The challenges of creating an AI data platform — Cameron Turner // Kin + Carta

Show Notes

• 01:23
  The challenges of creating an enterprise AI data platform

  Building an AI data platform poses challenges such as defining the use case, avoiding technical debt, and ensuring extensibility without being locked into a single technology provider. It's crucial to build a solution that can grow with the business over time.
• 05:06
  The importance of building systems that connect people

  Considering the human element is essential because data is a byproduct of human activity. To truly harness the power of data, it's crucial to empower individuals, enhance workflows, and build systems that facilitate human connections, rather than oracle-like systems.
• 07:54
  An overview of Louisa.AI and how it creates connections

  Louisa.AI, an early-stage startup, analyzes an organization's data to foster meaningful connections and enable "planful serendipity. While the goal is to enhance connections using data, it requires thorough data source assessment to get to that point.
• 08:56
  AI integration in enterprises

  AI integration into an enterprise should be seen as a spectrum, ranging from full automation to activities that remain fully analog and reliant on human interaction. Understanding where a specific business problem falls on this spectrum is crucial to avoid potential issues.
• 10:58
  The transformative impact of AI on the nature of work

  Automation mainly targets non-creative tasks, notably impacting knowledge workers. Manual tasks remain unaffected, and AI has improved the efficiency of creating machine learning models through automated machine learning, making it faster and more accessible.
• 13:08
  Advancements in supervised machine learning and human AI collaboration

  Models created quickly now primarily involve supervised machine learning, focusing on data with existing answers to predict and recommend for future data. Scalable cloud services enhance accuracy, but human involvement is required for interpreting results.
• 15:01
  Considerations when developing a bespoke AI data platform

  Building bespoke AI data platforms involves addressing data quality, data provenance, stewardship, harmonization, and latency considerations. It begins by understanding the desired data outcomes and then tailoring the technical stack accordingly
• 17:11
  Prioritizing extensibility over perfection in AI solutions

  Building the perfect solution isn't necessary, and long-term projects risk obsolescence due to the rapid pace of change. The questions that you need to answer will change over time, so prioritizing extensibility is crucial in adapting to evolving business needs.
• 17:46
  Using data to solve business problems

  Kin + Carta views data as a piece of the larger puzzle, impacting experiences, products, business change, and strategy. They assist companies with answering critical questions, creating long-term plans, and exploring cloud migration possibilities through data work."
• 19:15
  Solving customer pain points with data driven precision

  Addressing a specific customer pain point through data and connecting them to a solution is a hallmark of successful data projects. When data represents the problem and predicts the recommendations directly to a solution for the original stakeholder, it fuels business growth.
• 23:42
  Leveraging cloud opportunities for data cleansing

  The cloud migration process should include data modernization to fully leverage its potential. Cleansing and sorting data before migration is essential for maximizing the benefits of cloud technology.
• 24:32
  AI tool implementation and focusing on achievable goals

  Even with all the AI tools and technology on the market, its best to trust your instincts and not get caught up in the hype. Instead, focus on achievable goals and tangible business impact when it comes to implementing AI within your organization.
• 25:21
  Balancing perfectionism with pragmatism in AI data platform development

  The greatest challenge in creating an AI data platform is not striving for perfection but rather building a minimum viable AI data platform that delivers results quickly. Starting small, proving value along the way, and avoiding long, uncertain modernization projects are key to success.