Adaily: the present and the future. An interview by Aleksander Piskorz.

The text below is an interview conducted by Aleksander Piskorz with me, featured on the Tech Pill blog and published on August 5th. You can find the original version (in Polish) here.

Hey, Piotr! Can you tell me, what has occupied most of your time in the last few weeks?

Searching for inspiration to develop my product. I read X/Twitter, blogs, ProductHunt, and follow specific individuals on LinkedIn and Youtube. I also track all announcements from the AI industry and ponder how I can leverage them. It's always in the back of my mind, accompanying every conversation with a potential user or witnessing a new solution on the market.

Of course, it doesn't stop with just thinking - the last few weeks have also been about implementing various ideas with my partners.

Adaily is a tool set to revolutionize marketing. What were the biggest challenges in creating this tool, and how did you overcome them?

We started with exploring user needs and matching them with the current possibilities offered by AI. Anyone following the progress of generative AI knows that almost every month brings a mini-revolution pushing the market forward, making some products obsolete. Hence, anticipating the direction of change and building our unique "moat" (or advantage) that can't be easily surpassed is crucial.

This perspective is crucial as we now have to consider what OpenAI or Google might achieve in a year, rather than just our competitors.

What AI technologies and approaches have you applied to adaily to ensure quality results while maintaining a conversational approach?

In short, we combined the capabilities of GPT-4 with the LangChain framework. Our conversations are powered similarly to ChatGPT, which I believe is currently the market gold standard. LangChain has enabled us to incorporate vector-based databases, a revolution, in my opinion, as big as AI. Thanks to them, we can efficiently and effectively access information from vast databases.

For us, this has added significance: semantic searches often result in exploring entirely new creative directions.

Will adaily be available for small businesses and freelancers, or are you targeting it as a paid model only for large agencies?

Adaily will have two main subscription plans: individual and team. We aren't introducing more granularity right now, basing this decision on user behavior analysis. Therefore, by the end of August, users will be able to use the individual plan.

It will provide access to the tool and a complete knowledge base (2000 award-winning campaigns from 2019-2023). We're thinking of various marketing experts in agencies, businesses, and even freelancers. Later in the year, we'll introduce the team plan, especially for advertising agencies or client-side marketing departments. In this version, we'll add collaboration features, such as shared team chats or using a local knowledge base.

Are there competitive products in the market trying to achieve what adaily does? How do you distinguish yourselves in this segment?

There's no solution like ours, and I don't think there will be one for a long time. We dig our moat deeper every day. Of course, looking broadly, there are many AI martech products, but they address different areas. When considering the full spectrum of the marketing process, which starts with analysis and strategy and ends with execution, we strongly address the initial phases, while the vast majority of other products address the end. This has its benefits in the form of complementary solutions.

What are the main indicators you look at to gauge Adaily's success?

Currently, we have an open public demo, available without registration. So, quantitatively, we're focusing on analyzing the number of new and returning users, session duration, as well as geography and traffic sources.

We're thrilled with the significant interest from key markets: North America and Western Europe account for half of our traffic.

Additionally, on a qualitative level, we analyze all queries in the tool. I personally read tens of thousands of conversations trying to understand how to improve the product.

Which current limitations in AI technologies would you like to break through in the future? Are there specific technologies or tools you'd like to integrate with adaily in the coming months?

A big step in adaily's evolution is ahead: developing our proprietary model that will help generate new marketing ideas. We believe every company deserves marketing at the level of giants like Nike or Coca-Cola. We provide this opportunity through our proprietary database and tailoring to a client's expectations. This project is a high priority now and will be our most challenging yet.

Are there unusual marketing strategies you'd like to explore using your platform, but haven't yet?

I have a background in exact sciences. Mathematics has always been a fascinating way for me to describe reality. Contrary to appearances, marketing is no different. Looking back over decades of advertising, one can distinguish several archetypal solutions or strategies to talk about a product.

For example, the authors of Deck of Brilliance identified 52 such approaches, including "Challenge Yourself" and "Partner Up". Therefore, I believe we already cover or soon will cover virtually all marketing strategies.

However, there's a nuance: some companies perceive marketing differently, focusing perhaps solely on their mailing list, promoting orders, or creating B2B content. In this case, we will be adding more such campaigns to our knowledge base.

What are your thoughts on the role of personalization in marketing, and how can adaily facilitate this?

AI offers the possibility of granular personalization to the absolute. There are systems today that can generate thousands of ad creative variants and optimize them for performance.

However, one must remember that this is a very tactical approach. In many cases, marketing personalization might mean that a brand can communicate with its audience in the right way. So, first "what", and then "how". I deeply believe that tools like ours will help marketers better understand the creation of ad messages.

Do you think marketing is becoming more complicated or simpler thanks to artificial intelligence algorithms?

One of the main advantages of AI is automation.

Sure, at this moment we are bombarded with hundreds of new AI tools for marketing and how they can help. We, as adaily, are also contributing to this. However, it's a misconception that this complicates marketing. I'd call this phenomenon something else: it's democratizing, allowing all companies to have opportunities previously reserved only for giants.

Gather inspiration, generate ideas, create marketing plans, produce hundreds of creatives, and optimize them. Soon or later, it will cost tens of dollars a month, not tens of thousands. The aforementioned automation will allow a click of a button "make an ad for me", and this process will handle itself. Therefore, in the long run, AI will significantly simplify marketing, even if, under the surface, it becomes statistically more complex.

And finally, I'll ask... do you plan to add instructions for beginner users on how to interact with adaily? (laughter)

Yes, I promise (laughter). In all seriousness, we always have in mind that our tool should be for everyone. There's a pejorative term, but it makes a lot of sense: "idiot-proof". I would like a person with minimal marketing knowledge and technologically excluded to get value from using the tool.

We have already introduced many solutions, and we add more. For example, no matter how you ask a question, our bot will "translate" and interpret it. We'll soon introduce suggested questions. I personally do not believe that as a society we will become experts in prompting. Therefore, we need to create AI tools that are maximally "prompt-less".

Thank you so much for the conversation.

>> Try adaily yourself! Click here <<

Update: Later the same day, we actually added new features to help users interact with the chatbot more effectively, such as suggested questions that appear as tiles.