90% of your staff uses AI, just not the one you paid for
Discover why enterprise AI adoption falls short from a business strategy perspective and how flexible AI solutions deliver greater ROI and security.
Don’t scroll past just yet. I promise this is a three-minute read, max. If you get to the end and disagree, hey, at least you gained a different perspective and there’s always value in that.
Listen, on paper, your company’s AI strategy is ticking all the boxes. You bought the enterprise licenses and the data looks great. You’re not alone in this mindset. The MIT NANDA State of AI in Business Report shows that about 40% of organisations have bought enterprise LLM subscriptions. AI Strategy Transformation achieved, right? Not quite.
The exact same report reveals that over 90% of employees are regularly using personal AI tools for work. Most of them aren’t using the rigid corporate platform you paid for. Instead, they’re quietly turning to personal accounts and unvetted tools to get things done. It is not happening because your employees are trying to be rebellious, it is happening because they are trying to be efficient. Welcome to the Shadow AI economy.
The "one size fits all"
The biggest mistake we see right now is prescribing a blanket AI tool across an entire organisation without considering how different teams actually work. An HR manager, a backend developer, and a creative designer for instance have completely different roles, responsibilities, and tasks. Yet, leadership teams frequently hand them the exact same corporate LLM interface and expect it to help them all the same. If an AI tool isn’t malleable enough to adapt to a specific workflow, it’s not the solution. Rigid enterprise setups rarely fit use cases in practice. When you force a rigid tool onto a team with highly specialised needs, adoption plummets, and people quickly move beyond your tech stack to find something that works.
A reality check in practice
At Gen25, we recently worked with a large enterprise client that successfully deployed a massive enterprise tool. However, their design and creative teams needed something more. The design team needed ultra-high-quality image production. The corporate-mandated tool was not meeting that need. Instead of forcing it, we sat down with them, listened, and sourced two to three specialised AI imagery tools that formed a "trifecta" for their workflow. We didn't compromise on security, we ran those tools through the exact same rigorous enterprise vetting as the main platform. By accepting that the corporate tool could not do it all, we solved the problem, stopped wasting budget, and got the teams what they needed.
How we overcome this: the power of discovery
There is no one-size-fits-all solution for AI. Every business is unique, and every department within that business is unique. What works for HR will not answer the same needs of a design team. Here is how we get out of the adoption trap:
- Lead with discovery: Before mandating a tool, we do the groundwork to understand the highly differentiated needs across the business. Learn from the shadow AI usage.
- Deploy malleability: We focus on deploying tools that allow for deep customisation and flexibility. Platforms like DevRev succeed because they let teams build their own targeted agents and bend the tool to the workflow, not the other way around.
- Vet the alternatives: If a tool isn't meeting a team's needs, don't force it. Help them find the right tool, put it through the proper enterprise security checks, and make it official.
At the end of the day, a license does not equal adoption. If you want real ROI, stop looking at the technology and start looking at the people’s needs. Your people are already there with their personal AI tool, can you catch up? Give your people tools that adapt to them. So, how is your organisation balancing enterprise security with the actual workflow needs of different teams? Let’s talk about how we can build flexible, secure AI ecosystems.
