This interview is with Ben Miller, COO at Undetectable AI.

Ben Miller, COO, Undetectable AI

Ben, thank you for taking the time to share your insights with our audience at ChiefExecutiveOfficer.io. Could you start by telling us a bit about your journey to becoming the COO of one of the fastest-growing AI startups? What sparked your passion for AI and led you to this exciting role?

I was already working in tech and utilizing AI at my previous role, which was also a remote company. This gave me a lot of experience managing remote teams, building processes, and measuring results.

Your company's growth, achieving over 8 million users in its first year, is remarkable. Can you walk us through the key decisions and strategies, particularly regarding bootstrapping and leveraging offshore talent, that contributed to this rapid growth?

We built the initial platform on Bubble.io, which is the most robust no-code platform available on the market. This enabled our CTO to single-handedly stand up the entire initial site and immediately scale globally. It also meant a very short initial development time of just a few weeks. For our go-to-market strategy, we utilized a number of channels: organic SEO, TikTok organic growth, but especially helpful was an affiliate program, which gave a percentage of revenue to hundreds of affiliates in exchange for promoting our brand. Also, providing a free tool on our homepage for a very desired service kicked off a viral growth loop that brought us right to the top of Google within a matter of months. Due to being in a new and highly in-demand space, we grew quickly and made the crucial decision to make some key initial hires in the Philippines. Great people who are loyal, smart, hard-working, and immediately helped us scale up to manage a large and growing user base. Because our labor and infrastructure costs were quite low, this allowed us to forgo any outside funding, especially as revenue began to trickle in very quickly.

Building and managing a remote team of 55 employees from scratch is no small feat. How did you approach building culture and ensuring effective communication within a fully distributed team?

This starts from the beginning in a unique interview process. We typically put all new hires through a work test. So, for example, we may hire someone off Upwork and have them do some initial work to see how well we actually work together. I find this to be a much more accurate signal than simply looking at resumes and interviewing people, although that is also part of the process.

Once past this initial test, most of our employees use time-tracking software, TimeDoctor, so that we can accurately track the cost of different projects. This software also tracks work analytics to help us weed out the few bad actors who are not putting in the claimed hours.

In addition, I built out an OKR system to clearly set quarterly goals and track progress every week toward them. This also helps keep us focused on a culture of results, not simply activity levels and time spent.

When projects start to get off task, I review with the relevant team to make corrections and add further support when needed. We have everybody on Slack and do our best to build remote culture by celebrating birthdays and work anniversaries, putting out a weekly positive update on all the good things going on at the company, and similar things. We are also hosting occasional in-person meet-ups to further get to know each other.

As a top manager of this team, I also go out of my way every day to be very thankful and congratulate people for their hard work. Also, when some natural disasters hit the Philippines, we went out of our way to support the employees who were affected. This may seem like a small thing, but apparently many foreign companies do not do this, so it bought us a lot of praise and loyalty internally.

Also, our employees are happy because they can work fully from home and many of them are on flexible hours. As long as the work gets done, they can go pick their kid up from school or any of the other day-to-day things that people need to get done. And of course, as a foreign company, we pay them more than they would be able to earn locally and, from what I hear, are easier to deal with than many local companies.

Many startups prioritize fundraising, but you've successfully scaled your business through bootstrapping. What advice would you give to entrepreneurs considering bootstrapping as a viable option, particularly in the AI space?

There are so many tools today to get an MVP up and running quickly. But, of course, too many people focus on this and not enough on marketing before launch. Our affiliate strategy, I would say, was the number one early-stage move that we made to explode our growth, and it was only based on results, so we did not have to pay until people landed us real subscribers. If we had not gone the no-code route, another great option would have been to hire some developers in the Philippines or Pakistan to stand up the initial MVP at a very affordable cost. Probably, being in the AI space also made it a bit easier because it is in high demand and growing very quickly. Starting to work with new people on a part-time or work-trial basis also keeps people motivated and focused on quick results. And, when the initial work test does not go well, you are not locked into an under-performing employee or significant severance costs. Only after the initial work test do we talk about making the position full-time, so we are not leading them on and making expectations unaligned. Once we started to get a bit of revenue, working with almost exclusively offshore talent gave us the security to start to build a good full-time team at 80% cheaper than hiring people in the US. Even hiring Europeans, we have found, can be significantly less expensive. It was also important that we picked a very specific niche and ICP, so I highly suggest that other founders do this and really think through their marketing strategy before getting started on building.

With your hands in everything from sales and marketing to development and customer support, how do you prioritize your time and ensure each area receives adequate attention?

This is the most difficult part, definitely. It has been a slow process of finding good middle managers to help me interact less daily with all of the departments. But, over the last year, we are starting to get there and have some great-growing leadership in the CS, HR, development, and marketing teams. I try to stay very organized with product-management software called ClickUp. It allows us to set due dates, priority levels, and keep conversations about specific projects centered in one area for easy reference. We try to reserve Slack usage for more urgent, immediate concerns, as it is much harder to organize all of this information. But, even on that platform, I make sure to keep inbox zero every day and use a lot of the reminder functionality to remind myself to follow up on key issues after a week or a month. Similarly, I keep my email at inbox zero by aggressively unsubscribing from anything unnecessary. I have found that when leaders miss things or are unresponsive, it snowballs into a huge problem for the team, so this is a huge priority. Our OKR system, with weekly and monthly reviews, keeps us focused on the most important things to get done. We try to minimize meetings, and whenever one of my managers can attend something in my place, I push them to do that. Finally, we have very good analytics systems across the whole business, so that I can quickly scan them on a weekly basis and recognize any concerning trends before they become an issue.

