Mindspark helps employees work smarter with AI agents customized to their workflows

The top GenAI solutions had too many tradeoffs, so SHI built their own.

Challenge:

In 2022, SHI CEO Thai Lee presented her leadership team with a challenge: use the power of AI to achieve a 20% increase in productivity. This wasn't just another tech initiative; it was a response to the transformative user experience that was being pioneered by tools like ChatGPT. The problem was, there weren’t many commercially available AI solutions that could meet SHI's needs.

Early attempts with existing AI tools failed to gain traction; they didn't offer the kind of granular control SHI wanted, and the per-user pricing added up quickly. A deep dive into the numbers showed that most employees using AI for basic tasks like meeting summaries were only getting about $4.68 worth of value each month, while a tool like Microsoft’s Copilot cost a steep $30 per user. The team also found that employees were already experimenting with external AI tools like ChatGPT, which raised concerns about security and data privacy.

SHI needed a solution that would not only be cost-effective, but would also allow them to own their data, models, and infrastructure and build internal expertise around AI.

Solution:

The solution came in the form of Mindspark, an in-house-developed AI platform. Mindspark was built to be a collaborative partner, empowering employees to work more effectively.

SHI approached the development of Mindspark with a strategic, iterative process focused on usability, scalability, and impact. The first step was forming a dedicated AI team to design a flexible platform capable of integrating multiple AI models and adapting to diverse business functions. The developers ensured that Mindspark would function as a usage-based cost model, meaning SHI only incurred expenses when the AI was actively used. This approach established a direct correlation between cost and value, unlike third-party tools that imposed a fixed per-user fee.

To drive adoption and uncover high-impact use cases, SHI implemented a citizen development model, which enabled employees from all departments to create their own AI-powered agents. Mindspark provided a no-code environment where employees could design and deploy AI agents, encouraging an internal culture of AI literacy and innovation. This decentralized approach allowed SHI to scale Mindspark organically, with users developing and sharing AI-driven solutions tailored to their daily work.

SHI also developed governance structures and business intelligence dashboards to monitor AI adoption and effectiveness. These dashboards provided insights into which AI tools were being used most frequently, which departments were seeing the greatest benefits, and how AI was transforming workflows. By combining qualitative user feedback with quantitative usage data, SHI was able to continuously refine and optimize Mindspark.

A key component of Mindspark’s success was its modular design. Instead of being locked into a single AI model or vendor, SHI structured the platform to integrate multiple AI models, ensuring adaptability as technology evolved. Robust security measures were implemented to address data privacy concerns and prevent employees from resorting to unapproved AI tools.

Outcome:

The deployment of Mindspark at SHI has yielded significant productivity gains, cost savings, and cultural shifts in AI adoption. Mindspark achieved a 75% adoption rate across the company, with about 4,000 active users. Internal surveys revealed that Mindspark saves these employees an average of 3.5 to 4 hours per week, translating to an estimated 10% increase in overall productivity to date.

Mindspark’s decentralized development model led to the creation of over 300 custom AI agents, addressing diverse needs across departments such as sales, marketing, finance, and IT support. These AI-driven solutions have not only reduced workload burdens but also accelerated key business processes. For example, contract analysis, which previously took up to four hours per document, was reduced to just 30 minutes. In some cases, the efficiency gains were so striking that clients questioned how the team was able to respond to requests so quickly.

Mindspark has also dramatically lowered AI adoption costs. The average cost per user is just $2.75 per month, compared to the $20 to $30 charged by other popular AI tools. By shifting to a pay-per-use system, SHI has achieved a tighter correlation between AI investment and value creation, ensuring a higher return on investment.

Beyond cost and efficiency, Mindspark has empowered employees to take ownership of AI-driven innovation. By allowing individual contributors to identify and solve problems with AI, SHI has cultivated a workforce that is not only more efficient but also more engaged in the company’s technological evolution. The internal success of Mindspark has been so profound that SHI has now begun offering it as a product for external clients, leveraging its two-year head start in enterprise AI adoption.

SHI’s decision to develop Mindspark in-house has proven to be a transformative move, demonstrating the advantages of cost control, customization, and user-driven innovation. By prioritizing flexibility, security, and ROI, SHI has not only enhanced its internal operations but also positioned itself as a leader in enterprise AI solutions. Mindspark exemplifies how organizations can leverage AI to maximize efficiency, reduce costs, and foster a culture of continuous innovation—all while retaining full control over their AI journey.