All artists and designers experience a creative block, leading to lost time and revenue, but generative AI presents the opportunity to change the way designers start.
Opportunity
Creative block poses a significant challenge for artists and designers, impacting productivity and innovation. This issue is particularly critical in the design industry, where the ability to deliver fresh, innovative designs is paramount.
Case Study
Market Dynamics
2023 has seen the fast rise of generative AI, not only in terms of the performance of these models but also in the applications that are emerging to make use of this game-changing technology. However, we have yet to see big players emerge with real large-scale applications. In that sense, the timing seems right for innovative solutions from the gen-AI models. In the design space, open-source tools like Stable Diffusion emphasize customization, indicating a competitive landscape. Our product aims to blend AI-driven inspiration with user-guided customization, addressing creative block challenges and offering hands-on creative control.
Solution
Uncertainty, testing strategies, and roadmap for moving forward and risk mitigation
Areas of Uncertainty and Assumptions
| Assumption | Description | Plan |
|---|---|---|
| User Values Customization at the End of the Design Process | The assumption that customization adds significant value and is a crucial aspect of the design process hasn’t been explicitly tested | Conduct usability tests and gather feedback on the prototype to assess whether users indeed appreciate and find value in customization options |
| Cloud Costs Balance with Pricing | The chosen pricing model effectively balances covering cloud costs while remaining attractive to users | Conduct pricing surveys or A/B testing to understand user willingness to pay. Gather insights on what pricing points are perceived as fair and valuable by the target audience |
| Programmatic Generation of PSD is Technically Feasible | GSPhotoCo would be able to programmatically generate a compliant PSD file from multiple and varied image and text inputs | The team was able to validate this technical assumption via a short javascript application which generated a PSD compliant file. More information can be located here. |
Market Analysis
Competitive Analysis
Leading Signals
User Engagement Metrics:
- Increased engagement with customization features
- Positive feedback on usability and user satisfaction
- Growing subscription numbers and recurring revenue
Logistics
Why the Visual Studio and how to build it
Alternatives
The table below documents the potential alternatives to solving creative block for artists as well as the drawbacks that ultimately led us away from them.
| Alternative | Description | Drawbacks |
|---|---|---|
| Templates | Professionally curated designs and ‘plug and play’ options for any design type |
|
| Editor | A full-blown image editor, in the browser |
|
| Creativity Stimulator | A short game/video/meditation-like guide post-purchase that helps users initiate their ‘creative flow’ |
|
Costs
Generative AI has the potential to render GSPhotoCo’s traditional media repository antiquated, and possibly put the company out of business. The new technology changes the way designers work, meaning traditional methods of creation, such as templates and pure editors, will take a less prominent role in future design. However, generative AI is still in its infancy and requires high-quality data to get there. Current generative AI and stable-diffusion-esque models produce assets of varied quality which are uneditable in traditional editing tools. GSPhotoCo must leverage its curated repository of high-quality, human-made, images, fonts, and artistic assets to establish a foothold in the changing design industry before generative AI models can close the quality gap.
Technologies
The GSPhotoCo Visual Studio is a straightforward web application with distinct technical requirements to allow the project to be rapidly prototyped and quickly iterated upon. For this reason, Visual Studio should be a separate application from the existing GSPhotoCo, even if sharing some infrastructure, industry, and code (ie. authentication services).
Below is a non-prescriptive table of required technology layers, and practical selections mitigating risk by optimizing for low engineering lift (ie. rapid prototyping/rapid iteration).
| Level | Technologies | Description |
|---|---|---|
| Frontend |
|
Simple SPA statically served from a selected cloud provider (ie. AWS Amplify) |
| Backend |
|
Serverless backend for rapid prototyping, simple development and isolated deployments |
| Authentication |
|
Reuse authentication and session endpoints from existing GSPhotoCo cloud solution |
| Payments |
|
Implement payments through Stripe and PayPal; two most popular, secure, checkout methods |
| Storage |
|
Implement document database for rapid prototyping and simple blob storage for asset use and construction |
| ML Pipeline |
|
Industry-standard ML pipeline augmented with OpenAPI endpoint; trained existing repository assets |
| Development Operations |
|
Industry-standard Development operations, automated testing, and continuous deployment infrastructure |
Timeline
| Weeks 1-3 | Market research, user needs identification, competitor analysis, and project scope definition |
|---|---|
| Months 1-3 | Prototype design and development, basic AI integration, testing with partner users, and iteration based on feedback |
| Quarters 1-3 | Beta testing, feedback analysis, improvements, comprehensive launch strategy development, and market-ready product achievement |
