Foxbox Digital is an award-winning engineering-focused product development agency.

Inside the Box Foxbox Digital Blog

Versapay: Digital Payments Mobile Strategy

The democratization of software, product strategy, and the case for proper experimentation. 

Generative AI is making it easier than ever to build the wrong thing

Generative AI is making it easier than ever to build the wrong thing.

"The democratization of software” long promised by no-code software and AI-enabled development tools is soon to actually arrive, thanks to ChatGPT and AI development tools like GitHub Copilot. Businesses of all sizes can undoubtedly benefit from the ability to build digital products in a fraction of the time. But the problem is this: 

The decisions that shape high performing digital products are made long before a line of code is written.

So while the path to building and launching a product might be faster than ever, generative AI tools can also act as a high-speed train to the 

So how do you build the right thing? Here’s a few things to keep in mind.

It’s easy to fall in love with a slick interface. We’ve all done it. And the next wave of generative AI tools will only compound the problem.


When a few prompts and shiny drag and drop tools can churn out UX & UI mockups or even functional applications, it’s easy to create a digital product that appears well-designed on the surface, but whose foundation is shaky at best.

For the same reason you’d never design a car around the upholstery, AI-powered development tools can be your friend, but not as a replacement for things like user research, persona development, technical assessment and competitive landscape analysis.


Particularly when development can be so dramatically accelerated, the ability to test, experiment and challenge your own assumptions is more critical than ever. When building a digital product, being wrong early is a gift.


It’s how we discover that there’s more nuance to various target users than initially assumed. It’s how “happy paths” emerge through careful research and usage analysis. It’s how refined product roadmaps clarify the future development plan that mitigates cost while still making substantial progress.


Don’t squander the gift of being wrong early in the name of being done fast.

The age of generative AI will make the gap between the two wider than ever. A well crafted digital product solves the right problem the right way. Which takes an intimate understanding of your business, your customer and broader cultural context that AI will fall short of reaching.


When it’s easier than ever to build the wrong thing, we’ll see a proliferation of software in search of a problem to solve. Perhaps circling the problem, even approaching it or pointing to it, but not built soundly enough from the jump to actually solve it.


So should you invest in AI tools capable of writing software? Sure.


But remember: software’s not really what you’re after.

Ready to learn more about building a practical AI roadmap for your organization? Join our  free workshop: AI Roadmaps for the Real World on LinkedIn Live on April 18th.

Click here to  learn more and save your spot!

Embarking on a project is like setting sail on an epic adventure - a quest for success sprinkled with challenges, triumphs, and enough coffee to power a small army. 

Numbers Worth Knowing: How We Think About Project Metrics

Embarking on a project is like setting sail on an epic adventure - a quest for success sprinkled with challenges, triumphs, and enough coffee to power a small army.

As you can imagine, before we start our journey, just like setting sail, we have to make sure our rigging is right, our people are ready, and we at least roughly know our destination.  Recently, we helped a client work on a fast and furious front end project to bring their ideas to life with a very aggressive 3 month timeline. After delivery, we took some time to take a look at what it means to partner, even for a short time, simply to show what it takes to foster a collaborative environment.  We’ve crunched the numbers to show you what you could expect in working on a custom build product development project.

In this blog post, we're not just diving into the intricate metrics of our project but also infusing it with a generous dose of humor because, let's face it, laughter makes any journey more enjoyable, and, full disclosure, the author of this post loves idioms. Join us on this joyous ride as we explore the metrics that defined our project, where 9 dedicated Foxboxers and 6 client folks joined forces.

A team of 9 Foxboxers and 6 client folks – because, as the saying goes, "Two heads are better than one," and apparently, 15 heads are even better. Each person played a key role in delivering the end result, and as you can imagine, 15 people keeps a project manager busy!

1 meticulously crafted Figma file with dozens of iterations – it's like a treasure map, but for our application. X marks the pixel-perfect spot, and once again, collaboration and calibration gave us exactly what we needed to bring our client’s vision to life.

