The Convergence of Low Code No Code and AI/ML

The rise of low code no code (LCNC) development platforms is one of the most disruptive trends in enterprise technology today. LCNC tools enable both professional developers and non-technical business users to rapidly build web and mobile applications with little to no coding, using visual drag-and-drop interfaces and pre-built components.

As an artificial intelligence (AI) and machine learning (ML) expert, I see LCNC as a major enabler for scaling the adoption and impact of AI/ML. By dramatically simplifying the development process, LCNC expands the pool of people who can build AI/ML applications and the speed at which they can be deployed. At the same time, the latest LCNC platforms are integrating AI/ML capabilities in innovative ways that will reshape enterprise application development in the years ahead.

The Rapid Rise of LCNC

While LCNC platforms have been around for years, they have seen explosive growth recently. Gartner expects LCNC tools to be used for 65% of all application development by 2024, with 66% of large companies using at least four different LCNC tools [^1]. The global LCNC platform market is projected to grow from $13.2 billion in 2020 to $45.5 billion by 2025, a Compound Annual Growth Rate (CAGR) of 28.1% [^2].

Several factors are driving this rapid adoption of LCNC:

1. Acute shortage of skilled developers: There are an estimated 250,000 unfilled developer jobs in the US alone and over 1 million worldwide [^3]. LCNC helps organizations build the applications they need with the talent they have.

2. Accelerated demand for digital solutions: The COVID-19 pandemic greatly accelerated enterprise digitization efforts, with 69% of boards of directors accelerating their digital business initiatives [^4]. LCNC enables much faster development and deployment of new applications.

3. Increased business-IT collaboration: LCNC empowers business domain experts and "citizen developers" to directly build the solutions they need in collaboration with IT, rather than tossing requirements over the wall. 59% of organizations have active citizen development initiatives [^5].

4. Maturing LCNC platforms: The capabilities of leading LCNC platforms have expanded rapidly to enable development of enterprise-grade applications with robust scalability, security, and governance. Providers like Microsoft, Salesforce, ServiceNow, Google, Appian, Mendix, and OutSystems are heavily investing in their platforms.

AI Gets Low Code and No Code

The transformative potential of AI and ML is well established, but development of AI/ML solutions has largely remained complex and costly, requiring specialized data science and developer skills. This is starting to change with a new generation of AI-powered LCNC tools.

By encapsulating machine learning models and data pipelines in reusable components with easy configuration interfaces, LCNC platforms are beginning to democratize the development of AI/ML applications. Some of the ways AI and LCNC are converging include:

• Embedded ML building blocks: Major cloud platforms including Microsoft Azure, Google Cloud, and Amazon Web Services now offer pre-trained ML models as drag-and-drop components in their LCNC tools. For example, Azure‘s AI Builder provides customizable models for object detection, category classification, entity extraction, key phrase extraction, language detection, sentiment analysis, and more that can be embedded in Power Apps with no code [^6].

• Visual data science tools: New tools like Google Cloud AutoML Tables enable non-technical users to automatically build and deploy state-of-the-art ML models simply by uploading a data table and selecting a target column to predict. The tool handles data preparation, feature engineering, algorithm selection, and model tuning behind the scenes.

• Natural language interaction: LCNC platforms are leveraging natural language processing (NLP) to enable interacting with applications using plain English. For example, Salesforce enables building Einstein Bots to automatically engage with customers using natural conversations [^7]. Microsoft Power Apps Ideas uses AI to translate natural language input into data queries, formulas, and workflows [^8].

• Automated user experience: AI-driven design assistance is helping citizen developers create more effective, personalized, and accessible user experiences without deep expertise. For example, Google‘s AppSheet uses reinforcement learning to provide personalized next-step suggestions as you build your app [^9].

• Intelligent process automation: LCNC workflows can be enhanced with AI for intelligent document processing, image recognition, decision support, and predictive analytics. AI-powered LCNC platforms can observe user behavior to identify opportunities for automation.

By enabling more of the AI/ML development lifecycle to be handled with LCNC tools, the time, cost, and skills required to build intelligent applications will be greatly reduced. Gartner predicts that by 2025, 70% of new applications developed by enterprises will use LCNC or no-code technologies, up from less than 25% in 2020 [^10]. The research firm also forecasts that the number of citizen developers at large enterprises will be at least four times the number of professional developers by 2023 [^11].

Implications for AI/ML Adoption

The convergence of LCNC and AI/ML has profound implications for the adoption and impact of intelligent technologies across industries. Some of the key developments I expect to see include:

1. Accelerated AI deployments: Enterprise adoption of AI has been hindered by talent shortages and the complexity of traditional development approaches. LCNC will enable a much wider range of organizations to quickly deploy AI-powered applications for specific business needs. This will shift AI from a competitive differentiator to a basic necessity.

2. Citizen data science: While LCNC won‘t replace the need for professional data scientists, it will enable many more analytically-savvy business users to build basic ML models and embed them in applications. Leading LCNC platforms already provide AutoML capabilities to automatically train, tune and deploy models without requiring data science expertise.

3. Composable AI: As catalogs of pre-built, reusable ML models and data components expand in LCNC platforms, it will become increasingly easy to stitch together powerful AI workflows and applications. This composable approach makes the power of AI accessible without the underlying complexity.

4. Continuous learning everywhere: LCNC enables iterative development and real-time performance monitoring of deployed models. We‘ll see more AI/ML solutions that continuously learn and adapt as they are used, without needing manual retraining. LCNC will enable non-tech users to be directly involved in this iterative process based on business outcomes.

