Agentic artificial intelligence in the enterprise

Abstract

The next major wave of value creation from generative artificial intelligence (AI) will come from enterprises implementing agentic systems. Previous waves of AI innovation (such as machine learning) have proven very effective at automating and enhancing largescale processes which have a quantitative element. In the marketing and advertising industry, this has been evident in the impact of predictive modelling for targeting and personalisation. However, other parts of the marketing industry, such as business-tobusiness account-based marketing, have proven harder to improve, because these rely on a lower volume of decisions with much higher dimensionality, supported by unstructured data that includes e-mail exchanges, proposal documents, product descriptions and so on. Generative AI agents possess the potential to add real value in these areas as they can reason over such data very effectively and generate actions that can be automated through account-based marketing and customer relationship marketing systems, enhancing business-to-business sales and marketing effectiveness. While much has been written about the strategic importance of agentic AI (and major industry players such as Hubspot and Salesforce are pushing agentic capabilities hard), this paper provides a practitionerfocused guide to the underlying architectural patterns and implementation choices and explores the governance and agent management practices that organisations will need to implement to reduce risks associated with agentic AI. This article is also included in The Business and Management Collection which can be accessed at http://hstalks/business.