148 Jira issues developed, reviewed, merged, QAed, and closed – With a mix of user stories, subtasks, bugs, and chores, it’s crucial to always log the work you’re doing, both for transparency and for documentation.

An average sprint velocity of 64 – we're not saying we're faster than a cheetah, but have you ever seen a cheetah close a Jira ticket? Yeah, we didn't think so.

9 demo meetings were held – because we believe in transparency, and also, we had to show off all those beautiful UI features to someone other than ourselves. Without the demo, we would not be able to speak to our hard work, nor would our client be able to tell us what they’re seeing, and if what we built met their expectations.

32% automated test coverage achieved – like a warm, automated security blanket, ensuring our code doesn't get the chills. We believe in covering our code like we cover our late-night snack cravings – thoroughly and without compromise.  What does this look like?  Well,  we automed the happy path most of our users will take.  This is important to create a benchmark of what the tests will look like, and to free up our awesome QA engineer to file bugs and write test cases.

169 PRs expertly merged – because nothing says "team spirit" like a well-merged pull request, or in our case, a code-coordinated group hug.

125 manual test cases written and leveraged – proving that sometimes, a human touch is necessary, especially when the machines get a little too binary. We like to think of manual testing as the project's spa day – indulgent, meticulous, and oh-so-necessary, cucumbers on the eyes are optional of course.

2,300 Slack messages exchanged – because sometimes, the best project decisions are made with GIFs, emojis, and the occasional pun.

98 Daily Standups - Ever heard the phrase, "A stitch in time saves nine"? Well, spending 15 minutes collaborating daily is our version of that wisdom. While Slack is the heartbeat of our remote work, there's magic in face-to-face (well, screen-to-screen) conversations. When it comes to clarifying a user story, hashing it out in real-time during our standups beats waiting for those Slack messages to ping back and forth. Sometimes, a quick chat saves hours of back-and-forth typing!

As the project unfolded over 3 months, the metrics weren't just numbers; they were a testament to our unwavering focus and the output of every single one of the actions we took each day. From the inception where our diverse team united, to the meticulous crafting of the UI, and the comprehensive testing processes, resilience and collaboration were our secret sauce, and maybe now you can imagine what working together could look like for you!

Goodbye 2023!
You were good to us. Not only did we add eight new team members and five new clients, but 2023 marked the most time and energy we’ve ever spent working on a client that’s often easy to keep on the back burner: Ourselves. 


Foxbox: Goodbye 2023 & Hello 2024!

You were good to us. Not only did we add eight new team members and five new clients, but 2023 marked the most time and energy we’ve ever spent working on a client that’s often easy to keep on the back burner: Ourselves.

Here’s what it looked like for us in 2023:

The launch of AI Lab at the end of 2023 marked our most significant commitment to date in helping clients harness the power of AI. From design exploratories to determine how AI fits into your business, to incorporating AI-powered functionality directly into a digital product, AI Lab is designed to separate the helpful from the overhyped.

Bolstered by new methodologies and a renewed focus on client collaboration, our Product Inception offering has transformed how we lead businesses through the strategy, design and product blueprint process. Read more about 

We formed a new team within Foxbox dedicated to Product Maintenance. As a combination of engineers, QA specialists and specialized service managers, Foxbox Product Maintenance plays an invaluable role in giving our clients the proactive, predictable support they depend on for their production applications. Read more about our 

The healthcare industry continues to prove to be an area of strength for us as a digital product partner. We added two Healthcare clients to our roster in 2023, and continue to organize our teams to match our industry experience to the inherent complexities digital-first healthcare companies face. Read more about  

We reorganized our engineering teams around core specialties: web & backend, mobile, and QA. Our foxes felt the changes in efficiency, team morale, process and overall quality almost immediately.

We hosted our first live discussion about Generative AI in November. John Nicholson from In4mation Insights was kind enough to join us for a candid, practical conversation about all things Generative AI and the future of digital product development. Thanks to everyone who came out!

