It's only human to want to optimize the decisions we make on a daily basis. As technologists, we build digital solutions that save time and create efficiencies to optimize communication and resourcing. We rely on experts to help guide solutions but as data has become more democratized, everyone has access to... too much data. It's paralyzing.
Automation has promised efficiency at the expense of behavior change. This change if not designed with intention, will replace human capability instead of augmenting expertise. The value of technology is more than performance and efficiency.
Digital transformation is about people. A human centered approach means widening solutions across teams and business units and connecting people to open data pipelines. What if we approached digital transformation as enabling eco-systems of conversations between intelligent agents to support decision makers? How might we ask more outcomes driven questions and make fewer decisions as we search for unknowns?
As we move from trust to reliance in intelligent systems, design principles need to enable searching for what we don't know we don't know:
1 — Signal
H: Make me aware of when to take action and nudge me in the right direction based on behavioral prompts (e.g. loss aversion) that I can tweak. M: Predict micro-actions by inferring intent from situational data models.
2 — Awareness
H: Keep me accountable by monitoring progress. Help me recognize cognitive biases and understand the impact of actions taken by showing me potential outcomes that help me weigh decisions. M: Detect and extract signals from behavioral data models.
3 — Adapt
H: Give me meaningful anchors that help me change bad habits or avoid reactive decisions and help me adjust thresholds to fit situational needs. M: Present prescriptive models that can adapt to individuals with weights and thresholds.
4 — Learn
H: Alleviate uncertainty and build confidence in making the right, good decisions. Show me my trade-offs to help me understand implications of action or inaction. M: Generate predictive remedy models that describe potential outcomes. Find the good local optimum and show context of why.
E.g. There's a high likelihood of patient population X reacting adversely to therapies X,Y,Z based on these X number of publications.
H: Give me strategies within my ability to execute and fine-tune within guard-rails. M: Tuning parameters and train the data to get more accurate profiling models.
H: Build confidence that outcomes that are ethical and in my best interest. M: Embed policies as features of prediction models.
To learn how these design principles could influence your digital transformation initiatives or augment your digital tools, drop us a note.