Homebrew makes investments by consensus – it really works as a result of there’s simply two of us. We’ve achieved it this fashion for 2 causes – first, it really works internally given our model of decision-making and respectful however loud debate. Second, it issues to us externally that founders understand it’s at all times Homebrew making the funding – by no means a state of affairs the place certainly one of us was excited and the opposite one didn’t block it. To that finish, backing Regal, now a pacesetter in AI Powered Calls to your companies’ gross sales, help and operations, was fast to mutual settlement. Satya and I had been excited by the imaginative and prescient and the cofounders’ earlier work expertise collectively. Because the world of AI is shifting fairly quickly, I needed to examine in with Regal CEO Alex Levin to ask 5 Questions.
Hunter Stroll: In a current 5 Questions the particular person I used to be interviewing stated to take into consideration your profession as a narrative you’re telling about your self. What story does your profession inform?
Alex Levin: I bear in mind finding out the philosophy and psychology round consciousness in school and pondering I’d pursue academia. At some point I noticed that I may examine my subject for my entire profession and by no means get to the top of what was already recognized. And at that second, it was crystal clear to me that I needed to discover a profession the place I may get to the forefront of what was recognized extra shortly so I may assist contribute as an alternative of feeling like I used to be at all times behind.
After commencement, I dove headfirst into tech startups because it was clear they had been investing the long run day-after-day. I knew that someday I needed to be an govt, and even the highest govt. And I understood that if I didn’t know the way the sausage was made—if I didn’t know the way to construct expertise—I’d by no means be eligible for these high roles. So early in my profession, I prioritized being near engineering groups, educating myself the way to construct software program, and studying the way to be a product supervisor.
Once I got here throughout Marc Andreessen’s article, “Why Software program is Consuming the World” in 2011, it felt like a vindication of the route I had chosen, however I used to be already down that path.
I’ve skilled each giant and small corporations, and I’ve discovered that I thrive in early-stage environments. It’s like how they describe the skilled leagues of any sport: “issues transfer quicker”. Additionally, there’s extra alternative, and you actually really feel like should you’re not there someday the undertaking doesn’t transfer ahead. I really like that sense of accountability.
I’ve been lucky to tackle progressively senior roles, totally on the industrial facet, and expertise scale, together with reaching $1.5 billion in income at Angi. And now, I’ve taken every little thing I’ve discovered and put it to the check by beginning an organization myself.
Trying forward, I hope my story is one the place I proceed to tackle difficult alternatives, create one thing the place nothing existed earlier than, and construct corporations that actually change their markets. And, ideally, not simply as soon as—however many occasions.
HW: Much like you, I began Homebrew with a former coworker. How did your relationship with Rebecca evolve and do you bear in mind the ‘we’re going to do that!’ second?
AL: Whereas we had been at Helpful, Rebecca and I each had our first children, and I distinctly bear in mind speaking about beginning one thing collectively quickly afterwards. We spent months having basic conversations, however ultimately, we sat down and did a “Founder Relationship Quiz”. We went by way of a listing of 20 to 30 key questions—issues like our working kinds, how huge we needed the corporate to be, what an excellent exit appeared like, whether or not we needed to lift cash, what may go unsuitable, and what we had been most nervous about. After that, it actually felt extra actual as we knew we had been aligned.
It nonetheless took a while to land on Regal as the corporate we had been going to start out as we had two or three different concepts we significantly thought of (which is a narrative for an additional time). However once we talked to potential Regal clients about our imaginative and prescient, and the way we needed to alter the way in which corporations have interaction with their clients, the response was overwhelming. Folks had been pulling for it. And as soon as we felt that “market match”, we jumped in quick.
HW: Regal is an AI-driven firm that was based previous to the ChatGPT launch. How have the previous couple of years accelerated the roadmap? How has it challenged the corporate?
AL: Coming from a background in product and advertising earlier than Regal, Rebecca and I had been used to nice instruments that might faucet into information sources, personalize buyer interactions and even apply machine studying—it wasn’t actually AI again then—to do some fairly sensible issues. In the meantime, involved facilities, groups had been afraid to alter something of their software program for worry it could all break, not to mention attempt utilizing information for personalization or machine studying. So we knew it could be a little bit of an uphill battle to get contact facilities to alter.