Our foxes love to post about the tech they love:

Retrofit2: An Appium Implementation to Upload APKs to Browserstack

With 2023 in the books, it’s time to look ahead. My feet may be frozen. And my snow boots may be salty. But I couldn’t be more optimistic about what 2024 will hold.

We’re big believers that we’ve barely scratched the surface of what Generative AI is capable of. That’s why we’re doubling down on the 

 and hands-on experimentation necessary to help our clients stay ahead of the curve. We’re even building our own tools using OpenAI, including a 

project estimation tool built by our AI Lab.

Our people are our most important asset. The new operating model solidified in 2023 has made a tremendous impact on our ability to make our teams more cohesive, connected and positioned for personal growth. Together with our leadership team, I look forward to building on that momentum, creating new opportunities for meaningful time together and a more intentional approach to the culture, habits and expectations that shape us as a company.


We’re hosting a series of small and intimate gatherings designed to provide time and space for meaningful, candid conversations about digital products & AIGenerative AI and digital products.

Chicago on February 7 to answer your questions around AI

If you or someone you know is curious about working at Foxbox, and you don’t see a relevant open role below, don’t hesitate to reach out. We’re a remote-first company with big plans. We’re looking for top-tier talent to join our team.

We hear about Generative AI everywhere. I wanted to see the real-world applications for generative AI, so I admittedly felt like I was behind the 8-ball for not jumping on the bandwagon. Last quarter, I met with tech leaders to understand how they envision using Generative AI to solve real problems for customers. I was blown away by those conversations, and we’re putting what we’ve learned into practice.

Foxbox 2023 Q3 Update: Generative AI, Inc 5000 (again!) & more

If you’re trying to decide how to incorporate AI into your products, I hope you reach out. Let’s have a conversation to answer the questions you have around AI. 

Overall at Foxbox, growth has remained strong! We’re working with exciting new clients in healthcare, data, and the aviation industry. We’ve brought on new key team members that we’re thrilled to have on board.

 Inc. 5000 list for the second year in a row!


Our growth is a result of our team’s commitment to delivering exceptional work for our clients. We are proud to be one of the fastest-growing companies in the US.


The headlines about AI are everywhere for good reason. The value to be unlocked is VERY real. But AI tools are just that: tools. They’re only as effective as they are reviewed, refined and continuously realigned toward the product or experience being built.

With that in mind, I’m excited to share two updates on what AI means to us at Foxbox.

We’re building an AI service to help clients enhance their products with generative AI.I’ve been calling the team our “AI Squad.” We’ve assembled a team of Foxbox Senior Developers and AI experts to solidify an AI-development capability at Foxbox.Whether it’s building custom chatbots, automating manual processes, or synthesizing data to provide insights, we’ll soon be ready to show clients a path to harnessing AI that’s practical, efficient and carefully tailored to their business.

- We recently published the first iteration of the Foxbox AI Dev Playbook. We’re open-sourcing how we leverage AI to accelerate development for our clients. For developers looking to put promise into practice, it’s a practical snapshot of what we think are the most fruitful and feasible places to start.Will you be able to build custom software 10x faster? Probably not. But it will make you a more efficient and effective developer.Check out the 

 at the link below, and let us know what you think. We’ll be making updates periodically, so any and all feedback is welcome!

, our new Recruiting Business Partner at Foxbox. With over 14 years of experience in business, marketing, client and contract management, Rachel brings a unique skill set that aligns seamlessly with our culture.

Known for her ability to foster strong relationships and execute effective recruitment plans, Rachel collaborates closely with our hiring managers to understand how our client’s challenges are evolving, identify personnel needs to best meet those challenges, and engage highly-sought after talent. ls in identifying hard-to-find talent and engaging highly sought-after professionals.

Beyond her professional acumen, Rachel is also bilingual, fluent in both Spanish and English, and is a lover of the arts. Originally from Miami, Florida, Rachel and her loyal companion, Leo the dog, now call the suburbs of Atlanta, Georgia, home. In 2021, Rachel graduated with an MBA from Loyola University New Orleans School of Business.

