SaaS integration challenges explained by Prismatic

In this Q&A, we delve into the world of SaaS integrations with the CEO and co-founder of Prismatic Michael Zuercher. He shares valuable insights on the challenges companies face when implementing integrations, including security concerns, scaling difficulties, and the impact on core product development. Michael offers guidance on mitigating security risks and evaluates the crucial factors in the build vs. buy decision for integrations. He also provides a roadmap for companies starting their integration journey and explores the future of integrations, including emerging opportunities with AI tools. With his extensive experience in scaling software companies and working with numerous B2B SaaS businesses, Michael offers a unique perspective on how companies can navigate the complex landscape of SaaS integrations to drive innovation and meet customer needs efficiently.

ADM: SaaS integrations enable powerful new functionality but also have hidden pitfalls like cost overruns and timeline delays. What SaaS integration challenges often catch companies by surprise?

Zuercher: Some of the biggest pitfalls we see include security, scaling difficulties, lack of standardization, accumulation of technical debt, and disruption to core product development.

Connecting SaaS apps introduces myriad security concerns that require ongoing effort to address. Scaling from a few integrations to dozens and deploying them to hundreds of customers is a significant challenge that can't be solved by simply growing teams. Every external system has its own unique way of doing things, adding unexpected complexity. Rapidly delivering customer-specific integrations under pressure often leads to development shortcuts and technical debt. As integration demands grow, they can consume more and more development resources, making it harder to innovate on the core product.

Companies often underestimate how much these challenges will snowball as integration needs grow. What seems like a reasonable-sized effort can balloon quickly and have major consequences for competing priorities. Building on top of a mature integration platform can be a game-changer in avoiding these common pitfalls.

ADM: What are some of the security challenges associated with SaaS integrations, and how can developers mitigate these risks?

Zuercher: When it comes to SaaS integration security, it's all about mitigating the inherent risks of connecting multiple systems. Developers need to focus on implementing robust authentication measures like OAuth 2.0, securely managing credentials, integrating only with trusted apps, validating all incoming data, adhering to the principle of least privilege, and comprehensively logging and monitoring integration activity.

Adopting security best practices and partnering with an experienced integration platform provider can significantly bolster your SaaS integration security posture. The key is to make security a top priority in your integration processes from day one. 

ADM: Building vs buying integrations is a crucial decision for any scaling SaaS business. What key factors do the most successful companies evaluate when approaching this decision?

Zuercher: After scaling a software company whose product had hundreds of integrations and then working with hundreds of B2B SaaS companies making this exact decision, I definitely have some advice here!

For most teams, the first factor is considering which path most reduces the time their engineering teams spend on integrations overall. Make sure you think about this holistically across the full build/deploy/manage integration lifecycle. Don’t stop with evaluating the time it takes to build the actual integrations. Also, consider how you will run them in production, deploy them across your customer base, handle customer-specific configurations and credentials, log and monitor the integrations, manage security and scalability, and provide a customer-facing integration UX. Implementing a third-party integration platform does take some time upfront and may be overkill for a small number of simple integrations. For anything more, an embedded iPaaS is likely the most efficient route. It abstracts a great deal of boilerplate work, comes with all of that infrastructure built in, and provides tools that enable non-developers to take on much of the workload.

The second major factor to consider - and I can’t emphasize this enough - is choosing a path that actually enables you to build the full range of integrations your customers need. We’ve all had - or heard of - bad experiences with integration platforms that force a low-code-only way of thinking and throw up roadblocks when things get complicated. Anything can look good in a blog post or demo. If you do choose to pursue the buy route, I recommend two things. First, be sure to do an in-depth, hands-on proof of concept. Build an actual integration from your roadmap to ensure the platform you choose actually handles the real-world complexity of your typical integrations. Second, look for a platform that provides both low-code and full-code building experiences so you have full flexibility to build what you need in the way that suits you.

The third factor is picking a path that fits the way you build software overall. A major argument for keeping integrations in-house is being able to incorporate them into your existing tools and processes. You don’t want to end up with something that sits off to the side and requires an entirely different way of doing things than your core product. Fortunately, there are embedded iPaaS solutions that let you write integration code the way you see fit, use your existing logging tools, CI/CD pipelines, and so on. If you choose to buy, select one of those.

ADM: For companies early in their integration journey, what does an ideal integration development roadmap entail in terms of stages, priorities and time investments?