5. Scaling last-mile AI: Many organizations struggle to turn proofs-of-concept and siloed AI successes into scaled, systematic deployments. LCNC provides tools to quickly build last-mile interfaces and integrations to make AI real and actionable for frontline users in their daily workflows.

6. Reducing bias: Bias and lack of transparency are major challenges in enterprise AI deployments. LCNC platforms provide governed, documented processes to build and deploy models, helping to standardize best practices and reduce risks. Automated testing tools in LCNC can help detect potential fairness and bias issues before they become problems.

Getting Started with LCNC for AI/ML

For organizations looking to harness the power of AI and LCNC, some key best practices include:

1. Assess your LCNC maturity: Most organizations are already using some LCNC tools, but with varying levels of scale and sophistication. Systematically evaluate your current tools and practices to identify opportunities to expand citizen development and embed more AI.

2. Build your citizen dev community: Identify analytically-savvy and automation-oriented business users who can become champions for AI/ML with LCNC. Provide them training, support, and incentives to build solutions. Create forums for citizen and professional developers to collaborate.

3. Establish governance guardrails: LCNC should democratize development, but within appropriate guardrails. Implement processes for IT to manage platform security, data access, architecture standards, and reusable components. AI/ML solutions in particular require robust data governance.

4. Focus on quick wins: Start by rapidly developing solutions with LCNC for focused AI/ML use cases with measurable business impact. Common quick wins include call center chatbots, document processing automation, predictive maintenance, and customer churn prevention.

5. Compose, don‘t code: Take advantage of pre-built AI/ML capabilities in your LCNC platform rather than trying to develop complex models from scratch. Experiment with AutoML tools for custom model development before investing in more manual data science efforts.

6. Upskill your professional devs: Ensure your software engineering talent is trained on your LCNC platform so they can build any complex capabilities that can‘t be composed, and provide reusable components for the citizen developer community.

Risks and Challenges of LCNC for AI/ML

While the potential of LCNC to accelerate and democratize AI/ML development is immense, there are also challenges and risks to manage:

• Lack of explainability and trust: As citizen developers rapidly build and deploy AI/ML solutions without deep data science expertise, there is increased risk of undetected bias, fairness issues, and lack of explainability. Organizations must establish processes to document data provenance, provide transparency into model behavior, and continuously test for responsible AI.

• Data quality and governance: AI/ML solutions are only as good as the data they are trained on. LCNC tools make it easier to quickly connect models to diverse data sets, but this democratization introduces risks. Clear data governance policies, processes, and tools must be in place to ensure data quality, security, privacy, and regulatory compliance.

• Talent gaps: While LCNC expands the pool of potential builders, it still requires some technical and analytical skills that may be lacking in many organizations. Training and upskilling programs are essential to build communities of capable citizen developers and ensure professional developers can effectively use LCNC tools.

• Shadow IT: If not properly managed, LCNC initiatives can become a new form of shadow IT, with business units deploying AI/ML solutions without appropriate oversight and controls. IT should provide guardrails and collaborate closely with citizen developers to mitigate risks.

• Vendor lock-in: As organizations build mission-critical applications using a particular LCNC platform, it becomes harder to switch vendors. Carefully evaluate the openness, portability, and integration capabilities of platforms before standardizing.

The Road Ahead for LCNC and AI/ML

The convergence of LCNC and AI/ML is still in its early stages, but the trend is unmistakable. As LCNC platforms continue to add more AI/ML capabilities and AI/ML technologies become more accessible to non-experts, the speed and scale of intelligent application development will soar. Gartner predicts that by 2025, 10% of all data will be produced by generative AI models [^12].

For organizations that harness these trends effectively, the ability to rapidly fuse human and machine intelligence to solve business problems will become a major competitive differentiator. It will also elevate the roles of business technologists and citizen developers who can bridge the gap between domains and data. Longer term, I expect to see AI-driven development with LCNC further lowering barriers to building powerful applications.

At the same time, organizations pursuing LCNC and AI/ML strategies must proactively address challenges around trust, transparency, and ethics to unlock the full potential of these technologies. IT leaders have a critical role to play in balancing democratization and governance. Executed well, the fusion of LCNC and AI/ML promises to reshape enterprise innovation in the years ahead.

References

[^1]: Gartner, "Gartner Says Cloud Will Be the Centerpiece of New Digital Experiences"
[^2]: MarketsandMarkets, "Low-Code Development Platform Market"
[^3]: US Bureau of Labor Statistics, "Software Developers, Quality Assurance Analysts, and Testers"
[^4]: Gartner, "Gartner Says 69% of Boards of Directors Accelerated Their Digital Business Initiatives Following COVID-19 Disruption"
[^5]: Gartner, "Gartner Forecasts Worldwide Low-Code Development Technologies Market to Grow 23% in 2021"
[^6]: Microsoft, "Overview of AI Builder"
[^7]: Salesforce, "Einstein Bots"
[^8]: Microsoft, "Use AI to quickly add intelligence to your apps with AI Builder"
[^9]: Google, "Introducing AppSheet Automation, an intelligent model driven platform for business process automation"
[^10]: Gartner, "Gartner Says Cloud Will Be the Centerpiece of New Digital Experiences"
[^11]: Gartner, "Future of Apps Must Include Citizen Development"
[^12]: Gartner, "Gartner Identifies the Top 10 Strategic Technology Trends for 2022"