We're thrilled to have Rachel on board, and we're confident that her diverse skills and experiences will be invaluable in supporting Foxbox's continued success. Welcome to the team!

If you or someone you know is curious about working at Foxbox, and you don’t see a relevant open role below, don’t hesitate to reach out. We’re a remote-first company with big plans. We’re looking for top-tier talent to join our team.


New breakthrough AI systems are making headlines for good reason. 
The ability to accelerate software development, close skill gaps and even improve overall code quality all present extraordinary potential value. Having said that, AI tools are just that: tools. Particularly when building mission-critical digital products and experiences, these tools are only as effective as they are reviewed, refined and continuously realigned toward target outcomes and objectives.

The AI Dev Playbook - How to Accelerate Development with AI | Foxbox Digital

New breakthrough AI systems are making headlines for good reason.

The ability to accelerate software development, close skill gaps and even improve overall code quality all present extraordinary potential value. Having said that, AI tools are just that: tools. Particularly when building mission-critical digital products and experiences, these tools are only as effective as they are reviewed, refined and continuously realigned toward target outcomes and objectives.

So if you’re a developer looking for practical ways to integrate AI into your workflow, you’ve come to the right place. We designed this guide to demonstrate what we think are the most fruitful and feasible places to start.

 is an AI-powered tool that developers can use directly within code editors.

When utilized properly, it can dramatically accelerate development, improve code quality, and even help build new skill sets and programming language fluency.

Here’s how you can put it to work today.

 Because GitHub Copilot can understand the context of existing code, it automatically suggests accurate and relevant code completions in real-time.

 It’s the shortcut developers have always longed for. GitHub Copilot can quickly generate things like code templates, class definitions and function signatures, so repetitive coding tasks don’t prevent you from focusing on the core logic of your application.

3. Integrate External Libraries and APIs:

 When working with unfamiliar libraries or APIs, GitHub Copilot can accelerate your learning curve. With a simple prompt, it can quickly understand usage patterns, proper syntax and general best practices to make integration fast and painless.

 If you're venturing into a new programming language, GitHub Copilot can act as a teacher, guide and assistant simultaneously. As it generates code snippets and examples, it can help explain nuances to the language’s syntax and idiomatic patterns.

 Get to validation and refinement faster by testing your code under different scenarios and test cases. Based on the logic and context of your code, it can even suggest adjustments, further reducing the manual time and effort needed in typical test-driven development practices.

You can sign up for a 30-day trial directly with Github. Then simply prompt Copilot directly from your code editor and it will begin suggesting solutions immediately. As a developer, you’re still in charge. Know what you’re looking for, cycle through options to visualize them in context, and make more informed decisions.

 represents the evolution of AI-powered coding tools, building upon the foundation laid by GitHub Copilot. While both tools share the common goal of enhancing developer productivity and streamlining the coding process, CoPilot X takes things to a new level.

Unlike the original Copilot, CoPilot X harnesses the power of an enriched and more diverse dataset, allowing it to grasp a broader spectrum of programming languages, frameworks, and libraries. So no matter the context or programming language, CoPilot X is your astute sidekick.

Don’t throw the learning out with the debugging bathwater. CoPilot X goes beyond conventional coding assistance by offering insightful explanations with suggested code fixes. CoPilot X's explanations serve as valuable learning moments, helping you expand your coding prowess while remaining focused on the goals of the application.

CoPilot X Introduces a more sophisticated documentation process, bolstering efficiency at every step of the development lifecycle. Through automatically generated explanations, inline citations and precise answers to context-dependent questions, hurdles can be cleared faster and decisions can be made with more insight than ever before.

Engineers can ask clarifying questions or auto-generate code snippets, all directly from within the IDE. Based on the context fully contained within the repository, Copilot X is armed with everything it needs to deliver clear, accurate answers, all without interrupting your workflow.

