Triggering Autonomous Agent Action

In partnership with

Dear Reader…

Those of you who have experienced riding in an Autonomous Vehicle or AV’s will be directly familiar with AI enabled agents making decisions without human intervention - see my account of the experience - in case you haven’t. But what about in other realms - such as data engineering? How will autonomous decision making agents play a part in data management, and how will your role as a data engineering be augmented or sidelined?

A critical piece of the puzzle to enable autonomous agents is text-to-action, this is one of several advances that we are diving into as we look to what the future holds in 2025. Text-to-action technology is poised to significantly impact the design, build, and use of data and analytics systems. This emerging capability will transform how organisations approach data collection, analysis, and decision-making processes.

🧐 What is Text-to-Action?

This refers to an AI systems' ability to take natural language inputs and translate them into complex actions or software commands without human intervention. Simply put text-to-action enables you to describe tasks in plain language, which the AI then executes autonomously, much like you would instruct an employee to complete a given task. For example, a user could instruct an AI to "build a competitor to TikTok," and the system would generate the necessary code, deploy the application, and potentially iterate the application.

Eric Schmidt on the implications of text-to-action:

🙇🏼 Some not-so Obvious Implications

While the example above may seem a little fantastical, what we will see in 2025 are incremental steps towards making such a text-to-action prompt a commercially viable proposition. Amongst the wider implications are:

The Democratisation of Innovation

Text-to-action has the potential to revolutionise many aspect of a wide range of industries by making it possible for individuals and small businesses to launch projects that previously required entire teams of engineers. This democratisation of innovation is already leading to a surge of new applications and service providers for specialist services. On such application is bland.ai You can actually have it call you as a demo - and it is blazingly fast - I encourage you to check it out

Enhanced AI Capabilities

We can expect AI systems to handle increasingly complex tasks based on natural language inputs, with text-to-action capabilities.This advancement is closely tied to developments in large context windows (discussed last week) and advances in reasoning across LLM’s.

Software Development Augmentation

One of the most immediate applications of text-to-action is in software engineering. Data engineers can expect to see AI systems that can generate code and develop applications based on high-level descriptions. Even more interesting is the recoding of legacy applications that were built using now obsolete languages and there is no-one left that understands how the code works. This is fertile ground to be able test and prove the use of text-to-action AI agents for code development in a sandbox environment.

Business Process Transformation

Text-to-action will enable more sophisticated automation of business processes. AI systems will be able to execute multi-step tasks, from generating reports to potentially managing entire projects based on prompted instructions. Potentially handling routine tasks such as customer support, report generation, and data analysis, improving efficiency and monitoring for human-based errors.

Cyber Defence

In the realm of cybersecurity, text-to-action will enable security professionals to rapidly deploy complex defensive measures and incident response protocols using natural language commands. For instance, security teams could describe desired security configurations or threat mitigation strategies in plain English, and AI systems would automatically implement the appropriate actions across networks and systems. On the offensive side, attackers may leverage text-to-action to automate and scale sophisticated attacks, potentially generating tailored malware, crafting convincing phishing emails, or exploiting vulnerabilities with unprecedented speed and efficiency. This technology could lower the barrier to entry for cyber attacks, allowing less skilled This will likely accelerate the cat-and-mouse game between attackers and defenders, necessitating even more robust, AI-driven security measures.

Writer RAG tool: build production-ready RAG apps in minutes

• Writer RAG Tool: build production-ready RAG apps in minutes with simple API calls.

• Knowledge Graph integration for intelligent data retrieval and AI-powered interactions.

• Streamlined full-stack platform eliminates complex setups for scalable, accurate AI workflows.

Learn more about our production ready RAG tooling here.

👩🏽‍💻Data Systems Design

Text-to-action is part of what will revolutionise the design of data and analytics systems in 2025:

• Intuitive User Interfaces: Systems will be designed with natural language interfaces, allowing users to interact with data and analytics tools using plain English commands. This will make complex analytics more accessible to non-technical users.

• Automated System Architecture: AI-powered text-to-action capabilities will enable the automatic design of data pipelines, storage solutions, and analytics workflows based on high-level descriptions of business requirements.