From the very starting in 2020 we had a imaginative and prescient of higher business-customer interactions powered by buyer information + ml/ai + software program for groups to make adjustments simply. In 2020 the avant-garde AI was for aiding human brokers, not autonomously dealing with duties, however as we began constructing, we knew buyer information and orchestration needed to be on the core of our platform, hoping that someday, autonomous AI Brokers may take over the shopper dialog as an alternative of a human.
For years, although, AI simply wasn’t adequate. However we stored pushing on it as a result of human brokers trigger points as a result of they produce other motivations, weren’t at all times well-trained, may give up, or simply have an off day. The general public would possibly assume people are the gold normal, however anybody who has truly operated contact facilities with human brokers is aware of it’s truly actually troublesome, and there are many challenges with human brokers.
Then, a couple of yr and a half in the past, every little thing modified. LLMs lastly reached a stage the place they might perceive and generate language and make choices much like people. And we acquired to some extent that our AI agent demo fully opened our eyes as to what was potential. We began specializing in creating an “omni-agent” system, one the place each AI and people may function seamlessly. Construct the insurance policies, scripts, orchestration, guardrails as soon as, and deploy them to each human and AI brokers.
A yr and a half in the past, not many corporations had been prepared for AI. They nervous about hallucinations, buyer reception, and whether or not it could actually work. However AI brokers have superior extremely quick, and the platform we spent 4 years constructing to assist human brokers take the appropriate motion truly gave us the best-in-market platform to function AI Brokers and we now have an enormous benefit over corporations simply beginning to construct AI brokers from scratch. So we’re seeing a seachange at client corporations – everyone seems to be evaluating AI Brokers in 2025.
One of the best a part of this shift is that it’s made our GTM easier. To maneuver human brokers into Regal required ripping and changing their present contact middle software program. Now once we promote Ai Brokers, we roll out with out touching their contact middle software program, accelerating the gross sales course of and implementation.
HW: Content material advertising dominates the enterprise business today, a lot of it about Synthetic Intelligence. Because you’re down on the sphere with the precise corporations shopping for these merchandise, what’s one thing that we’re lacking? Share an ‘earned perception’ that you’ve got from the work at Regal?
AL: Rebecca and I are usually not naturally public folks, so for years, we didn’t focus a lot on advertising. That was a mistake. With our newly constructed advertising staff (who’ve strongly inspired us to get on our recreation and step up), we’re lastly getting in entrance of extra clients, and Rebecca and I are beginning to embrace constructing in public.
One easy earned perception is about the way to begin testing AI Brokers in manufacturing. Each firm needs to start out small. And most gravitate to small use circumstances (like out of hours) to check. I at all times advise in opposition to that as even when that succeeds, it received’t have a cloth impression on the corporate and you will have to start out once more with a bigger use case. As an alternative, choose the most important use case for voice (like your principal inbound or outbound calls) and check your AI Agent on 1% of calls to start out. Then because it succeeds, scale as much as have impression instantly.
HW: What’s your recommendation for founders who could be beginning corporations immediately within the basic AI area – let’s concentrate on those that are utilizing AI to unravel issues for patrons (utilized AI) versus decrease within the stack base fashions or deployment infra.
AL: Rebecca and I are founders and angel buyers within the AI Voice area. It’s important to assume each firm may have entry to the identical LLMs and voices. The problem, then, is to construct an organization that thrives regardless of this actuality. Mentioned in any other case, construct a “thick” workflow or software layer that may present worth it doesn’t matter what the LLM and voices do.
A lot of what Regal focuses on, and what we put money into as angels focuses on role-specific or business particular workflow instruments for generative AI to unravel issues that LLMs alone can not. For instance, integrating AI into buyer information techniques, and enabling actions (like processing funds, updating CRMs, or sending SMS messages). Or making certain AI stays inside its supposed scope, supplies correct info, and acknowledges when it ought to say, “I don’t know.”
These challenges exist throughout industries and roles, and received’t be solved by LLMs alone.
Thanks Alex! To demo Regal, or be taught extra about their product, head to their web site. And if you wish to be part of a quick rising firm that cares not nearly what they’re constructing, however how they’re constructing it collectively as a staff, properly, Regal is hiring.
Get these posts by way of electronic mail
Bluesky has the JUICE – comply with me there