Skipping the pull-request template is how the explanations for code changes get lost in translation. CoPilot X is setting out to transform the pull request process by leveraging its advanced code analysis. When a PR is initiated, Copilot X auto-generates descriptions and a precise summary of the changes made. Communication is dramatically streamlined. Momentum keeps moving forward.

 is an advanced natural language processing conversational model that can be directly used by developers for tasks such as conversational agents, customer support, and content generation. It provides an interactive and intelligent chatbot experience.

As a senior developer, you possess a wealth of knowledge and experience in solving complex problems. ChatGPT can enhance your capabilities by providing real-time assistance and valuable insights.

Think of it as your sounding board, field guide and assistant all in one.

Engage in dynamic discussions with ChatGPT to explore different approaches and solutions. Describe your project challenge or objective, and ChatGPT can generate creative ideas, alternative strategies, and potential solutions. This collaborative brainstorming process can help you uncover new perspectives and overcome roadblocks.

Developer: "I'm facing a performance issue in our application. What are some possible optimizations we can consider?"

ChatGPT: "There are several optimization techniques we can explore. One approach could be to implement caching mechanisms to reduce database queries. Additionally, optimizing algorithms and leveraging asynchronous processing can also improve performance. Let's dive deeper into each option."

When confronted with complex code issues, ChatGPT can serve as a helpful companion. Describe the problem you're encountering, share snippets of code, and seek guidance on potential solutions or debugging techniques. ChatGPT can provide suggestions, point out possible errors, and offer best practices based on its extensive training in programming knowledge.

Developer: "I'm getting a 'NullPointerException' error in this section of code. How can I troubleshoot and fix it?" ChatGPT: "NullPointerException usually occurs when you're trying to access a null reference. To debug this, you can start by examining the line mentioned in the error message. Check if any variables involved are properly initialized or if there are any missing object assignments. Additionally, you can use debugging tools to step through the code and identify the exact point where the exception is thrown."

ChatGPT can serve as a valuable research assistant, helping you find relevant documentation, articles, and resources to deepen your understanding of specific topics or technologies. Simply provide the context or ask questions, and ChatGPT can suggest authoritative sources or summarize relevant information.

Developer: "I need to integrate a payment gateway into our application. Can you provide me with some documentation or tutorials?"

ChatGPT: "Certainly! For integrating payment gateways, you can refer to the official documentation of popular providers like Stripe, PayPal, or Braintree. They offer comprehensive guides, SDKs, and sample code to assist you throughout the integration process. Additionally, I can fetch specific tutorials or articles if you have any particular requirements."

ChatGPT is also invaluable tool in Proof of Concept (POC) prototyping, allowing you to gauge viability of potential solutions with remarkable speed and accuracy.

Lets explore an example scenario where ChatGPT can help generate an index.html file to calculate the distance between two points using JavaScript.

Engage with ChatGPT and clearly outline the objective of your POC.

Developer: "I'm working on a Proof of Concept (POC) and need to generate an index.html file. The objective is to calculate the distance between two points using JavaScript. Can you assist me in creating the initial code structure?"

ChatGPT: "Certainly! I can help you with that. Let's start by creating the basic structure of the index.html file with the necessary JavaScript code to calculate the distance between two points."

Engage in an iterative conversation with ChatGPT to generate the code for your POC. Describe the logic and steps involved in calculating the distance, and ChatGPT can provide suggestions and help you refine the code.

Developer: "To calculate the distance, we need to use the Pythagorean theorem. We'll need the x and y coordinates of the two points. Let's start by creating two input fields for the coordinates and a button to trigger the calculation. How can we implement this in JavaScript?”

ChatGPT: "To achieve that, we can start by creating the HTML structure with the input fields and button. Then, we'll use JavaScript to fetch the values entered by the user, calculate the distance using the Pythagorean theorem, and display the result. Here's an example code snippet to get you started..."

Example using Chat GPT to generate a javascript code for distance calculation

Continue the conversation with ChatGPT to address any additional questions or modifications to the generated code. Refine the code based on your specific requirements and proceed with testing the POC to ensure it meets the intended objectives.

