Over the past year, we’ve all seen the hype around AI in the enterprise reach a new level. Every platform, from HR tools to ITSM suites, now claims to have “AI built in.” For many leading complex IT environments, that promise feels both exciting and exhausting.
As CEO of Voxxify, I spend a lot of time talking to CIOs and IT leaders who are navigating the reality of multi-vendor, multi-cloud ecosystems. And what I’m hearing is clear: adopting AI in a SIAM environment isn’t just a technology challenge – it’s a governance challenge.
That’s the conversation I’ll be bringing to Service North 2025, in our session on AI and Automation in SIAM. Because while AI offers huge potential to improve service integration and experience, there’s a growing issue that most enterprises haven’t fully recognised yet – what I call AI sprawl.
The Hidden AI Sprawl Behind Your SaaS Stack
Recently, I came across a public breakdown from a popular SaaS provider showing which foundation models they use across different features. It included 3-4 versions of GPT & Claude, all spread across Azure and AWS.
Now multiply that across your ecosystem.
The average large enterprise uses hundreds of SaaS applications, many now embedding AI in ways that aren’t always transparent. Even if only a quarter of those apps use generative AI, that could mean over 200 different models quietly influencing your data, your workflows, and your user experiences.
It’s the new “shadow IT” – only this time it’s “shadow AI.”
Why This Matters for SIAM
In a Service Integration and Management (SIAM) model, you’re already coordinating multiple suppliers to deliver one seamless experience to the business. Add in dozens of AI-powered services, each using different foundation models, and that coordination challenge grows exponentially.
A few things start to happen:
- You lose visibility into which models are touching which data.
- You face inconsistent behaviours across providers – one chatbot sounds human, another sounds robotic, and both are meant to represent your brand.
- You struggle with compliance – because each model, vendor, or cloud sub-processor may have different data retention or residency policies.
Without real oversight, you can’t be sure that your integrated service ecosystem is delivering a consistent, secure, and ethical experience.
What Enterprises Need to Do Next
There’s no single fix, but here’s what leading organisations are starting to do:
- Create an AI Bill of Materials (AI-BoM) – Just as we once tracked software dependencies, we now need visibility into which AI models, vendors, and infrastructures our services depend on.
- Embed AI governance into SIAM – Extend your supplier governance model to include AI behaviour, transparency, and accountability.
- Standardise AI-enabled integration – Require service providers to use agreed-upon APIs, data schemas, and ethics standards for AI features.
- Build literacy at every level – Your integrators, suppliers, and even business stakeholders need a shared understanding of what responsible AI looks like.
- Stay people-centric – AI should enhance human decision-making, not replace it. Focus on outcomes that improve employee experience and service quality.
The Real Opportunity
AI isn’t the enemy of SIAM – it’s the next evolution of it. When managed well, AI can automate integration points, surface insights faster, and personalise support in ways we couldn’t before.
But to get there, we need transparency, trust, and collaboration across the entire ecosystem. The future of service integration isn’t just multi-supplier – it’s multi-model. And success will depend on how well we align those models around human experience and business outcomes.
That’s what I’ll be exploring at Service North 2025. If you’re attending, let’s connect – I’d love to hear how your organisation is tackling AI adoption across your provider landscape.
See you in Manchester.
Steve Fleming, CEO, Voxxify
