For many enterprises, the promise of AI has outpaced reality. Data sits in silos, and tools are deployed without a coherent strategy. Fortude, a global digital solutions company, now positions itself at the centre of that gap.
Echelon sat down with Chief Technology Officer Surangana Sarathchandra, SVP of Digital and Microsoft Solutions Gaurika Wijerathne, and VP of Products and Automation Harshana Kuruppu at Fortude to explore how the company is making AI practical, proactive, and grounded in strong data foundations.
What drove Fortude’s shift from project-based delivery to building across the enterprise, and where does innovation fit in?
Surangana Sarathchandra: Four years ago, we were doing contained engagements: an ERP implementation, a software component for a specific business unit, a dashboard.
Fortude now cuts across the entire enterprise.
We start with an advisory and map the digital landscape, identifying which processes can be optimised, automated or eliminated. From there, we layer in core solutions: ERP, enterprise software, data and analytics platforms. An ERP upgrade becomes an opportunity to improve data consumption, introduce automation or deploy AI agents.
Efficiency gains depend on the task. Steps requiring human interaction stay as they are. Low-value work, like structuring documentation, can be automated. In some cases, gains of 70–80% are achievable. Across a full process, 20–30% is a reasonable baseline.
What does it actually take to innovate from the inside out, and how does Fortude stay disciplined rather than chasing hype?
Harshana Kuruppu: Innovation starts with understanding and solving our customers’ pain points. Teams on the ground surface pain points that feed our product funnel. We prioritise this carefully before committing to a full product journey.
The second starting point is internal: a programme called ‘Customer Zero’, where we build solutions to address our own pain points. We take these solutions into production internally and, where applicable, lift and shift them to our customers. Neither path is straightforward. When we built Fortest AI (AI-enabled automated software testing for ERP systems and business apps built in-house), we had to go back to the drawing board twice before a successful third attempt.
Charlie and Fortest represent a shift from AI that assists to AI that acts. What does that look like in practice?
Harshana: Charlie began as a simple RAG (Retrieval-Augmented Generation) agent. It has since evolved into Fortude’s core AI framework, with a suite of AI agents for enterprise automation.
Fortest addresses a critical and costly ERP pain point: manual testing. It automatically records business process flows and executes them to ensure that existing functionality remains intact and nothing is broken during changes or upgrades.
A release management agent takes this further. Monthly vendor updates typically take customers two weeks to analyse, and now that’s compressed into hours.
With Fortest, we are now transforming ERP testing efficiency at scale by reducing release analysis time from 2 weeks to just a few hours, cutting complex script creation from 12 hours to 15 minutes with AI, and bringing manual testing cycles down from weeks to hours. This has been proven across our global customer base, with real gains in speed, quality, and delivery confidence.
Do you agree that AI is only as powerful as its data? How does that fit into Fortude’s story?
Gaurika Wijerathne: Yes. Data is the foundation, but it’s only part of the equation. What truly matters is how that data is structured, refined, and given meaning. Raw datasets alone have limited value.
The real impact comes from layering. Starting with raw inputs, progressing to a structured middle layer, and ultimately enriching that data with semantic and business context: it’s this progression that enables AI to deliver – whether it’s accurate predictions, robust forecasting, or uncovering patterns that drive better decisions.
That said, not every AI use case is complex. For example, a Copilot assisting with ERP data entry doesn’t depend on a complex data platform. The type of AI solution determines how much infrastructure and data maturity are needed.
We make sure to get the foundation right from day one and start with an assessment and a clear strategy. We build only what’s needed, and we build it right. This ensures the ecosystem can scale when the time comes to support advanced AI workloads.
How does Fortude help organisations build data foundations for AI, and how do partnerships like Microsoft amplify that?
Gaurika: For most enterprises, it starts with a clear understanding of their current technology landscape and business priorities.
Without that clarity, AI initiatives quickly become fragmented – teams work off different data sources, there is no single source of truth, and security risks multiply. We are already seeing this play out across the market today.
At Fortude, we take a platform-agnostic approach, working across AWS, Databricks, and Snowflake. However, Microsoft plays a central role in many of our engagements. Its integrated ecosystem – spanning Azure, Microsoft Fabric, and Copilot, combined with strong enterprise-grade security and compliance capabilities, enables organisations to build and scale AI solutions within a well-governed, trusted environment.
This is particularly important as AI moves from experimentation to production. With Microsoft, customers are able to unify their data estate, apply consistent governance, and operationalise AI models at scale, without compromising security or control.
While the technology stack may vary, the principle remains constant: establish a secure, unified, and well-governed data foundation first.
That’s what allows AI to move beyond isolated use cases and deliver sustained, enterprise-wide impact.
When an agent acts autonomously inside an ERP, who is accountable, and how does Fortude approach responsible AI deployment?
Surangana: Not all agents operate the same way. Some are task-based, executing defined steps. Others work autonomously inside a workflow. In both cases, guardrails are non-negotiable. We work with customers to define what an agent can do, what it cannot, and when it must seek human approval.
Take order processing. An agent can validate submissions, check inventory, verify credit limits and flag bundling opportunities, all within configured boundaries.
Responsibility extends further. We honour each organisation’s access and authorisation model: a sales rep restricted to one region interacts with an agent that respects that boundary. Data privacy is built in from the start, whether GDPR, EU AI Act, ISO 27k and 42k compliance, regional data protection laws or financial encryption standards are applied.
For enterprises still navigating the gap between AI ambition and execution, Fortude’s message is clear: the technology is ready, but the foundation has to come first. As the company continues to scale its ecosystem globally, it isn’t just building AI solutions; it is building the conditions for AI to actually work.