Whether it's generating code structures or refining implementation details, ChatGPT serves as a valuable companion in the iterative process of POC development, enabling you to quickly test ideas and assess their potential before investing significant resources.

API or Framework integrations can bring AI-enabled technology even deeper into your solution. While these integrations present exceptional upside and exciting possibilities, it’s critical to closely evaluate whether your particular solution requires it in order to avoid unnecessary complexity and potentially time-consuming tasks. These are a few of the AI integrations we think are most promising:

 OpenAI provides an API that allows developers to integrate ChatGPT into their applications. By using the API, developers can send requests to the ChatGPT model and receive responses, enabling seamless integration of AI-powered chatbot capabilities.

 Microsoft Cognitive Services offers a suite of APIs for computer vision, speech recognition, natural language understanding, and decision-making. Developers can integrate these APIs into their applications to add AI capabilities in a modular manner.

 AWS provides a range of AI services with APIs that developers can integrate into their applications. Services like Amazon Rekognition, Amazon Polly, and Amazon Comprehend offer pre-trained models accessible through APIs for image analysis, text-to-speech conversion, and natural language processing, respectively.

 Google Cloud AI offers various APIs that developers can utilize for different AI tasks. APIs like Google Cloud Vision, Google Cloud Speech-to-Text, and Google Cloud Translation allow the integration of AI capabilities related to image recognition, speech recognition, and language translation.

 Microsoft Azure AI provides APIs for computer vision, speech recognition, natural language processing, and more. Developers can integrate Azure Cognitive Services APIs and Azure Machine Learning into their applications for adding AI functionalities.

 TensorFlow is an open-source machine learning framework that provides tools and libraries for developers. It allows direct usage for building and deploying machine learning models in tasks like image recognition, natural language processing, and deep learning.

 PyTorch is a popular open-source machine learning library that offers a flexible and dynamic approach to building neural networks. Developers can directly use PyTorch for research, prototyping, and production deployments.

 Keras is a high-level neural networks API written in Python that runs on top of TensorFlow, Theano, or Microsoft Cognitive Toolkit. It offers a user-friendly interface for developers to directly build and train neural networks.

The AI Dev Playbook - How to Accelerate Development with AI

Last quarter we acquired four new clients across telehealth, data & analytics, aerospace, and e-commerce. We’re honored to be chosen as their technical partner to design, develop and bring their products to market.

LocalVR has been a disruptive force in the vacation rental industry since 2016, but arduous financial processes and backend systems created a bottleneck to growth. We partnered with Stripe to build a new platform, dramatically improving operational efficiency and overall speed to market. To the tune of:

​​Happier end-users. Stronger systems. Faster growth. Wins all around. Thank you to Stripe for being an incredible partner in this process, and to LocalVR for your trust and collaboration at every step. 

We’ve created two new awards called The Foxies as a way to recognize our team members who have gone above and beyond.

 - Awarded to a team member who shows excellence in their role. A person who consistently delivers results, showcases skills, and produces quality work.

Awarded to a role model at Foxbox. A person who is collaborative, positively impacts others, and helps to create a great culture at Foxbox!

If you or someone you know is curious about working at Foxbox, and you don’t see a relevant open role below, don’t hesitate to reach out. We’re a remote-first company achieving big things. We’re looking for top-tier talent to join our team.

Sales Executive - Minneapolis or Chicago (remote)

Sr. Backend Engineer-Python LATAM (Remote)

The majority of our clients come through referrals. If you know of anyone who is looking for a technical partner to design and develop a new digital product, replatform an existing product, or maintain a production application, we’d love to meet them.

Inside the Box

Articles, updates, and ideas you can put to use.

Generative AI is making it easier than ever to build the wrong thing

Numbers Worth Knowing: How We Think About Project Metrics

Foxbox: Goodbye 2023 & Hello 2024!

Foxbox 2023 Q3 Update: Generative AI, Inc 5000 (again!) & more

The AI Dev Playbook - How to Accelerate Development with AI

Foxbox Founder's Update - 2023 Q2