Creating Virtual Assistants: Chatbots that Work (MEAP Edition, Version 1)

Building a useful virtual assistant takes more than teaching software to recognize a phrase. It involves decisions about conversation flow, language understanding, supporting systems, and how success will be measured. Andrew R. Freed’s Creating Virtual Assistants: Chatbots that Work introduces those moving parts in this Manning Early Access Program edition, with an emphasis on making the underlying ideas approachable.

This Version 1 MEAP edition presents the work in progress, rather than the later completed book published as Conversational AI. Its contents map out a practical path from core concepts to building, training, testing, deploying, and improving assistants.

More than chatbots

The opening chapters distinguish three kinds of virtual assistants: conversational assistants, command interpreters, and event classifiers. That broader view helps explain why some assistants guide a user through a multi-step exchange, while others handle a single command or classify a message without any dialogue.

Freed also introduces the roles of the interface, dialog engine, natural language understanding, and orchestration. Together, these concepts offer a useful way to think about how an assistant interprets a request, chooses what to do, connects with other services, and responds.

From foundations to the working outline

The contents are organized into three parts: foundations; building and training; and maintenance. The listed chapters move from creating a first assistant and designing effective processes and dialog to developing rules and classifiers, training, measuring, testing, deployment, and ongoing improvement. A final chapter is devoted to moving an assistant to voice, and an appendix offers a machine-learning refresher.

Accessible technical grounding

The introductory material assumes familiarity with decision trees, process flows, and basic branching logic such as if statements. It presents machine learning as part of the subject without requiring readers to arrive with advanced mathematical training. The focus is on understanding how the pieces of an assistant fit together and what different contributors bring to its development.

Who may find it useful?

This early-access edition is aimed at readers interested in developing virtual assistants, particularly software developers and people working alongside development teams. Beginners can use the opening material to build a conceptual foundation; readers with prior experience may prefer to consult the chapter outline for topics relevant to their work.

If you want a grounded introduction to the systems behind chatbots, voice assistants, and automated message classification, this MEAP edition offers a clear starting point—and a view of the larger development process it sets out to explore.