Snaplogic – Aaron Kesler, Product Manager, Snaplogic -emporkence Series

Aaron Kesler, AI/ML Product Manager at Snaplogic, is a certified product leader with more than a decade of experience building scalable frameworks that bring design thinking, the work to be done, and product discovery. He focuses on developing new AI-driven products and processes while guiding aspiring PMs through his blog and directing strategy, execution and customer-centric development.

You have already had an entrepreneurial journey, starting with college, then continuing to be acquired by Carvertise. How did these early experiences affect your product mindset?

This is a very interesting moment in my life. My roommate and I started Stak because we were bored with our courses and wanted real-world experiences. We never thought this would lead us to be attracted to poster startups in Delaware. This experience really affected my product mindset because I naturally tend to talk to businesses, ask their questions and build solutions. I didn’t even know what the product manager was at the time – I was just doing the job.

On Carvertise, I started doing the same thing: Work with their clients to understand the pain points and develop solutions – before I got the PM title. As an engineer, your job is to solve technical problems. As a product manager, your work turns to finding the right problems, i.e., issues worth solving, because they drive business value as well. As an entrepreneur, especially without funds, your mindset becomes: How to solve someone’s problem in a way that helps me put food on the table? Early scratches and hustle and bustle kept me browsing different shots all the time. Whether you’re at a self-funded startup, a VC-backed company or a healthcare giant, Maslow’s “basic needs” mindset will always be the foundation.

You talk about your passion for aspiring product managers. What advice do you want when you break into a product?

The best advice I’ve got, and what I’ve got for aspiring PMs – “If you always look at it from a customer’s perspective, you’ll never lose arguments.” That line seems simple, but powerful. This means you need to really understand your customers – their needs, pain points, behaviors and circumstances – so you will not only have opinions meetings, but also have insight. Without this, everything becomes a hippo (the opinion of the highest paid person), a battle with more power or louder opinions. With it, you become the one people turn to for the clear ones.

You have previously stated that every employee will quickly work with more than a dozen AI agents. What will the future of this ai-aigment look like in daily workflow?

Interestingly, we are already in the reality of working with multiple AI agents – we have helped DCU plan, build, test, secure customers and put dozens of agents on the workforce that helps them. It’s fascinating that companies are building an organizational chart for each employee based on their needs. For example, an employee will have an AI agent specifically targeting certain use cases (such as the one used to draft epic/user stories) that helps with traction requests for encoding, prototypes or issues, while another person analyzing customer feedback – all of which are approved and carefully planned – because it determines on the backend who can replace access to humans, not visitors, and not clients. There will be a human in the loop for the foreseeable future, but they will remove duplicate low-value tasks so that people can focus on higher levels of thinking. In five years, I expect most teams will rely on agents the way we rely on Slack or Google Docs today.

How do you recommend companies to bridge the AI ​​literacy gap between technical and non-technical teams?

Starting small, there is a clear plan of how to adapt your data and application integration strategy to get it hand-held and open to iterating from the original goals and methods. Find the problem by being curious about the mundane tasks in your business. The highest value problem to be solved is usually the boring problem solved by the heroes without even numbers every day. We learned a lot of these best practices when we set up agents to assist our Snaplogic finance department. The most important way is to make sure you have secure guardrails for what types of data and applications that certain employees or departments can access.

Then, companies should treat it like a college course: simply explaining key terms, giving people the opportunity to try tools in a controlled environment, and then dive deeper. We also know that it is OK to not know everything. AI is developing rapidly, and no one is an expert in every field. The key is to help the team understand what is possible and give them the confidence to ask the right questions.

What effective strategies do you see for AI UPKILLING go beyond the general training module?

The best way I see is to get people to do it. Training is a great start – you need to show them how AI can actually help the work they are already doing. From there, think of it as an approved method of covering it or shadow agents, as employees are creative to find solutions that can solve super specific problems. We have access to our field and non-technical teams with Snaplogic’s proxy AI technology, which removes the complexity of enterprise AI adoption and empowers them to try to build something and ask questions. This exercise leads to a real learning experience because it is related to their daily work.

Do you think companies that adopt AI tools without the right skills are at risk – what are the most common pitfalls?