Zuercher: The first step is to determine whether to build integrations and related tooling in house or implement an integration platform. That decision determines the rest of the integration development roadmap.

Teams that build everything in-house spend from two weeks to three-plus months building each integration, depending on its complexity. Some of that time is spent building the actual integration code to move data between systems, but much of it goes to boilerplate work like setting up infrastructure, provisioning resources, and creating the customer configuration UX.

Teams who use an embedded iPaaS typically invest several weeks upfront purchasing and implementing the platform. They can then build each integration in as little as several hours, with more time required for highly complex integrations. We see companies build integrations up to eight times faster after implementing an embedded iPaaS. They also report an improved ability to accurately estimate integration timelines.

Three key questions can help teams prioritize which integrations to build first: 

• Which integrations are most needed by customers and prospects?
• Which integrations will extend your product's functionality and serviceable addressable market (SAM)?
• Which integrations will increase your competitive advantage?
   

It’s common to make exceptions within that framework to build an integration needed to win an in-flight deal or retain a key customer.

ADM: As AI capabilities grow more sophisticated, what emerging integration opportunities do you foresee between SaaS platforms and AI tools like multimodal models?

Zuercher: Today, AI systems like OpenAI are already being utilized as endpoints within integrations, and we're seeing integrations play a crucial role in collecting the vast amounts of data needed to train large language models (LLMs) and other AI systems. Some low-code integration designers have even begun incorporating AI prompts to provide users with interactive, real-time support during integration building.

Looking ahead, we expect AI to become increasingly adept at predicting integration patterns and offering more robust assistance in integration development. Multimodal AI models, which can understand and generate content across text, images, speech, and video, create possibilities for more natural, context-aware interactions within SaaS applications. Integrating these AI capabilities will enable SaaS platforms to offer more personalized, efficient and user-friendly experiences, driving up engagement and productivity. While the full impact of AI on the integration landscape remains to be seen, it's clear that the opportunities are vast and growing.

ADM: Low-code integration builders aim to democratize development for non-technical users. How might these tools expand the integration possibilities for less technical business teams?

Zuercher: Low-code integration builders certainly democratize integration development. They’re a critical component of both traditional iPaaS, used by businesses to build internal integrations, and embedded iPaaS, used by B2B SaaS companies to provide integrations to the other products their customers use. 

Traditional iPaaS has revolutionized integration development within businesses by enabling line-of-business users to automate workflows and quickly address business needs while reducing the burden on IT and developers.

For embedded iPaaS, however, democratizing integration development within B2B SaaS companies requires providing code-native integration building options in addition to low-code integration builders. Low code enables non-developers to take on much of the integration burden, while a code-native approach gives developers the flexibility to tackle highly complex integration scenarios and incorporate integrations into their existing workflows and CI/CD processes. 

By providing both low-code and high-code tools in a single platform, integration solutions can empower less technical teams to build simple integrations autonomously while also enabling developers to take on more sophisticated projects. This versatility expands the overall integration possibilities for the organization and allows each team member to contribute in the way that best suits their skills and preferences.

ADM: Looking ahead, what do you view as the most transformative integration capabilities on the horizon for SaaS businesses? How could they drive the next wave of innovation?

Zuercher: The integration landscape is evolving rapidly, and there are several transformative capabilities on the horizon that have the potential to drive innovation for SaaS businesses. One exciting area is the emergence of AI assistance in integration platforms, which will likely become increasingly adept at predicting integration patterns and assisting with integration development. 

Another game-changer is the continued evolution of integration solutions to serve specific needs. Embedded iPaaS solutions purpose-built for B2B SaaS are emerging as a key driver of innovation, enabling software companies to quickly build reusable, configurable integrations and deliver them to customers as a seamless part of their product.

However, the most transformative integration solutions will be those that support both low-code and high-code development, catering to the full spectrum of technical aptitude. This versatility will empower teams across the organization, from business users to professional developers, to contribute to integration projects based on their skills and expertise.

Ultimately, the integration capabilities that will drive the next wave of innovation will be those that empower SaaS businesses to be more agile, responsive, and adaptable to ever-changing market demands.

About Michael Zuercher

Michael Zuercher is CEO and co-founder of Prismatic, a leading embedded integration platform (embedded iPaaS) for B2B SaaS companies. A second-time founder, he previously scaled a software company whose product had hundreds of integrations serving thousands of customers. He passionately believes that B2B SaaS companies need a better way to build and manage integrations.