we rely on them and will bring in additional investors like them for future rounds to build something special and lasting.

How many people are at your company now?

We're 30 now, mostly product and engineering. We're focused on scaling product and adding customers at a good clip. We're globally distributed with team members in New York, Dallas, California, Italy, Hong Kong, and Kochi, India.

How was pitching your AI agents platform and how are you different in the retail merchandise realm with AI?

There are agents for everything now, but not highly specialized agents for merchants. Merchandisers and planners spend a lot of time in spreadsheets, which are their tool of choice despite legacy and modern systems.

Chief merchandising officers say it's a lot of art, science, and math to pick trends and plan inventory, balancing warehouse and store stock, especially around events like Black Friday.

Our biggest competitor is spreadsheets. The main problems are efficiency—spending 50-60% of time in spreadsheets limits strategic thinking—and getting the right inventory to the right store at the right time.

Existing tools are rigid and lack flexibility to adjust constraints. Many use blackbox AI, but verticalized agents for specific problems allow reasoning and act like additional staff.

Organizations have limited operating expenses but flexibility in software. Our agents provide collaborative intelligence, working alongside merchandising, financial planning, allocation, and strategic planning teams.

They act as teammates, not replacements, delivering better results more efficiently.

What are the main concerns or questions you get asked about your tool? Do clients worry about job replacement?

The fear that AI agents will take jobs is overblown. An MIT study showed 95% of AI implementations in Fortune 500 companies failed. This is a new industrial revolution requiring change management.

Agents excel at pattern recognition, repetitive tasks, and workflows—tasks merchandisers don't want to do. They want to think strategically and creatively, working with product and finance teams.

Agents help teams do what they want to do, leading to higher sell-through, more efficient teams, and longer-term planning.

In some cases, low-level customer service agents might replace jobs, but agents help people level up to higher-value roles.

In our space, agents enhance rather than replace because operating expense budgets don't expand rapidly.

If-then logic is big with AI, but the human aspect and creativity in retail planning can't be replaced by agents.

There's more to retail than deterministic unit-based operations; trend analysis and creative decisions require human input.

Elon Musk predicted a utopian society without work, but I don't believe that will happen in my lifetime or in our space anytime soon.

Milan Fashion Week exemplifies the creative, ethereal nature of fashion that agents can't replicate.

How have reasoning models and AI improvements helped with strategy? Do they assist humans by doing leg work?

Reasoning is key. Early on, our reasoning was technical, based on data like historical sales and inventory, but customers found it too technical.

We improved reasoning engines to have real business conversations that feel like talking to a person, which increased adoption and success.

Better reasoning differentiates our agents from legacy ML models and early LLMs. For enterprise customers, data must be safe, secure, and scalable.

The magic happens when users realize agents won't take their jobs but will make their jobs much better.

How do you approach working with clients to determine the best AI agent solutions for them?

We focus on very specific problems in a limited vertical. We have about 20 agents solving 20 specific problems in retail merchandising, financial planning, line planning, and allocation.

We are a foot wide and a mile deep, focusing deeply on these problems rather than broadly.

Our agents talk to each other and can integrate with other enterprise agents like Salesforce, ServiceNow, and Workday. CIOs will orchestrate multiple agents from different brands on the same protocol.

We build agents based on customer patterns and needs, adding new agents to the pipeline as required.

I like the phrase 'a foot wide and a mile deep' because agents and AI generally don't excel at generic tasks but do well when focused.

A big misconception is that AI agents can do everything well generically, but specialization is key.

Many retail executives don't fully understand what AI agents do, even though they know they need AI.

CEOs expect AI deployment or risk going out of business in five years. But many companies don't know what to do beyond that.

The MIT study showed many companies dipping toes in AI without clear goals or KPIs. Now the focus is on deploying AI use cases that deliver real ROI.

Are we in an AI bubble? There are many bubbles discussed, such as Nvidia valuations and data center build-outs. We're early in the technology curve with a long way to go.

Is there an evaluation bubble? OpenAI and Anthropic are highly valued private companies, Nvidia is highly valued publicly. The next phase is application layers with winners and losers.

There may be overvaluations, but that's typical in tech cycles. I'm not a financial analyst, so I don't focus on that.

