With the sheer volume of tools, models and hype surrounding AI these days, it can be confusing, maddening, frustrating or (D) all the above for IT service providers (ITSPs) to even know what to do or where or how to do it.
But make no doubt, ITSPs can start applying AI today, not just as a buzzword but as a transformative force for operational efficiency, smarter data strategies and real ROI, says Wasim Ali Khan, chief digital officer and general manager of the Digital Transformation & AI Office at FUJIFILM Business Innovation Asia Pacific.
From everyday use cases like intelligent document processing to strategic advice on building successful AI practices, Khan believes there is a grounded roadmap for ITSPs ready to move from curiosity to capability. He’ll speak about the subject 16 September at the GTIA ASEAN Community meeting in Jakarta.
We asked Khan to share his thoughts on AI opportunities and challenges for ITSPs, and how to sort through all the hype. Here’s what he had to say.
First, a note from Khan:
So, when I was asked to respond to this Q&A, given the core topic was all AI related, one thought that crossed my mind was do I just plug such questions into a LLM model ChatGPT tool, but that would still miss out on the true customized personal response. As such, these are not auto generated responses and from my own tongue, [Wry smile].
How can ITSPs start applying AI today?
If the perspective of starting to apply AI today is from the ITSPs being key AI service provider to enterprises, the first thing to consider is that the whole enterprise AI space is vast. Therefore, selecting the right positioning and targeting the key offering categories to focus on becomes vital.
That positioning selection is often derived by criteria such as level of manual process steps vs. automation: Can AI increase efficiency, level of process/ task complexity, ability to demonstrate ROI and value, and does using AI transform a clients business process? The AI value topic is a key theme.
What skills/resources do they need to get started?
The consistent approach for all enterprise AI projects irrespective of the industry or process improvement category, is to ensure a strong solid reliable data management and data connections strategy. The data management part is fairly obvious given quality data will still provide greater value. But more so with LLM-based AI offerings, we find different data integrations or data augmentations occurring.
Therefore, having skills in using the right tools for data aggregation and integration are key skills that are the logical starting point for many enterprise projects. Having strong relationships with group data teams becomes ever more critical, as is mapping out the various data input ecosystems for starting projects.
What is a customer business problem that MSPs can solve now using AI (and how)?
Speaking from my own experience, I’ve seen a lot of change in the OCR, capture category. There was an evolution to intelligent document processing (IDP) and now more LLM OCR-driven offerings. If ITSPs can use that as a logical starting point to support clients, this will often bear fruit, as the AI newer tools, allow better data understanding, the data context and defined metadata labelling, but by applying LLM tools, it allows for more additional business process capabilities such as data validation, data risk management as well as generative AI summaries.
The core factor here is this OCR problem often occurs across industry, processes and tasks. As such this is a good, everyday problem clients often still have. No doubt this will change as more digital native content is tagged at source, but you’d be surprised how much data input is still manual!
How can ITSPs sort all the AI tools/hype to know what’s best for their organization?
The best way to tackle the AI hype question goes back to the age-old concept of value. I.e. does using the latest generative AI tools consume so many tokens that the cost ROI becomes negative. Does using AI enhance and transform my business process, or can simpler process engineering help instead? Or by applying newer AI tools, does it realize tremendous time, efficiency and costs savings that the payback is clearly obvious. The best way then is to understand those current state processes and understand where the value and ROI using AI is clear.
A good example I see is the process of tender evaluations undertaken by large procurements departments. Typically, a human will analyse a tender response, then assess this based on different weighting criteria, must review supporting information, different data sets could be text heavy for risk or numerical financial information. But with the latest agent and generative AI tools, that data can be analysed in seconds/ minutes and a report fired out with the key outcomes, which can be fine-tuned. Such examples show that its reality and not hype of AI!
What’s the biggest challenge ITSPs face around AI today?
There are many challenges that MSPs and ITSPs face around AI. There is an abundance of choice with different models. Do you focus on a particular stack or model and get trained on those. Also, the ability to stay updated on all the new technologies and model developments by the various hyper scalers is quite a task. More and more projects now rely on a more agnostic approach to AI model selection, as large enterprises may have subscriptions to Gemini, DeepSeek, Claude and an MSP/ ITSP will need to adapt accordingly.
Also, concepts such a vibe coding where code is generated automatically simply from natural language prompts is meaning different skills sets and roles are taking precedent. Whereas, in the past developers would be more critical, now the data scientist role may require more focus. So, understanding which roles to focus and invest on becomes ever more important.
A difficult question to answer as there are so many perspectives it can be addressed from.
What’s your best advice for ITSPs about developing a successful AI solution or practice?
This requires a multi-dimensional response. ITSPs need to have the right tech AI offering focus but also be able to augment that with the right softer client relationship management skills, which themselves need to be extensible.
Aspects such as strong stakeholder engagement not just with IT and the business sectors but also addressing responsible AI/governance are vital starting points. Then building out strong offerings with good AI use case examples is also vital. Quantifying the results and success metrics is also key and calculating saving with credible baselines is again important.
This all may sound standard to most services deals undertaken by ITSPs. The key differentiating factor now is, that most of the above capabilities are often embedded with generative AI platforms, that can seamlessly provide the data. Or using agentic AI to build fast workflows to quickly provide client solutions. The AI tools available are the key difference to yesteryear and that’s where the true transformation lies. This means for hungry MSPs by applying such tools to inefficient process should yield strong value! Value, there’s that key word again.
Follow us on LinkedIn! #WeAreGTIA
