🤝Toronto AI Product Design: Scoping My First Engagement

When a new client approaches me for an AI product design engagement, one of the immediate challenges is often a lack of clarity around their actual needs versus their perceived solutions. They might present a vision for an AI agent, for instance, without a clear understanding of the underlying business problem or the current state of their organizational data and processes. This initial ambiguity, if not meticulously addressed, can lead to scope creep, mismatched expectations, and ultimately, a project that fails to deliver meaningful value.

My role, particularly in these early stages, shifts from a pure designer to something more akin to a strategic partner and investigative journalist. It's about asking the right questions, listening intently, and co–creating a foundational understanding that serves as our north star throughout the engagement. This upfront investment in precise scoping is perhaps the most critical determinant of a project's eventual success and its impact on the client's business outcomes.

Understanding the Core Problem, Not Just the AI Desire

Clients often come with a specific AI solution in mind—a financial services copilot for loan officers, perhaps, or a generative UI for report creation in a government agency. My first step is always to gently pivot the conversation back to the fundamental problem they are trying to solve. What inefficiencies exist? Where are the bottlenecks? What user pain points are currently unaddressed or poorly addressed?

For instance, an enterprise might believe they need a 'smart' search engine for vast internal documentation. Upon deeper conversation, we might discover their actual problem isn't search per se, but rather a lack of structured knowledge, inconsistent metadata, and a fragmented content creation process. The AI solution then shifts from a pure search play to something more encompassing, potentially including a knowledge management strategy with AI–assisted curation and content generation, all designed within a new information architecture.

Deconstructing the Current State and Desired Future

Once the core problem is articulated, we map out the current state. This involves understanding existing workflows, the technology stack in place, data availability and quality, and––critically––the human users and their current capabilities. For a healthtech client exploring a clinical decision support tool, this means understanding how doctors currently access patient data, the regulatory environment for data usage, and the existing decision–making protocols.

Simultaneously, we define the desired future state, not just in terms of technical capabilities, but in terms of measurable business impact and improved user experiences. What does success look like? How will we know if the AI product has achieved its objectives? This often involves defining key performance indicators (KPIs) and establishing baselines against which we can measure progress. For an AI product, this invariably includes considerations around accuracy, interpretability, and trust, leading into how we might design for human–in–the–loop processes and robust evaluation frameworks.

The art of scoping an AI product engagement lies in transforming vague aspirations into concrete, actionable steps with measurable outcomes.

Identifying Opportunities and Constraints for AI

With current and future states defined, we can pinpoint where AI genuinely offers a unique advantage. Not every problem needs an AI solution, and a significant part of my role is to guide clients away from expensive, unnecessary AI implementations. I look for areas where AI can automate repetitive tasks, augment human capabilities, or generate novel insights at scale.

Constraints are equally important. These include technical limitations—can their existing infrastructure support a large language model deployment? Or data limitations—do they have the volume and quality of data required for effective model training? There are also organizational constraints, such as internal AI literacy, budget, security requirements, and the often–overlooked change management hurdles for adopting new AI tools within an established enterprise.

Defining Project Deliverables and Milestones

Based on the preceding discovery, we work together to define concrete deliverables. These might include user research reports focused on human–AI interaction, detailed AI UX flows, wireframes or prototypes for generative interfaces, a technical feasibility report, or even a roadmap for an AI design system. Each deliverable is linked to a specific phase of the engagement and has clear acceptance criteria.

Establishing clear milestones and regular check–ins is crucial. This iterative approach allows us to validate assumptions, mitigate risks early, and adapt the scope as new information emerges. For highly sensitive projects, such as those involving government agents or financial copilots, this also means dedicating substantial time to discuss ethical AI considerations, bias mitigation strategies, and the design of accountable systems before any development begins.

Structuring the Engagement and Pricing

Finally, with a clear understanding of the problem, desired outcomes, constraints, and deliverables, I can structure the engagement. This involves proposing a phased approach—perhaps an initial discovery and strategy phase followed by an iterative design and prototyping phase. I provide a transparent breakdown of the effort involved, the associated investment, and the expected timeline.

My goal here is to present a scope that is both comprehensive and flexible. Comprehensive enough to address the identified problem effectively, yet flexible enough to accommodate the inherent uncertainties of AI development. It’s about building trust through transparency and ensuring both parties have a shared mental model of what we are setting out to achieve and how we will measure its success.

Scoping an initial AI product design engagement, especially in complex enterprise or government contexts, is a nuanced exercise in strategic thinking, empathetic listening, and pragmatic planning. It sets the stage for a productive partnership and ensures that the innovative potential of AI is directed towards truly valuable and sustainable outcomes, rather than just chasing the latest technological trend. When we get this right, the foundation for truly transformative products is laid, and that’s a rewarding space to operate in.

Related reading

Designing AI-Native Websites Versus Chatbot Bolt-Ons · Designing AI Lead Qualification for B2B Services · When Adding AI to a Small Business Stack Isn't Prudent

Where this shows up in my work

AI consultant — Toronto & Los Angeles · Hire a senior service designer