TuringBots are going to be a game changer for development teams, with early adopters claiming increased software development productivity of 15 per cent to 30 per cent, Forrester reveals in a new report.
“Microsoft has released Copilots for non-software-development-related roles and activities, but they are not TuringBots; they are assistive bots for all business and technical users across the enterprise,” Forrester said. “Google Bard and OpenAI ChatGPT are both business and development assistive bots.”
TuringBots are specifically designed for the software development lifecycle (SDLC), delivering “automation and semi-autonomous capabilities to plan, analyze, design, code, test, deliver, and deploy while providing assistive intelligence on code, development processes, and applications.” To examine their potential, Forrester spoke to industry experts from the likes of IBM, GitHub, OpenText and Globant.
Based on those conversations, and with the ballooning generative AI usage, Forrester predicts a journey of only two to four years before seeing a concrete impact of TuringBots in the market. Business and technology management teams, leveraging TuringBots, will be able to boost speed and productivity with improved natural language processing, deliver crucial knowledge to developers and generate product development assets in seconds.
The entire SLDC is set to gain even more in efficiency as low-code, cloud-native development, and value stream management vendors move to infuse generative AI into their products, Forrester said. At each step of the lifecycle, TuringBots can help:
- Analyze and plan software projects
- Generate code and automate design and models
- Improve software testing
- Accelerate and automate DevSecOps delivery
- Drive product development with data insights
Currently, products like Atlassian’s Intelligence, Figma, Microsoft Sketch2Code, Amazon DevOps Guru, Red Hat Ansible Lightspeed, Google Duet, CloudFabrix and more come very close to delivering one or two of the above capabilities, but they will continue to evolve very fast.
The full potential of TuringBots will unfold over three time horizons, the report concludes, predicated by key factors such as AI innovation speed, enterprise risk management practices, governance, cultural acceptance, and AI regulation, as well as a lack of trust in IP protection and security.
The three phases are as follows:
- Short term (1-2 years) – Product owners, testers, and developers will leverage TuringBots as assistants, ordering them to accomplish tasks, with strict reviews and controls still in place.
- Mid term (2-4 years) – TuringBots will mature to the point of autonomously supporting business people, become members of the team and have more interactive collaboration with humans. AI regulations will get approved, TuringBots will start to comply, and human developers will increasingly trust them.
- Long term (five plus years) – TuringBots will scale at the enterprise level, the SDLC will become a practice of the past as human product managers interact with TuringBots through language and visual graphic designs. TuringBots will do all the hard work, and human developers will review, refine, and supervise.
