VENDORiQ: Zapier’s Evolution – How Low-Code Agentic AI Could Deliver a Revolution in Automation

Why it Matters

The focus on agentic AI architectures by low-code vendors, such as Zapier, reflects a broader industry trend. This approach leverages AI, particularly ‘reasoning AI’, to decompose complex requests into smaller, actionable tasks. These tasks are then routed to existing low-code processes, each retaining its specific integrations. 

This method aligns with the premise that AI’s primary value in enterprise contexts may stem from its ability to orchestrate existing linear processes and to perform information search and summarisation within these workflows. 

Organisations frequently encounter challenges in automating end-to-end processes due to the inherent complexity of integrating disparate systems and the need for dynamic decision-making. Agentic AI, within a low-code framework, presents a potential solution by providing a layer of intelligent orchestration. 

For instance, a customer support agent might initiate a process to resolve an issue; an AI agent could then determine necessary steps, such as retrieving customer history from a CRM, accessing a knowledge base for troubleshooting, and initiating a support ticket in a separate system. 

This is an evolution of traditional process automation, where the AI component introduces adaptability to diverse or unstructured inputs. Major enterprise software vendors are also investing in this domain, often positioning AI-driven solutions to enhance the consumption and value of their core offerings. 

For organisations already leveraging low-code platforms for agility and rapid application development, the integration of agentic AI could extend their automation capabilities. However, a key consideration for organisations will be the transparency and configurability of these AI agents, ensuring they align with business logic and compliance requirements. While Zapier notes its updated AI by Zapier includes a prompt builder, indicating a degree of control over AI behaviour, the depth of this control and its suitability for complex enterprise scenarios will be critical for adoption. 

Who’s Impacted?

• Software and Enterprise Architects: Assess the strategic implications of adopting AI-directed processes, particularly in terms of integration with existing infrastructure, data governance, and long-term scalability. 
• Automation/Business Process Management Teams: Evaluate how AI reasoning services can augment or redefine current automation strategies, identifying opportunities for increasing flexibility in processes.
• Application Development Teams: Understand the new paradigms of building and managing AI-driven workflows within low-code platforms, potentially shifting from rigid process definitions to more adaptive, AI-orchestrated components. 

Next Steps

• Conduct pilots to assess the efficacy of agentic AI within a low-code environment for a specific, well-defined business process. 
• Evaluate the transparency and auditability features of agentic AI solutions to ensure accountability and compliance with regulatory requirements. 
• Develop internal expertise in generative AI engineering and AI configuration to maximise the effectiveness of new agentic capabilities.
• Engage relevant stakeholders, including legal and compliance teams, to address data privacy and ethical considerations associated with semi-autonomous AI workflows.