TuringBots Will Disrupt The Software Development Approach Forever — Here’s How
We just published Forrester’s The Top 10 Emerging Technologies In 2023 report, and TuringBots are not onlyin the top 10 technology trends but are expected to deliver ROI much earlier than predicted in earlier research that we published for clients, all thanks to generative AI.
Generative AI Advances TuringBots Maturity Much Faster Than Expected
What’s happened? AI, in the last six to 12 months, has progressed at the speed of light: from single models to multimodels, with large language and foundational models capable for the first time in the history of AI to effectively generate assets: text, images, code, and more. This progress can be summarized in two words: generative AI (genAI). In 2020, when Forrester first coined the term “TuringBots,” we predicted a long journey — between five to 10 years — before seeing a concrete impact in the market. In 2023, we’ve shortened that journey to two to four years. Today, we are seeing value and benefits already, and TuringBots are improving at an outstanding pace.
OpenAI leads The GenAI Disruption But Is Not Alone
Large language models (LLMs) and multimodels have also opened the door to more comprehensive abstractions across different domains. Suddenly, AI not only could generate with images from scratch (DALL·E) and summarize text with ChatGPT, but even the developers of ChatGPT found out almost by accident that the same LLM could also generate code from human natural language into many, many programming languages. How could that happen? Because, in the training data, OpenAI accessed repositories on the internet with trillions of lines of open source code (many living in the GitHub open source repository) … and, oops, what an interesting use case! AI can write code and help build software faster! But OpenAI is not the only player building TuringBots. Microsoft, its strategic partner, also is, as well as IBM, Google, Amazon, and smaller players such as Tabnine, Replit, and others. They’ve all focused their genAI efforts on software development and IT operations use cases.
It’s GenAI (TuringBots) In The Loop, Not Humans!
We believe that TuringBots and humans will go through major collaboration maturity phases, pushed by the quickening pace of AI technology innovation. Generative AI just boosted the market takeoff of TuringBots to what they are today. As AI technologies mature, we predict that TuringBots will traverse three key phases: From current state to future state, TuringBots will start initially (phase 1) “in the loop,” with human stakeholders of development teams in the driving seat; in the next phase (phase 2), more accurate TuringBots will become peer members of the team, with more interactive collaboration between humans; and TuringBots in the next phase, or phase 3 (and not necessarily the last), will become more autonomous but still supervised and orchestrated by humans. You will find more details on how the three horizons unfold for enterprise adoption over the next five-plus years in the Forrester report, The Future Of TuringBots.
This blog follows earlier blogs and research that we’ve written about TuringBots. We are also developing a rich stream of research on genAI and TuringBots — covering the state of TuringBots, the architecture of TuringBots, and the next-generation vendor landscape — so stay tuned. If you are using any genAI TuringBots in any of the phases of software development, whether you are a vendor building TuringBots or a client using them, do reach out to me at dlogiudice@forrester.com to connect and share your experience. In addition, if you are a Forrester client, schedule a guidance session, inquiry, or strategy session if you have any questions or need any assistance in defining your own TuringBot strategy.
