Humanly reviewed by
Angie Navia from Jalasoft’s AI Practice Group shares lessons from building an internal AI chatbot in a thought leadership article published in Unite.AI.
Jalasoft announced today that Angie Navia, Full-Stack Developer and member of Jalasoft’s AI Practice Group, has published a thought leadership article in Unite.AI, a leading global publication covering artificial intelligence and emerging technologies.
The article, titled “Technology Alone Does Not Guarantee Adoption: Lessons from Building an Internal AI Chatbot,” explores a real-world engineering experience building an internal AI chatbot designed to help users navigate a new enterprise application. The piece examines the technical architecture behind the system and the deeper lesson that emerged: successful AI adoption depends as much on interaction design and user expectations as it does on the underlying technology.
In the article, Navia describes how the chatbot leveraged vector embeddings to ground responses in internal documentation, improving accuracy and reducing hallucinations. However, despite a technically sound implementation—including context management, document retrieval, and backend integrations—the project ultimately revealed that clear boundaries, guided interactions, and expectation-setting are essential to sustained user engagement with AI systems.
“From an engineering standpoint, the system performed well,” Navia explains in the article. “But it lacked intentional interaction design. Building conversational AI is not only a technical challenge. It is also an interaction design challenge.”
The article also highlights key technical and operational lessons from the project, including:
- Strategies to reduce hallucinations using embedding-based document retrieval
- The complexity of managing conversation history and token limits
- Challenges related to multilingual AI performance
- The importance of clearly defining the scope and capabilities of AI agents
- How user expectations can evolve beyond the system’s intended role
These insights contribute to a growing conversation across the industry about the gap between AI experimentation and real organizational adoption.
“At Jalasoft, our engineers are constantly exploring how AI technologies can be applied responsibly and effectively in real-world software environments,” said a representative of Jalasoft’s AI Practice Group. “Angie’s article reflects the type of practical engineering insight our teams bring to clients building AI-enabled products and platforms.”
The article is part of Unite.AI’s Thought Leaders series, which features perspectives from experts shaping the future of artificial intelligence.
Readers can access the full article here: https://www.unite.ai/technology-alone-does-not-guarantee-adoption-lessons-from-building-an-internal-ai-chatbot/
About Angie Navia
Angie Navia is a Full-Stack Developer at Jalasoft with more than five years of experience building production software applications and integrating AI capabilities into modern platforms. She has completed the IBM Generative AI for Software Developers Specialization and actively applies AI tools and techniques in her development workflow.
About Jalasoft
Jalasoft is a nearshore software engineering company providing high-performing development teams to organizations across the United States and beyond. With more than two decades of experience and a proprietary talent development ecosystem, Jalasoft delivers scalable engineering capabilities across areas such as AI, software development, quality assurance, cloud, and data platforms.
