Solving Problems, Not Just Selling, with a Problem-First Mindset

Welcome to "AI Meets API," the podcast exploring the powerful intersection of artificial intelligence and API-driven integration. In this insightful episode, host Ajay Konda, CGO and co-founder of Prowess Software Services, spoke with esteemed guest Jake Morgan, Senior Director of Solution Engineering and Financial Services at Salesforce, who brought deep insights from his 12 years of experience in finance integration and AI-driven automation.

The discussion centered on how organizations can revolutionize their approach to technology by adopting a problem-first mindset and strategically leveraging APIs and AI.

The Problem-First Philosophy: Reimagining the World

Jake Morgan is a strong advocate for focusing on the problem statement rather than the product side. His core philosophy is selling the problem, not the product. He warns that many opportunities are missed when teams focus too heavily on existing tools—likening it to having a hammer and trying to find all the nails, sometimes forcing things that should be screws to be nails.

Morgan emphasizes the need to focus on the desired outcome and play the "art of the possible," especially since customers often don't know what they don't know.

A key strategic viewpoint for AI adoption involves completely reimagining existing processes:

• Instead of designing a process for humans and then attempting to "shoehorn AI", organizations should think as if they started day one with AI.

  
  

• The goal is to design the process with AI capabilities in mind first, and then strategically inject the human element where necessary.
  

The Evolution of APIs and Real-World Success

Morgan’s career journey provided a unique background, having worked across various IT operations and industries, including implementing software distribution systems for major global companies like Goldman Sachs and Morgan Stanley around the year 2000. He later moved to BMC software, helping CIOs earn a seat in the boardroom by developing the Business Service Management (BSM) concept based on the ITIL framework.

Morgan eventually joined Mulesoft as one of the early technical employees (the sixth SE in the company). In those early days, Mulesoft evolved beyond purely focusing on integration patterns and ESBs to significantly change the landscape by embracing API management and implementation.

A crucial component of this shift was recognizing that technology alone was insufficient. Success required addressing people, process, and technology. This led to the establishment of the Center for Enablement (C4E), designed to foster a community and evangelize the use of APIs across global organizations, including entities like Coca-Cola and Annheiser-Bush.

A standout customer success story involved Annheiser-Bush:

• For a small Proof of Concept (PC) involving a beer garden app, a third-party mobile company defined their API needs (the Experience API) using the RAML spec.

• They were able to mock up the functionality and secure executive approval in a single 45-minute meeting, with changes being made dynamically in the UI and published.

• Mulesoft then took that top-down specification and rapidly implemented the Experience, Process, and System APIs.

• This approach allowed them to complete the entire PC in one week, a task that would normally take a month due to complex coordination and iteration.

This pattern of API adoption continues to replicate across enterprises, unlocking the entire business as a platform. For example, at Exxon Mobile, integration via APIs allowed the business side to dictate the type of oil needed for production for the very first time, rather than just processing whatever came out of the ground.

Navigating the AI and API Trends

Currently, the intersection of AI and APIs is viewed as the "Wild West," reminiscent of the internet's early days when powerful tools existed (like access to 100 pages of content via modem in 1994), but no one knew exactly how to capitalize fully on it (pre-web browser).

Morgan stresses that data is fundamental. Achieving autonomous automation requires data exposure, which must be made available primarily through an API of some sort.

For API practitioners, AI tools are already accelerating work significantly:

• Tools like Anthropic’s Claude are powerful for proof of concepts (PCs) and mocking. Morgan noted that he used Claude to rapidly design an underwriting API in OAS 3.1 and then implement that API in Mulesoft, even generating the necessary mock data and database scripts.

  
  

• Future innovation focuses on adopting AI into API management. This includes creating self-aware agents that can dynamically create or pull policies to consume, understand queries from approaches like GraphQL, and even flag opportunities for improvement in an API based on how it is being used.

  
  

Solving Adoption Hurdles and Measuring Success

A critical challenge in API adoption is overcoming the belief that "everything should be an API". Morgan advises:

• APIs should be created only when the goal is reusability.

• Enterprises should find use cases that demand an API be used three or four times before it is created.

• If an API is not being used, it should be deprecated or merged with others.

  
  

Successful adoption requires internal marketing to evangelize the power of APIs and having a central place (like Mulesoft Exchange) where people can easily find, consume, test, and request enhancements for available APIs.

When measuring success, organizations should start by analyzing their project backlog to find integration overlaps, which drives reusability. Key Performance Indicators (KPIs) include tracking:

1. API creation: How many APIs were created in initial projects.

2. Usage frequency: How often those APIs are being used (reusability).

3. Growth rate: How fast net new APIs are being added or how often existing APIs are being changed.

   
   

Morgan cautions against relying solely on "low-hanging fruit" projects early on, as they are not sustainable for proving ROI. Instead, project timelines should be managed by staggering larger, high-ROI projects (identified in the project backlog) with smaller initiatives.

These APIs, created for standard business processes, then become the capabilities that AI agents tap into for gathering information, contextual summarization, and driving intelligent automation.

The Role of Integration in Application Rationalization

Integration also plays a key role in the current trend of application rationalization (cutting down the number of applications used). This is directly tied to legacy modernization. Mulesoft has long capitalized on this segment by creating API facades in front of legacy systems like mainframes and AS400s, unlocking valuable data.

Integration is also central to M&A strategies, helping streamline environments where different integrations tools have accumulated. Morgan noted one customer identified over 200 different integration tools they were using, highlighting the massive need for rationalization.

In closing, Morgan noted that AI has "staying power," unlike technologies such as blockchain, and will continue to evolve rapidly. Ultimately, the goal of AI remains consistent with the earliest debates in the 1950s: human augmentation.