AI is rapidly changing the business landscape. What are some key trends you're observing in the AI industry, and how are these trends shaping your company's strategy?

The biggest trend is that the core models are getting more and more powerful. And these are extremely difficult to compete with, as all of the incumbents have massive budgets and staff. We have taken a different approach, which is to create bespoke tools for one specific purpose, and use affiliates, paid ads, SEO, and PR to drive traffic to these tools. We are confident that even though ChatGPT will keep getting more powerful, a more purpose-built tool will continue to get great SEO traffic and can solve problems more quickly for end users. For example, you can certainly use ChatGPT to write a blog article. But to do a good job of it you will need to learn a ton about effective blog articles, do a lot of prompt engineering, and it could take you months to be able to produce something of high quality. Instead, you can just pay us a couple dollars for a finished blog article, we have already done the engineering and the research and optimized it perfectly for you. Our strategy is to keep building out lots of different tools like this and drive traffic from organic SEO as well as paid ads and social.

One of your areas of expertise is SEO and paid advertising. What are some effective digital marketing strategies that have yielded significant results for your AI startup?

First of all, it is very important that we use WordPress for our blog. The course site is hardcoded, and it is much quicker to get perfect SEO with all of the great plug-ins that WordPress provides. Key in this is TranslatePress, which is a plug-in that uses machine learning to automatically translate your blog into as many languages as you want.

When we first implemented this, we did 40 languages, which massively increased the total number of blog articles. In the coming months, we found that these articles very quickly began to rank well all over the world.

We also hired an expert full-time SEO person in Brazil who coordinates a team of article writers and established some strong SOPs for high-quality blog articles that will rank well. Through our affiliate program, we got tons of backlinks from a wide variety of sites, which helped to boost our SEO profile quickly. We also hired a PR person to get placements in major media outlets, which also provided some very high-quality backlinks that help SEO.

Another big focus has been on the speed of the site, which is crucial for SEO. We use Kinta as our WordPress host, which has a global cloud. Sitting on top of that is Cloudflare, which further helps to improve speed. Finally, we use a service called Prerender, which optimizes speed specifically for SEO bots to further boost our profile.

For strategy, we do extensive keyword research to find long-tail search terms that we can easily rank on the first page for. We are currently publishing three times a week and looking to increase that to daily this month. As our domain authority has grown greatly in the last year, it is now very easy for us to rank highly, especially for longer-tail search terms.

For paid advertising, there really was no silver bullet. I would just encourage people to try lots of different platforms and lots of different ads and monitor them closely to weed out underperforming ones. Paid ads take time and patience but can be very scalable once you find specific ads and platforms that provide a good ROI.

For example, even though we initially started with Google Ads and still spend a lot on there, we found that TikTok, Snapchat, and Meta gave us some very attractive returns, and so we have been growing our spend on those platforms more quickly.

One thing I would say about ad creative is it's important to be controversial. Our product is kind of controversial, and so that helps a lot in getting engagement, especially on TikTok.

Customer support is crucial for any business, but even more so for a rapidly growing tech company. Can you elaborate on your approach to building and managing a high-performing customer support team remotely?

We hired our entire team in the Philippines, which allows us to have 24/7 live-chat support. This is crucial. When we upgraded to the system from our email-based system, we improved our average first response time from 48 hours to five minutes. If you can catch people immediately when they are having a problem, they are less likely to churn or leave negative reviews online. We have a manager of the team who sets the schedule and does a weekly meeting with the team to break down specific chats and coach them on how they could have been handled better. We also have analytics which show us the overall customer satisfaction, as well as for each specific agent. This helps us identify under-performing agents and coach them towards better performance, as well as give bonuses to our top agents. In addition, we have many shortcuts which have been reviewed by the management team to make sure that the responses to the most common questions are well-thought-out and crafted. Finally, there is an AI layer on the live chat, which is linked to a database of blog articles about support. When people start a new chat, they get automatically recommended to articles that could help solve their problem without talking to an agent. We have found this to also be crucial in catching bugs. When the site has a major issue or downtime, we find out about it within five minutes from the CS team. Any new common bugs which pop up are directed to me and I review them and forward appropriate ones to the development team to get fixed very quickly.

Looking ahead, what are some of the biggest challenges and opportunities you foresee for your company and the AI industry as a whole, and how are you preparing for them?

The biggest challenge is that there is a ton of competition because the space is growing so rapidly and the profitability of companies like ours is very good, with impressive margins. So everybody wants in on this action. Thankfully, as a first-mover, we have already built up a very powerful domain, which protects us from many startup competitors and allows us to rank on the first page of Google for new tools. To keep building on this, we are doubling down on our SEO efforts and our rate of shipping new tools. By diversifying into a number of different tools, that also allows us to de-risk if any single one of them faces too much competition. It also allows us to have one credit platform and one tech infrastructure to support multiple products, which keeps us lean and capital-efficient. And when users are excited about multiple tools of ours, they are less likely to churn, even if they stop using one of them. Another big risk that we are always monitoring is regulatory changes all over the world. So far, this has not affected us too much as none of our current tools involve video, audio, or images, but it is a constant threat, especially when you see some of these stringent regulations that tend to be created in Europe. Diversifying into different tools has also helped de-risk this to some extent, but with over 10 million users now, we are already at the scale that we could become subject to some of these regulations at some point.