How We Built An AI Launchpad in Under 20 Days on Amazon Web Services
Revolutionising AI Consulting: The WeBuild-AI Journey
At WeBuild-AI, we're on a mission to transform the consulting landscape through the power of artificial intelligence. Our recent launch event marked a significant milestone in our journey, and we're excited to share our story with you. The following blog is an excerpt and summary of all the key topics discussed during our live-stream launch on LinkedIn on the 25th September 2024.
Why We Built WeBuild-AI?
Six months ago, we had what we call our "coffee cup moment." We asked ourselves a crucial question: "Can we build an AI Native Consultancy that uses generative AI, autonomous agents, and cloud to transform the consulting experience and 10x outcomes for customers? And in doing so, can we use these same tools and capabilities to operate our own business?"
This question set us on an exciting path, not just for ourselves, but for our customers as well. We envisioned a future where we could disrupt the economics and commercial models of professional services, delivering 10x outcomes for customers across the globe. Namely, we believe that 80% of enterprise applications can be built 70% faster and operated at 30% of their current total cost of ownership, driven by AI tooling and highly skilled SME’s.
Our research, which included interviews with over 100 C-Suite, Product & Engineering Leaders, revealed that many partners were delivering Gen-AI engagements using traditional approaches with unclear ROI. We also found that data quality was a key challenge, with significant volumes of unstructured data inaccessible to employees. These insights further solidified our belief in the need for an AI-native approach to consulting.
2. Building Pathway with Generative AI
To bring our vision to life, we decided to create Pathway, our AI Launchpad. We wanted to use generative AI not just as a capability, but as a product to accelerate our own development process. This approach allowed us to:
- Reimagine the design sprint for the AI age.
- Achieve over 200 deploys in just 20 days and build the core foundations of Pathway.
- Evolve the 5-day design sprint process, that takes you from concept to working production code with AI.
In terms of the AI-powered design sprint process, we have now encapsulated these techniques into the DNA of how we deliver to customers. We have broken this down as follows:
In this phase, we leverage AI to rapidly gather and analyse vast amounts of data from various sources. Our Gen-AI capabilities in Pathway can quickly sift through academic papers, industry reports, and market trends to provide comprehensive insights. This allows us to:
- Identify key market trends and opportunities.
- Understand user needs and pain points more effectively.
- Generate data-driven hypotheses for product features.
- Accelerate the discovery phase, reducing it from weeks to days.
We always use AI to assist in the creative design process, helping us to:
- Generate multiple UI/UX design concepts based on research insights at warp speed.
- Create wireframes and prototypes rapidly.
- Suggest colour schemes, typography, and layout options.
- Optimise designs for accessibility and user engagement.
This AI-driven approach allows us to explore a wider range of design possibilities in a fraction of the time it would take using traditional methods.
Our AI tools play a crucial role in accelerating the coding process. In doing so we now use generative AI to assist us in:
- Generating boilerplate code and basic structures.
- Suggesting code completions and optimisations.
- Identifying potential bugs and security vulnerabilities.
- Automating repetitive coding tasks.
- Providing context-aware documentation.
This not only speeds up development but also helps maintain high code quality and consistency across the project.
We harness the power of AI to enhance our testing processes. Examples of how we now do this include:
- Automatically generating test cases based on code analysis
- Performing intelligent fuzzing to identify edge cases and vulnerabilities
- Analysing test results to prioritise bug fixes
- Conducting performance testing and suggesting optimisations
- Simulating user behaviour for more comprehensive UX testing
Through the experience of creating Pathway, AI-driven testing now allows us to achieve higher test coverage, identify issues faster, and ensure a more robust final product. Each of these points are practices and methods we now take forwards into our Pathfinder Playbook to transform and deliver outcomes for customers faster.
By leveraging AI in every step of our development process, we were able to significantly accelerate our delivery timeline. For our MVP, we ran this process four times, focusing on one core capability each sprint. This AI-powered approach not only speeds up development but also enhances the quality and innovation of our solutions.
3. Core Pathway Architecture
Our Pathway architecture is built on a robust foundation of AWS services, including:
- AWS Control Tower: Provides us with a rapid secure landing zone capability hooked into AWS Security Hub and AWS Config for product security, compliance and monitoring.
- AWS Bedrock: Provides access to foundation models and APIs for building generative AI applications. Enabling us to deliver a multi-model, multi-modal experience for customers.
- Amazon S3: Handles secure object storage for our data and application components.
- Amazon DynamoDB: Offers fast and flexible NoSQL database services
- AWS Lambda: Enables serverless compute for our functions and API’s.
- AWS DataZone: Provides us with a modern automated data governance approach which leverages generative AI.
- Amazon Q Developer: Enabled us to develop new application and infrastructure code in days, not weeks. Whilst it also helped us in creating technical documentation as we went along.
We used GitHub Functions for CI-CD and rapid delivery to production. This DevOps based approach ensured scalability, reliability, and rapid iteration. Whilst our Pathway architecture is designed to be cloud-native, ensuring scalability, resilience and operational excellence.
4. What We Have Built into Pathway
The development process culminated in the delivery of a number of core Alpha features for Pathway enabling us to create a powerful AI Launchpad with several key components:
Pathfinder offers flexibility and AI power. By supporting multiple AI models, customers can choose the best model for each specific task, ensuring optimal performance and results. The multi-modal capabilities allow for seamless integration of text and code, streamlining research, product design , requirements and development processes. Whilst RAG integration enhances the AI's ability to provide accurate, context-aware responses, while the prompt generator helps users craft effective queries, maximising the AI's utility.