Your approach of focusing narrowly and deeply makes sense because major LLM players aim to be generically good but not deeply specialized.

What's your long-term vision for the company in this industry?

There are many retailers globally needing to move products and plan inventory. We have a long runway to grow, currently focused on luxury and specialty retail.

We'll expand into big box, hard goods, grocery retail, and regions like Asia-Pacific where I've spent 10 years.

The total addressable market and serviceable available market are large, and we'll revisit growth in three years.

When you launched the AI Agents podcast, few people knew what an agent was. How did you prepare to enter this new market and educate people?

We came out of stealth in January. Major players like Salesforce, ServiceNow, and Microsoft laid the groundwork for enterprise understanding of agents.

If we had launched a year earlier, no one would have known what purpose-built agents were, but now awareness is widespread in Fortune 500 and 1000 companies.

Podcasts and discussions help spread knowledge, making it easier to explain and sell agents.

Was it difficult to explain agents to customers initially?

There was a lot of fear among merchants that agents would take their jobs. You have to dig deep and show agents are here to help, not replace.

These conversations still happen frequently, and you have to demonstrate your product is helpful, not like Skynet.

Do you think fear of AI will dissipate? When?

Technology transformations always cause fear. When cars came, buggy drivers were fearful. This will happen here but will dissipate over time. I can't predict when.

Movies like I, Robot create sensationalism and dystopian fears that aren't realistic. We should approach AI practically.

Much fear comes from sensationalism and movies, but those are just stories. We won't know real impact until it happens.

Is it difficult to plan hiring? Do you consider AI usage when interviewing?

This is my second startup. You have to be willing to work very hard. Every hire is critical, whether product, engineering, sales, or customer success.

At Intello, we eat our own dog food. Everyone uses AI tools all day long, and if you're not using Gemini or Chat GPT hourly and you're in sales, you probably won't fit.

If you're in marketing and not using AI marketing tools, you won't fit. If you're an engineer and haven't used cursor or Claude, you likely won't pass the first interview.

We have Slack channels tracking tool usage and measure it weekly. Tool usage is monitored closely.

If AI usage isn't in your DNA, this isn't the place for you. You have to embrace it quickly or get out of the way.

You must build a team aligned with the company's ethos and future projection. Many companies say AI won't take jobs but don't follow through.

We practically measure AI usage. In interviews, if candidates aren't interested in leveraging AI tools, they don't advance.

What are your favorite personal AI tools?

I used Gemini Pro 3 this morning, which is awesome. For presentations, I use Gamma, building custom decks quickly.

We also use cursor and Claude. I don't have a single favorite since I use them all, but Gemini is used most because we're a Google shop.

Internally, when creating agent tools, do you bounce between models?

We built our own reasoning engine and workflows and integrate with all major foundational models. We're not focused on any one model.

Some customers use DeepSeek on Azure for cost-effectiveness, others use OpenAI. Our multi-tenant strategy includes Gemini, Claude, and OpenAI.

We're agnostic on foundational models. There's room for all to be great.

Which model is best for research? Gemini's latest release is the best I've seen. It's very solid.

Earlier Geminis and Bard were bad, which hurt adoption. GPT-5 was overhyped and underdelivered, but 5.1 was solid.

Sam Altman acknowledged the misstep with humor. Setting expectations properly is important.

Outside of LLMs, what AI-augmented tools do you like?

I'm a big fan of Text Cortex for content and Crisp AI for noise cancellation during calls, which transcribes and reduces background noise effectively.

Meeting transcribers are underrated, especially for industries wary of recording calls due to security concerns.

What are your favorite non-model tools?

I love Gamma for presentations. Years ago, building pixel-perfect PowerPoints was tedious, but now you can create killer decks in minutes with prompts.

Is there anything else you'd like to plug besides your product?

Look at various agents specific to your problem. Don't try to solve everything with one big platform. We're part of Salesforce's Agent Force partner network, and others will have similar offerings.

Where can people learn more about you?

Visit intell.ai or reach out on LinkedIn. Intell.ai is the fastest and easiest way to contact us.

Make sure to like this video on YouTube, leave a review on Apple Podcasts and Spotify, and share it widely to help people understand AI agents like the cool ones we're building at Intello.ai.

Thanks so much for watching and we'll see you in the next one. Peace. Thanks, Demetri.