Ashish Surti, Executive Vice President & Chief Digital & Information Officer at Colt Technology Services, on navigating post-acquisition complexity, reimagining enterprise architecture, and AI with purpose.
Colt Technology Services, a global digital infrastructure provider, is no stranger to complex connectivity. Its recent acquisition of Lumen EMEA—adding 3,800 enterprise customers and an extensive fibre asset network—highlights more than just increased bandwidth. It signifies a strategic shift towards a deeper digital transformation.
At the helm of this shift is Ashish Surti, Executive Vice President and Chief Digital & Information Officer (CDIO), leading the integration of a sprawling IT estate, overseeing large-scale cloud migration, and exploring the real business value of generative and agentic AI.
Colt, a leading pan-European fibre provider, now extends its presence across key global markets: Tokyo, Singapore, Hong Kong, Australia, and North America. The company says that it is in “Chapter Two ” of its journey, which is shaped by the Lumen EMEA acquisition, which significantly strengthens Colt’s European footprint and customer base.
In an exclusive conversation with Jatinder Singh, Executive Editor, CIO&Leader and R. Giridhar, Group Editor, 9.9 Group, Surti reflects on the complexities and payoffs of enterprise-scale transformation, touching on everything from data migration and AI governance to architecture alignment and customer trust. Here are excerpts from the interview.
CIO&Leader: You recently completed a significant acquisition of a fibre asset company. What are your biggest challenges from a technology or information systems perspective?
Ashish Surti: This acquisition of Lumen marks Colt’s first major venture of this scale, and the need to scale our IT platforms by 60% almost overnight is a substantial shift for us. Our typical focus in the telecom industry is optimizing existing assets, networks, or IT infrastructure, so this marks a pivotal transition.
We have strategically decided to reimagine our entire technology platform to meet this challenge. We aim to build a future-ready architecture that aligns more closely with our evolving business needs and scales efficiently. We are leveraging cloud technologies to modernize our core inventory platform, which is critical for managing opportunities, bookings, and service delivery.
We did not acquire any of Lumen’s technology systems. Instead, we focus on migrating the data of approximately 3,800 customers from Lumen’s systems into our own. This requires significant upgrades across our platforms to ensure a seamless transition.
Our long-term goal is to operate the newly acquired business on a single, unified technology platform. Many organizations struggle with technical debt following acquisitions due to the complexity of integrating disparate systems. We are taking a proactive approach by consolidating operations early on, which will simplify our processes and help us avoid common post-acquisition pitfalls.
CIO&Leader: The challenge, of course, is that this is an operating business. How do you manage data migration and AI integration and ensure consistency without disruption?
Ashish Surti: Data migration is inherently complex, particularly when moving between platforms—even from the same vendor—due to differing data structures and attributes.
In our case, the business we acquired was formed through five separate acquisitions and had never fully consolidated its technology landscape.
As a result, we’re dealing with five different inventory systems, multiple pricing models, and various monitoring and ordering platforms. Our task is to extract, transform, and integrate data for 3,800 customers into Colt’s systems.
One of the biggest challenges is data quality. At Colt, customer data flows across several systems—from delivery to assurance to billing. To ensure accuracy, we must consolidate and reimagine how customer records and service links are structured so that everything aligns with our internal frameworks. This often involves rekeying and redefining customer relationships to ensure seamless service and billing continuity.
We are dealing with several terabytes of data and only a 15% is relevant to be migrated. Some data adds no value, and some may not exist in our current architecture. In such cases, we may need to generate or reformat data to make it usable synthetically.
We are approaching the migration on a product-by-product, customer-by-customer basis, which adds further complexity. To manage this effectively, we have implemented structured processes and technologies that enable real-time data transformation upon acquisition.
Throughout this journey, our priority is minimizing customer impact. We focus on clear, proactive communication to help customers understand what to expect and feel supported during the transition. At Colt, our customer-centric approach is a core differentiator, and we are fully committed to preserving that experience during this migration.
CIO&Leader: What is your approach to the 85% of data that isn’t immediately operationally relevant?
Ashish Surti: We are taking a deliberate approach. That data won’t be discarded, it’ll be archived. Initially, we considered building overlay tools to make it accessible on demand. Still, we’re exploring how generative AI can enable natural language queries on archived data, eliminating the need for rigid dashboards or manual retrieval processes.
This is one of GenAI’s most compelling use cases for us. It allows business users to ask plain-language questions about historical data and get actionable insights without relying on IT. If we later find a broader use case, we’ll selectively reintroduce that data through ETL processes. It’s about reducing operational burden while preserving long-term value.