• Adaptive Design: Analytics systems will be designed to continuously evolve based on user feedback and changing business needs, with text commands triggering automatic adjustments to the system architecture.

👷🏽‍♂️ Data Systems Build Impact

The build process for data and analytics systems will see transformations along these lines with text-to-action:

• Accelerated Development: Text-to-action will dramatically speed up the development of data and analytics solutions. Developers will be able to generate complex code and data models by describing desired functionalities in natural language, shifting their attention the quality of the code being generated and overseeing the process.

• Automated Integration: Building integrations between different data sources and analytics tools will become more streamlined, with AI interpreting text instructions to create necessary connectors and data transformations.

• Enhanced Quality Control: AI-powered text-to-action systems will automatically generate test cases and perform quality checks based on natural language descriptions of expected system behaviour.

📊 The Impact on the Use of Data Systems

The way data product customers use data and analytics will undergo significant changes, with:

• Democratisation of Analytics: Text-to-action will make advanced analytics accessible to a broader range of users within organisations. Employees without technical expertise will be able to perform complex analyses by describing their requirements in plain language.

• Real-time Insights: Users will be able to request and receive real-time insights by simply asking questions about their data. The system will automatically perform the necessary data retrieval, analysis, and visualisation based on the text input.

• Automated Decision-Making: Text-to-action systems will enable more sophisticated automated decision-making processes. Users will be able to define complex business rules and actions using natural language, which the AI will then execute autonomously.

• Enhanced Customer Insights: Text analytics capabilities will be supercharged by text-to-action technology, allowing businesses to gain deeper insights from customer feedback across various channels. This will enable more personalised customer interactions and improved product development.

• Streamlined Reporting: Generating reports and dashboards will become as simple as describing the desired output in natural language. The system will automatically collect relevant data, perform necessary analyses, and create visually appealing presentations.

⚠️ Risks of Text-To-Action

With great power, comes great responsibility and text-to-action as a power brings with it significant risks…

• Data Privacy and Security: As a system becomes more autonomous in data handling, ensuring compliance with appropriate privacy regulations and developing new more robust security measures will become part of a Data Engineers role.

• Quality Assurance: Validating the accuracy and reliability of actions taken by AI based on text inputs will require new approaches to quality control, data quality and governance procedures.

• Ethical Considerations: At an Enterprise level the establishment of clear guidelines for the use of text-to-action systems to prevent misuse and ensure responsible AI deployments will be a high priority in 2025.

Text-to-action technology is set to revolutionise the design, build, and use of data and analytics systems. It is part of a number of key advances that we will be discussing as we turn the page on 2024. Preparing yourself for these advances is something we would encourage you to do over the Holidays this year.

🔮 Preparing for the Future

Data professionals will see an evolution in their roles to focus more on defining high-level strategies, that are aligned with business requirements, as well as overseeing AI-driven business processes. Some ideas to stay ahead in this field include:

• Developing Natural Language Processing Skills: Understanding and implementing NLP techniques will be crucial for working with text-to-action systems. Developing your prompting skills, as well as building in guide rails in prompts will become a critical success factor.

• Focus on AI Ethics and Governance: As AI becomes more autonomous, understanding the ethical implications and implementing proper governance will be essential. the unintended consequences of poorly thought through prompts become even greater as we give agents more autonomy.

• Enhance Cloud and Distributed Computing Knowledge: Text-to-action systems will likely rely heavily on cloud infrastructure and distributed computing for processing complex tasks. Your understanding of how to optimise deployment will become important to reduce the computing overhead associated with letting AI Agents lose on systems deployment.

• Stay Updated on AI Legislation: Keep abreast of evolving AI regulations and ensure compliance in your projects. As regulation rolls out around the world such as the EU AI Act covered this article earlier in the year.

Text-to-action promises to be a highly consequential development for Data & AI Engineers in 2025: Are you prepared? We would love to hear your perspective, please weigh in and join the conversation at the Data Innovators Exchange.

Coming up next week, a deep dive into Agentic workflows and their impact in 2025.

That’s a wrap for this week.

Last of all check out the recently launched Enterprise AI Engineering Classroom and Resource Hub. You will find materials to get you started using IBM Watsonx in an enterprise environment.

Thank you