The biggest risk I see is the substantial governance and/or data that violates data security, which can lead to expensive regulatory fines and the possibility of putting customers’ data at risk. However, some of the most common risks I see are companies that adopt AI tools without fully understanding what it is and without the ability. AI is not magic. If your data is a mess, or your team doesn’t know how to use these tools, you won’t see value. Another problem is when organizations push adoption from top down without considering the actual execution of the work. You can’t just push something out and expect it to stick. You need champions to educate and mentor people, and teams need strong data strategies, time and environment to place guardrails and learning spaces.

At Snaplogic, you are working on new product development. How does today’s AI incorporate your product strategy into your product strategy?

AI and customer feedback are at the heart of our product innovation strategy. This is not only adding AI capabilities, but also rethinking how we provide customers with more efficient and easy-to-use solutions that simplify how they interact with integration and automation. We are making products with power users and non-technical users, and AI helps bridge the gap.

How does Snaplogic’s agent tools help businesses build their own AI agents? Can you share use cases that have a lot of impact?

AgentCreator is designed to help teams build real enterprise-level AI agents without writing single lines of code. It eliminates the need for experienced Python developers to build LLM-based applications from scratch and uses natural language tips to develop teams in finance, HR, marketing in just a few hours, and create AI-driven agents in hours. These agents are tightly integrated with enterprise data, so they can do more than just respond. Integrated agents automate complex workflows, understand and act in real time through decision making, all in a business environment.

AgentCreator has been a game changer for our clients like Independent Bank, which uses AgentCreator to launch voice and chat assistants to reduce IT help desk ticketing backlog and free up IT resources to focus on the new Genai program. Additionally, welfare management provider APTIA uses agents to automate one of its most resource- and resource-intensive processes: welfare elections. Since AI agents simplify data translation and verification across systems, it now takes several hours of backend data input to take several minutes.

Snapgpt allows integration through natural language. How about this democratized access to non-technical users?

Snapgpt is our integrated adverb and is a great example of how Genai can break the barriers in enterprise software. With it, users from non-technical to technical can describe the results they want to use simple natural language prompts (requiring to connect two systems or trigger workflows) and build integrations for them. Snapgpt goes beyond building integration pipelines – users can describe pipelines, create documents, generate SQL queries and expressions, and use simple prompts to convert data from one format to another. It turns out that what once was a developer’s cumbersome process turned out to be something employees throughout the enterprise could access. It’s not only about saving time, but about transferring who can build it. As more and more people in your business contribute, you can unlock iterations and more innovations faster.

What makes Snaplogic’s AI tools, such as Autosuggest and Snapgpt, different from other integration platforms on the market?

Snaplogic is the first generation integration platform to continuously unlock the data value of modern enterprises at unprecedented speed and scale. With the ability to build cutting-edge Genai applications in just a few hours without writing code, and Snapgpt is the first and most advanced Genai-powered integrated co-pilot, organizations can greatly increase business value. Other competitors’ Genai features are lacking or non-existent. Unlike most competition, Snaplogic was born in the cloud and was specially manufactured to manage the complexity of cloud, on-premises and hybrid environments.

Snaplogic provides iterative development capabilities, including automatic verification and read modes, which enables teams to complete projects faster. These features allow more integrators of different skill levels to get up and running quickly, unlike competitors that require most highly skilled developers, which can greatly slow down implementation. Snaplogic is a high-performance platform that processes over 400 million documents per month and can effectively move data to data lakes and warehouses, while some competitors lack support for real-time integration to support hybrid environments.

What do you most excited about the future of product management in an AI-powered world?

What I am most excited about the future of product management is the rise of one of the latest buzzwords to enhance the AI ​​space “Vibe encoding”, that is, the ability to build working prototypes using natural language. I envisioned a world where everyone in the product trio (design, product management and engineering) uses tools that can translate ideas into real-time functional solutions. Rather than relying solely on engineers and designers to bring your ideas to life, it is better to create and iterate everyone quickly.

Imagine a real-time solution on a customer call and at this moment using its actual data prototype. We can work with them to create the solutions they propose, but instead find better ways to solve the problem. This transformation will make the product development process extremely collaborative, creative and consistent. This excites me because my favorite job is to build with other people to solve meaningful problems.

Thanks for your excellent interview, and readers who hope to learn more should visit Snaplogic.