CIO&Leader: How are you handling unstructured data such as contracts, meeting notes, and emails?
Ashish Surti: While the majority of our workflows are structured, spanning delivery, assurance, and billing, there’s a layer of unstructured content we can’t ignore: contracts, change requests, meeting summaries, and so on.
We are actively exploring how to integrate this into our systems. GenAI also has real promise here, helping us summarise notes, extract metadata, and connect this information to operational records. It’s part of a broader strategy to bring structured and unstructured data into a usable, searchable format.
CIO&Leader: Is all this data ending up in a centralised data lake?
Ashish Surti: Not by default. We are using advanced technologies to enhance our data management and operational efficiency, but we’re cautious about scale. There is a cost to storing and maintaining vast volumes of data, so we’re taking a relevance-first approach.
We’ll assess which parts of the archived data justify inclusion within six months post-migration. If a clear use case arises, we’ll ingest selectively. If not, we won’t expand the lake unnecessarily.
CIO&Leader: Could you share how you leverage AI and automation to enhance your service delivery, especially in light of your past collaboration with Infosys and earlier data transformation initiatives?
Ashish Surti: During our acquisition process, we identified the need for a significant technology transformation. We are replacing some of our legacy inventory systems and migrating several components to the cloud, which requires us to manage a diverse set of platforms tailored to different functions.
We quickly realized that our internal IT team, though highly capable, would not be able to scale at the pace required to meet these new demands. That’s where our partnership with Infosys became instrumental. Over the past 18 months, they have worked closely with us to support our scaling efforts.
For instance, we needed 40 new products to our technology portfolio to enable quoting, order processing, and delivery of new services. Infosys has helped us build these capabilities while supporting our SAP HANA program, focusing on technology configurations and system integration.
About three years ago, we partnered with EY to reimagine our technology landscape. That strategic roadmap is now being implemented with Infosys’s help, turning vision into execution.
Infosys is critical in our data and customer migration programs this year. Their support in AI and automation is helping us streamline operations, improve scalability, and enhance our overall service delivery.
CIO&Leader: How do you think agentic AI will enhance the software development life cycle, and how are you supporting that?
Ashish Surti: Agentic AI has the potential to significantly enhance the software development life cycle by making processes more intelligent, efficient, and scalable. For instance, AI can assist in capturing business requirements, translating them into technical documentation, and generating test cases and automation scripts. This accelerates the pace of change and enables IT teams to deliver value more rapidly.
We’re beginning to see tangible outcomes from AI initiatives, whether through generative AI or other advanced models. The promise of agentic AI is especially compelling, as it allows us to build systems that can scale organizational capabilities beyond what was previously possible.
Currently, we are actively exploring multiple pathways to implement agentic AI. This includes evaluating both targeted applications and broader horizontal solutions. We’re also considering strategic collaborations with major technology providers like Google to shape our roadmap.
In parallel, we are piloting several internal use cases, particularly in IT management and monitoring, to assess the practical impact of these technologies. These initiatives are already revealing strong potential for real business value.
That said, it’s crucial to embed robust risk management and governance frameworks from the outset. We must ensure lifecycle oversight of the AI agents we deploy, continuously monitoring their behavior, effectiveness, and organizational impact.
CIO&Leader: What were your key considerations as you transitioned to a new enterprise architecture? And how will the new architecture enable your AI initiatives?
Ashish Surti: Those are important questions. When I stepped into my role about three years ago, I found that we had accumulated several IT solutions, each implemented to solve specific business problems. Over time, this led to a fragmented landscape of point solutions, some integrated and others siloed.
One of our primary objectives was to improve the employee experience, as a better employee experience ultimately translates to a better customer experience. To do this, we had to reduce the number of disparate systems employees interact with during core business processes. We needed to modernize our foundational platforms to establish a reliable infrastructure, streamline workflows, and eliminate duplication and redundancy.
We are now focused on consolidating our applications into integrated suites, particularly within our CRM platform. This will enable teams to operate within a unified environment for marketing, sales, opportunity management, and reporting. This shift will enable us to create a “single pane of glass” for users and unlock richer insights across functions.
A key enabler of this transformation is data integrity. For AI to deliver meaningful and reliable insights, it must be grounded in consistent, trustworthy data. Without that, there’s a real risk of inaccurate or misleading outputs, which many call “AI hallucinations.” We want to avoid taking AI-generated responses at face value without validation.
We are rationalizing our systems and datasets to support our AI goals and we are ensuring that our data lake management practices align with our internal frameworks to maintain accuracy and security. This will serve as our primary source of truth and ensure the data feeding into our AI systems is accurate, secure, and dependable.