Navigator Prompt Library:
The Navigator Prompt Library is a game-changer for efficiency and knowledge sharing. It provides a curated set of industry-specific prompts, allowing customers to quickly leverage best practices and proven approaches. The ability to add, manage, and filter prompts enables organisations to build their own institutional knowledge base, preserving expertise and accelerating problem-solving across teams. Over time, the Prompt library will become increasingly agentic in nature.
The Knowledge Hub serves as a single source of truth, capturing and preserving institutional knowledge that might otherwise be siloed or lost. Over time, we will introduce AI-powered search capabilities making it easy for employees to find the information they need quickly, improving productivity and decision-making. Version control ensures that everyone is working with the most up-to-date information, while collaborative editing features foster knowledge sharing and refinement. By integrating with external sources, the Knowledge Hub also helps organisations stay current with industry trends and best practices.
5. What the Pathway Reinforced For Us
Our journey in building Pathway and launching WeBuild-AI has reinforced our assertions about the power of generative AI, whilst providing us with further valuable insights:
Enhanced T-shaped Skills: While product and agile methodologies remain crucial, our AI tools have enabled people to develop broader skill sets than we initially anticipated. This has led to more versatile team members who can contribute across various aspects of projects.
Importance of Subject Matter Expertise: Despite the power of AI, we found that SME skills and knowledge remain crucial, especially for subjective and detailed work. The combination of AI capabilities and human expertise yields the best results.
Component-Based Development: We discovered that developing components yields better results than focusing on whole screens at a time. This approach allows for more flexibility, reusability, and faster iterations.
Rapid Iteration and Deployment: Our ability to achieve over 200 deploys in just 20 days highlighted the power of our CI-CD pipeline and the agility of our development process. Those organisations who have shifted to a DevOps approach will adapt much quicker than their competition to the AI space race.
AI-Powered Design Sprints: Reimagining the design sprint for the AI age has significantly accelerated our development process, allowing us to go from concept to working production code in just 5 days. We now want to share this experience with our customers.
Data Quality Challenges: Our research revealed that data quality and accessibility remain significant challenges for many organisations. This reinforced the importance of our data integration capabilities in Pathway.
Need for Ethical AI Practices: Throughout our development process, we recognised the critical importance of establishing guardrails to demonstrate AI ethics. This has become a core part of our approach and a key consideration in all our work. We have built an initial set of safety controls that will need to expand further in time.
6. What's Next: Future Roadmap and Features
Looking ahead, we see exciting developments on the horizon:
Agentic AI Dominance: We believe that agentic AI will become dominant, with 2025 potentially being the "Year of Agents." This shift will enable more autonomous and sophisticated AI applications.
Industry Consolidation: We anticipate a consolidation of new AI product and model providers, leading to more standardised and robust offerings in the market.
Enterprise AI Adoption: We expect to see enterprises increasingly using AI to uplift existing technology and streamline processes, as well as launching AI-first business units or entirely new ventures.
For Pathway specifically, we plan to add several new features over the coming months:
Accessibility Features: Ensuring that our platform is accessible to all users, regardless of any impairment or ability is crucial for inclusivity and maximising the democratisation of Pathway and AI in general.
Enhanced Team Collaboration Tools: Improved collaboration features will facilitate better knowledge sharing, increase productivity, and enable more effective problem-solving across teams.
API Integrations: These integrations will streamline workflows, allowing for seamless incorporation of design tools into the AI development process.
Knowledge Base Analysis (e.g., GitHub and SharePoint): This feature will enable organisations to leverage their existing knowledge repositories, enhancing the AI's ability to provide context-aware solutions.
Expansion of Navigator Prompt Library: Covering more industries will make Pathway more versatile and valuable across different sectors.
Verified/Rated Prompts System: This will help users identify the most effective prompts, improving the overall quality and reliability of AI-generated outputs.
Enhanced Search Capabilities: Improved search functionality will make it easier for users to find relevant prompts and information, increasing efficiency.
Conversion of Chained Prompts to Agents: This advancement will enable more complex, multi-step tasks to be automated, further enhancing productivity and capabilities.
7. Become a Design Partner
We're excited to announce that we're looking for design partners for our 2024 cohort. As a design partner, you'll get:
- Early access to Pathway's latest features
- Direct collaboration with our product team
- Strategic insights to keep your organization at the forefront of AI adoption
By becoming a design partner, you'll have the opportunity to shape the future of AI consulting and gain a competitive edge in your industry. You'll be at the forefront of AI innovation, helping to refine and enhance Pathway to meet the evolving needs of businesses in the AI era.
If you're interested in being part of this exciting journey and contributing to the next generation of AI-powered consulting tools, we invite you to join us as a design partner.
Sign up to become a Pathway design partner by visiting: https://www.webuild-ai.com/pathway-signup
8. In Closing
At WeBuild-AI, we're not just embracing the AI revolution – we're driving it. Join us on this exciting journey as we redefine what's possible in the world of consulting and enterprise AI. Together, we can unlock the full potential of AI to transform businesses and create unprecedented value.
For those of you who were able to watch the stream live, thanks for joining us. For those who weren’t able to, we hope you enjoyed the replay above.
