Accelerate Product Teams with Generative AI
Learn how AI can be used throughout the product lifecycle to augment human performance and create better products faster.
From optimistic (it’s going to make me much faster at my job) to panic-inducing (it’s going to take all our jobs), there is a wide range of feelings and opinions on the impacts of generative AI. However, AI’s ability to enhance human output has especially exciting potential in the world of digital product development. Leveraged strategically throughout the product lifecycle (ideate, design, build, test, deploy, and run), AI can generate, enhance, and evaluate various types of content to augment and accelerate human capabilities, enabling teams to create better products faster.
Definitions
Digital product development is a complex, continuous, and iterative process that involves many activities, from ideation to deployment. The activities are happening concurrently as the team delivers value to their customers. Product teams face various challenges and trade-offs: balancing functionality, performance, cost, time, and quality. Product teams should always be looking for ways to improve how they deliver value through their value chain while balancing these factors.
Generative AI is a branch of AI that can create new content (e.g., audio, code, images, text, simulations, and videos). Generative AI can explore many possible solutions to a given problem and find the optimal one based on predefined criteria, and it can produce novel and nonintuitive solutions that humans may have missed. Generative AI has already been used to design drugs, materials, chips, and structures that are more efficient, effective, and sustainable than conventional methods. It has also been used to augment and accelerate creative work in marketing, media, and entertainment.
Now that we have a common understanding of what we’ll be discussing, let’s dive into ways in which generative AI can accelerate product development!
Ideate
During ideation, product teams generate and evaluate ideas for new products or features; it requires creativity, collaboration, and customer insight. Here’s how AI can help:
- Generate new ideas based on customer needs, market trends, or competitor analysis. Product teams can use ChatGPT to brainstorm new product names, slogans, descriptions, or use cases.
- Enhance existing ideas with additional details or variations. For example, DALL-E 2 is a generative AI system that can create images based on natural language text (additional tools include Midjourney and Stable Diffusion). Product teams can use DALL-E to visualize their ideas or generate different versions of their concepts.
- Develop product OKRs. Generative AI systems can be very effective in ideating potential metrics by which to evaluate the performance of your product. ChatGPT has demonstrated this across both consumer-facing and internally facing applications, and even platform and infrastructure teams.
Design
Design activities include creating prototypes or mock-ups of products or features. These activities require technical skills, aesthetic sense, and user feedback. Generative AI can help product teams in several ways:
- Generate prototypes or mockups with additional features or functionalities. Tools like Adobe with embedded generative AI can rapidly visualize the designer’s or product owner’s ideas, or enhance ideas using style transfer to apply to different styles or themes to a product design. Or the designer can use tools like Midjourney or Dall-E to generate mock-ups or wireframes as Alley Lyles-Jenkins highlights in her post here.
- Draft user stories, acceptance criteria, and tech specs. ChatGPT is fairly effective in getting 70%–80% of a user story complete with acceptance criteria and technical specifications. The product owner or business analyst can then review and supplement it with specific business rules/logic. I’ve found that it can reduce time spent writing user stories by 50%–60% and remind me of minor details that I otherwise might have missed.
Build
During build, the teams are developing and implementing their products or features. Teams are writing code, and understanding across various technology stacks is likely required. Generative AI can help product teams in several ways, with tools like GitHub Copilot roaring out of the gate:
- Generate code based on natural language specifications or examples. For example, generative AI systems can use code synthesis to translate natural language descriptions into executable code or use code completion to suggest possible code snippets based on the context. Some people have even used ChatGPT to build entire applications. The playing field of different skill sets is partially leveled with nontechnical people having the ability to instantly produce viable, working solutions.
- Enhance code with additional functionalities or optimizations. Generative AI systems can use code optimization to improve the performance, efficiency, or quality of code by applying various techniques such as refactoring, parallelization, or compression.
- Evaluate code based on quality assurance or testing results. For example, generative AI systems can use bug detection to identify and fix errors or vulnerabilities in code, or use code completion to suggest possible code snippets based on the context.
Test
Testing is where product teams verify and validate their products or features. This requires testing skills, user feedback, and debugging tools. Generative AI can help product teams during the testing phase in several ways:
- Generate test cases based on user requirements or scenarios. ChatGPT can quickly generate test scripts or scenarios based on user stories or specifications.
- Enhance test cases with additional parameters or variations. Generative AI systems can also use data augmentation to create synthetic data that can be used for testing purposes, or use data mutation to introduce noise or errors into the data to test the robustness of the product.
Deploy
Product teams launch and distribute their products or features as part of a deployment, with varying degrees of frequency. Deployments bring the broader ecosystem into focus, but are not a point-in-time activity. Marketing strategy, training, and user adoption are other key activities related to deployment and are excellent candidates for acceleration with generative AI:
- Generate deployment scripts based on deployment environments or configurations. Generative AI systems can use code synthesis to create deployment scripts that can automate the deployment process across different platforms or environments.
- Enhance deployment scripts with additional features or functionalities. Generative AI systems can, as mentioned earlier regarding the build phase, use code optimization to improve the deployment speed, reliability, or security by applying various techniques such as compression, encryption, or obfuscation.
- Draft communications to stakeholders. Generative AI can be used to draft messaging based on vision and key themes, accelerating the work that the change management team is doing.
Run
Ideally, the product team is monitoring and maintaining the product they built. Run activities require monitoring skills, maintenance strategies, and user retention. Product teams can leverage generative AI in the run phase to:
- Generate and enhance monitoring dashboards based on monitoring metrics or goals. Generative AI systems can use data visualization to create interactive dashboards that display the key performance indicators (KPIs) or objectives and key results (OKRs) of the product. Tools such as Generative BI can provide insights into the product performance, usage, or behavior, and generate recommendations for improvement or optimization.
- Provide tier 1 support to users. Foundational models can be enhanced with specific data to provide the first level of support to users. Better yet, they can support digital humans that provide a more empathetic experience.
Conclusion
Generative AI is a powerful technology that can meaningfully accelerate product development. It can help product teams through many activities of the product development lifecycle by generating, enhancing, and evaluating various types of content. It can also produce novel and nonintuitive solutions that humans may have missed otherwise.
However, generative AI also comes with some challenges and limitations. It has potential to generate inaccurate, biased, or unethical content that may harm users or society. Product teams need to be careful and responsible when using generative AI and always verify and validate its outputs. One tip for avoiding negative outcomes is to ask ChatGPT to reference sources to ensure you can cite effectively and validate any facts. Users also need to be careful about what data or information is entered into the tools. Finally, a product team not delivering value or heading in the wrong direction is not going to be fixed by AI; AI may actually create more spin for dysfunctional teams.
Generative AI is not a magic bullet that can replace human creativity, intelligence, or judgment. It is a great tool that can augment and accelerate human capabilities and enable us to create better products faster.
Slalom is a global consulting firm that helps people and organizations dream bigger, move faster, and build better tomorrows for all. Learn more and reach out today.
