Building BD with AI: A Conversation with Luka Mrkic, Head of BD at Espressio

Luka Mrkic spent six years in tech — across telco, ecommerce, and web3 — before joining Espressio as Head of Business Development. He’s not a developer, but he’s built his own AI-powered BD system from the ground up using tools like Clay, OrangeSlice, and the Anthropic stack.

—

We sat down with him to talk about what it really means to build business development with AI: where the technology helps, where human relationships still matter most, and what every skeptic should try first.

You’ve built your career on relationships. what pulled you toward an AI-first BD role?

Luka: I've been working in tech for last 6 years (telco, ecommerce, web3), and I’ve been using ChatGPT since it came out. I'm not a developer or a tech person, besides some HTML and CSS I can't write any code. But with agentic systems and especially the Anthropic stack, I saw it as an incredible opportunity for everyone, including non-tech people, to start building.

Of course, not everyone needs to build unicorns, and surely they won't. Small wins are what matters. Agents and workflows that make your life smoother, more organised, more efficient. Both professional and personal.

And by doing that you actually get to learn about all these things you can build right now. Karpathy said: The hottest new programming language is English. if that doesn't get you to open a terminal, I don't know what will.

Where does AI actually change the BD playbook, and where does it still come down to the human?

Luka: BD is about trust. People need to see you, hear you. You need a chance to crack a joke and build rapport. AI can't replace that. What it can replace is everything around it. Prospecting, research, outreach, proposal prep. All the repetitive stuff that eats most of your week. You streamline that, and you actually get to focus on the one thing that matters: conversations and relationship building.

You’re now running BD with a tool that automates the research and proposal side. what does your day actually look like?

Luka: We called it RevenueOS, and I keep wondering where was this my entire career ha-ha. It handles pre-call research, meeting intelligence, same-day proposals. What used to take me a full day of prep now takes a fraction of that, and the rest of the time goes where it should. Conversations, follow-ups, closing.

Loading tweet...

View Tweet

A lot of sales people think AI makes outreach feel generic. how do you keep deals personal at scale?

Luka: That's the double-edged sword. If you use spray and pray, even with a bunch of AI automations, it's not going to work. People can feel it. You need to set up a system that spikes the relevance of your outreach.

The whole point of BD is connecting your value to your prospect's pain points. With tools like Clay or OrangeSlice you can do that at scale without sounding generic, and then you train your agent on your voice, how you write, how you talk. The output should read like you wrote it yourself after doing the research.

When the conversation starts, that's where you take over. AI does the legwork, you close the deal. And that's what we tell our clients too. We're not replacing your BD team. We're giving them back the hours they waste on prep so they can do what they're actually good at.

What’s the one metric you’re focused on in the first 90 days?​​​​​​​​​​​​​​​​

Luka: Revenue - only thing that matters. It depends on the stage though. If you're early or just starting with an AI-driven BD system, you need leading indicators. Response rate tells you if outreach is landing.

Objection patterns tell you if positioning is off. Meeting-to-proposal conversion tells you if your discovery is working. These are the signs you're on the right track before revenue catches up. 

You've mentioned RevenueOS. What's the bigger picture with Espressio?

Luka: RevenueOS is only one piece of what we're building at Espressio. We help teams scale what they're already good at, whether that's sales, marketing, or ops. Make them more efficient, win back time. That's the whole point.

We actually implemented all of this at Lunar Strategy first. We've been running a web3 marketing agency for 7 years, 300+ projects delivered.

When we started building AI systems internally, we realised pretty quickly that other teams need this too. That's how Espressio was born.

What's the biggest mistake companies make when they start using AI for sales?

Luka: They build on top of a broken process. If your ICP is wrong, your messaging is weak, your follow-up is inconsistent, AI just scales the mess faster. You need to fix the fundamentals first, then automate.

What would you say to a BD lead who's skeptical about AI tools?

Luka: Pick one workflow, the most repetitive, time-consuming thing in your week. Automate just that. You'll know within two weeks if it's real. In my experience, the skeptics who actually try it tend to become the loudest advocates. And don't be afraid of the command line.

You don't need a fancy UI to create amazing stuff. Give it a go, trust me.