In this episode of Innovators Inside Podcast, Ian Bergman speaks with Ingrid Curtis, CEO of Sparq, about why so many artificial intelligence initiatives fail to create meaningful business impact.
The problem is often not the technology itself. Companies launch pilots without identifying the right business challenge, continue iterating without shipping, or focus only on small efficiency gains instead of larger growth opportunities.
For founders, operators, and innovation leaders, the conversation offers a practical framework for moving AI from experimentation into production.
Here are the five key takeaways from their conversation:
Many companies begin their AI journey by looking for projects that are already defined and easy to launch. Ingrid describes these as “shovel-ready” projects.
The problem with this approach is that the project may be convenient without being strategically important.
Instead, leaders should identify what Sparq calls “problem zero.” This is the business problem that would create the greatest impact if the company could solve it.
That may involve:
Increasing revenue
Improving the customer experience
Transforming a critical operational workflow
Creating a meaningful competitive advantage
Removing a major barrier to growth
The strongest AI strategy does not begin with asking where AI can be added. It begins with asking which business outcome matters most.
Automating one step in an existing process may create incremental value, but it does not necessarily lead to transformation.
AI gives companies an opportunity to reconsider how work should happen from beginning to end. Instead of improving one task inside an outdated workflow, leaders should step back and examine the entire process.
This matters because AI agents, natural language interfaces, and new software development tools may remove the need for traditional interfaces and handoffs.
The right question is not simply, “How can we automate this task?”
It is, “If we designed this workflow today, using current AI capabilities, how would it work?”
One of the biggest reasons AI pilots stall is that teams keep improving them without putting them in front of real users.
Ingrid compares this cycle to “doom scrolling.” Teams continue prompting, refining, and rebuilding until the solution becomes more complicated or less useful.
Her recommendation is to recognize when a solution is good enough to test. In many cases, 85 percent may be enough to move forward.
Shipping does not mean releasing an unfinished product without safeguards. It means getting the solution into the hands of real users, collecting feedback, and allowing actual behavior to guide the next investment.
A smaller solution in production can teach an organization more than a polished pilot that remains isolated from the business.
Sparq uses an ACT framework to help organizations move AI solutions into production:
Awareness
What does the system know about the business?
This includes the data, context, signals, and information the AI needs to make useful decisions.
Capability
What should the system actually do?
Leaders must define the workflow, reasoning process, actions, and intended outcome.
Trust
How can the system operate safely in production?
This includes governance, observability, accountability, security, and clarity about human oversight.
Many AI pilots focus heavily on capability while neglecting awareness and trust. Without the right context and safeguards, even an impressive prototype may never become a dependable business tool.
Many organizations begin with low-risk AI use cases such as document processing, coding assistance, administrative work, or workflow automation.
These projects can create value, but they are often limited to efficiency gains.
As Ingrid explains, companies cannot cut their way to growth. The larger opportunity is to use AI to identify new revenue, improve products, create better customer experiences, and rethink how the business competes.
Human creativity remains essential here. Competitors may have access to the same AI models and similar data, but they will not necessarily make the same strategic decisions.
The organizations that win will combine AI capabilities with human judgment, industry knowledge, and a willingness to challenge existing assumptions.
AI transformation is not primarily a technology problem. It is a strategy, leadership, and organizational change problem.
Companies must choose meaningful challenges, rethink workflows, ship solutions to real users, build trust, and connect AI investment to measurable outcomes.
The goal is not to become AI-native overnight. It is to create a faster and more adaptable way of working, one production-ready solution at a time.
Listen to the full conversation with Ingrid Curtis to learn how Sparq is helping organizations move from AI pilots to real business transformation.
Have a question for a future guest? Email us at innovators@alchemistaccelerator.com to get in touch!
Timestamps
01:20 🎙️ Meet Ingrid Curtis, CEO of Sparq
02:10 🧩 Where strategy, engineering, data, and AI meet
05:09 📉 Why AI investment is not always producing ROI
06:33 🎯 Choosing the right business problem
09:13 🤖 How AI is changing software development
11:21 🔄 Moving beyond traditional agile development
12:53 🔍 Finding your organization’s “problem zero”
14:11 🏢 Why AI transformation stalls inside enterprises
16:01 🧠 Human creativity as a competitive advantage
18:53 📊 Where AI is creating value today
21:45 🚀 Moving from problem zero to production in 90 days
23:18 ⚙️ The ACT framework: awareness, capability, and trust
25:25 💡 Building an entrepreneurial approach to AI
26:38 🛠️ Helping teams learn through experimentation
29:16 📱 When AI iteration becomes “doom scrolling”
31:05 🧭 What an AI-native company could look like
34:37 👥 How AI may change organizational accountability
39:01 🤝 What makes an ideal innovation partner
40:28 🏗️ Challenging assumptions before building
42:19 🔮 What AI transformation could look like one year from now
44:42 🌍 AI as an existential threat and opportunity