Are Businesses Paying For The Same AI Features Multiple Times?

After looking at how businesses have been testing AI and slowly moving into the next phase after the “testing” era of the technology, a different issue has since come to mind.

A typical AI stack now costs $1,236 per employee every year, according to Lorka AI. Writing tools, design software, research platforms and meeting assistants can build into a four figure annual bill very quickly.

How Much Are Businesses Spending?

Lorka AI found that many AI products moved from free or low cost access in 2023 to monthly subscriptions that often cost between $20 and $30. Premium tiers can cost as much as $200 per month.

The company found annual costs of $384 for writing tools, $552 for creative software and $300 for research and organisation products and those subscriptions often cover tasks that overlap.

Businesses spent the last few years signing up for AI products. Many are now reviewing what employees actually use and what software earns a place in the budget.

The issue is not always the price of a single subscription. The issue is paying multiple subscriptions for work that can be done through one platform.

Are Businesses Paying Three Times For One Capability?

Aigars Pilmanis, Founder at VolRadar, reviewed his own software stack after costs started growing, saying, “I noticed that the collection of tools for VolRadar increased in size and cost because of redundant AI subscriptions – this happened before I conducted an audit of the system.”

Many businesses purchase AI products department by department. Marketing teams choose one platform and operations teams choose another, for example…

“It is common for three different tools to include a general model in their packages. Because of this a company pays three times for a single capability.

“Many AI features are simple interfaces for a model that is already available through another service. It is less important to choose a specific model than it is to avoid paying for the same logic multiple times,” Pilmanis said.

A business can easily spend money on three subscriptions that produce nearly identical results.

Which AI Tools Deserve A Place In The Budget?

Pilmanis uses a simple method to review his software spending. He says, “For this task I list every AI tool and describe the specific function that only that tool performs. If a tool has no unique function, I cancel the subscription.”

That method forces businesses to justify every subscription. If a product performs the same work as another tool, there is little reason to keep both.

Lauren Murphy, CEO of Friday Initiatives, believes businesses should understand everything already operating within their organisation before purchasing new products.

Murphy says, “Before choosing another AI model, businesses should first understand their own environment. The challenge is rarely a lack of AI capability; it’s a lack of visibility into where AI is already being used, what data it can access, who owns it, and whether it genuinely delivers unique value.

“The most effective organisations start by mapping their AI estate. That means understanding which tools are in use, what data flows between them, where there is duplication, and which models are actually solving distinct business problems.”

A review of licences, subscriptions and workflows often identifies software that serves the same purpose.

“Once you have that visibility, it becomes much easier to consolidate vendors, reduce overlapping licences and focus investment on the tools that create measurable outcomes,” Murphy added.

What Does A Lean AI Stack Look Like?

Businesses often assume that more AI subscriptions produce better results but cost reviews are say something else, as Pilmanis explains, “The most affordable configuration is typically one powerful general model and two tools that are truly specialised.

“There is no need for ten interfaces that all send requests to the same API.”

Paul Clapp, co founder of Priority Pixels, has faced the same issue through web development, SEO, paid media and content production work. He says, “The problem is that most AI tools now do most things. ChatGPT writes copy. So does Gemini. So does Claude. Businesses end up paying for three of them because someone in marketing chose one, someone in ops chose another and nobody ever had the conversation about how they use it and what they actually needed.”

Clapp prefers selecting one primary platform and building company knowledge around that product.

“The way we approach it: pick one model for your core work and get good at it. The cost differences between frontier models are small. The switching costs between them are not.”

“The question worth asking is not which AI tool is best. It is which one your team will actually use consistently. Stack too many and you get none of them used well.”

Murphy believes spending discipline will become more important as AI budgets grow, saying, “The right AI model isn’t necessarily the most advanced or expensive one. It’s the one that fits your data, governance requirements and business objectives. In many cases, organisations can achieve the same result with fewer tools, lower costs and less risk simply by understanding their existing landscape before adding new technology.

“As AI adoption matures, the winners won’t be the businesses with the most AI tools. They’ll be the ones with the clearest view of how those tools connect to their data, processes and decision-making.”