The reason most businesses go looking for pricing information before they will speak to a provider is a reasonable one. Nobody wants to get on a discovery call, build some optimism about what AI could do for their business, and then find out the numbers make no sense for them. Wanting to know the cost upfront is not awkward, it is the sensible first move before committing time to a conversation.

The honest answer is that cost depends on what you are building, how complex your systems are, and how clean your data and processes are when you start. "It depends" is only useful if you know what it depends on, and most of the AI consultancy market gives you nothing to work with on that front. Pricing pages, where they exist at all, either quote ranges so wide they are meaningless or ask you to book a call before they will discuss numbers at all.

This article works through the three cost categories involved in any AI and automation project, gives you the real numbers for each one, and explains what moves the price up or down within each range. By the end of it you should have a clear enough picture to know whether working with a consultancy makes financial sense for your business before you speak to anyone.

What a build actually costs

Any AI and automation project involves three categories of cost: the build itself, the platforms that power the solution once it is live, and the ongoing maintenance. Understanding all three before you speak to anyone puts you in a position to evaluate what you are being quoted rather than simply accepting it.

On the build, the UK market spans a wide range depending on who you are working with and what you need. At the lower end, niche providers who have templated one or two specific solutions such as a chatbot or a voice agent can deliver builds from a few hundred pounds. At the upper end, enterprise consultancies like Deloitte or Accenture start their discovery processes alone in the tens of thousands, and their full builds are priced for organisations with hundreds of employees and complex technology estates. Most small and medium businesses sit well between those two points and are not well served by either.

For a business in that middle ground, build costs are driven by three factors:

  • How many systems need to connect, and how well documented those connections are
  • Whether the solution needs to make decisions and take actions based on variable inputs, or whether it simply moves data from one place to another
  • How clean and consistent the underlying processes and data are before work starts, as a solution built on unreliable data or an undocumented process costs more to build and delivers less once it is live

These three factors are what determine where a build sits in the range. A single clean workflow connecting two systems with well-documented logic is a fundamentally different piece of work to a multi-system build that requires the solution to reason about exceptions and integrate with several platforms simultaneously. Our own builds reflect that honestly:

  • Straightforward single-workflow builds start from £1,500
  • Complex multi-system builds with decision logic and multiple integrations can reach £10,000 or more
  • The factors above are what move the number, and we agree the price before anything starts

What platforms will cost you

Once a solution is built and deployed, the running costs sit entirely with you, not with the consultancy that built it. This is worth understanding before you commit to a build, because platform costs are ongoing and they belong to your business from day one.

They fall into three layers.

The first is the orchestration platform. This is the tool that connects your systems and runs the workflows once they are live. Which platform sits here depends on how the solution is built, and the cost implications are different depending on which route is taken.

For solutions built on dedicated workflow automation platforms such as n8n or Make, you pay a monthly subscription. n8n Cloud runs from £20 per month at the Starter tier through to £578 per month at the Business tier, with most SMB deployments starting on the Pro plan at £50 per month and moving upwards as build volume and complexity changes. Most medium-sized businesses begin with the Business plan at £578 per month and can scale to Enterprise plans and pricing if this fits their AI roadmap. Make has a comparable structure. The right tier depends on how many workflows are running and how frequently they execute.

For solutions built natively within the Google ecosystem, the orchestration cost depends on which tooling is used. Google AppScript is free to use, though execution costs and model calls sit outside that and are charged on consumption. More complex Google-native builds may involve tools such as Zenphi, which starts at around £100 per month and scales from there depending on the number of processes being automated.

For solutions built within the Microsoft ecosystem, Power Automate is included in most Microsoft 365 business plans for basic flows. The majority of AI-capable Microsoft applications, however, including Copilot Studio and AI Builder, require Azure credits or an Azure subscription on top of the existing Microsoft 365 licence. If your business is on Microsoft 365 and expects AI-integrated workflows, that Azure dependency is worth understanding before a build begins, as it adds a cost layer that is not always visible from the Microsoft licensing surface.

Consumption costs

The second layer is consumption costs, which are genuinely pay as you go with no fixed fee. To understand how these work it helps to know what a token is, because this is how AI models measure and charge for the text they process. A token is roughly equivalent to four characters of text, so a short paragraph is around 100 tokens and a full page of text is around 500. Every time your solution sends information to an AI model to process, classify, or generate a response, you are charged based on the number of tokens involved on both sides of that exchange: the text going in and the text coming back out. The more the solution runs, the more these costs are. Voice transcription works the same way, charged per minute of audio processed rather than per token.

On every build we select the most appropriate model for the job rather than defaulting to the most powerful and expensive option available. A simpler model that handles the task reliably costs a fraction of a more capable one, and most business automation tasks do not require the latter.

Storage

The third layer is storage, and it deserves more attention than most build conversations give it. If your solution generates or processes documents, recordings, logs, or any other data, that data has to live somewhere, and the cost of storing it grows over time as the solution runs. Most small businesses are well within the storage allowances already included in their Microsoft or Google plans at the point a solution is first deployed, but that position can change as the volume of processed data accumulates.

Two things are worth thinking about before a build starts rather than after. The first is what actually needs to be stored and for how long. Not everything a solution touches needs to be retained indefinitely, and defining a clear data retention policy at the outset keeps storage costs predictable rather than letting them drift upward unchecked. The second is how stored data is structured and labelled. AI systems that need to search or query stored data to answer questions or complete tasks will cost more to run if the data is poorly organised, because the model has to process larger volumes of text to find what it needs. Good structure and clear metadata reduce the number of tokens consumed in every retrieval operation, which directly reduces the ongoing consumption cost of the solution.

Across all three layers, total platform running costs for a deployed solution sit between £50 and £300 per month for most small businesses. That range depends on how many workflows are running, how frequently they run, and which platform environment the solution lives in. Before any build starts, we calculate a specific estimate for your deployment, not a ballpark figure but a number you can plan around.

Maintenance

The third cost category is maintenance, and it is the one most businesses underestimate when they first ask what AI and automation costs. Like owning a car, a solution needs ongoing attention: platforms release updates, the connections between systems change, AI models get replaced by newer versions, and as your business evolves the automations built around it need to evolve too. The question is who handles that work, when, and at what cost.

If you are on a Keep engagement, maintenance is included. We monitor your deployed solutions proactively, handle updates before they cause problems, and keep everything running without you needing to think about it. As the solution stack grows over time, the maintenance load grows with it, which is one of the reasons Keep is structured as a monthly service rather than a series of one-off jobs.

If you have commissioned a one-off Fix build, every solution comes with a three-month warranty as standard, meaning anything that breaks or needs adjusting in that period is handled at no additional cost. After the warranty period, support is available on request, and the level of ongoing attention a solution needs varies considerably depending on what it does. A single clean workflow connecting two well-maintained systems needs very little. A multi-system build with decision logic and AI reasoning will need more periodic attention as platforms update and model versions change, and before your warranty expires we will always give you an honest assessment of what your specific solution is likely to need going forwards.

What BIQ charges and why

There are three stages to working with us, each with its own fixed pricing structure. None of them involve hourly billing, and none of them involve a price you find out after the work has started.

The first stage is Find. If you do not yet know where AI and automation could genuinely add value to your business, a Find session is where that picture gets built. Over two hours we work with you and your team to map where the friction is and what it is costing you, then return a detailed diagnostic with sized and priced opportunities you can take forwards with us, consider independently, or compare with other providers. Find costs between £1,000 and £1,500.

The second stage is Fix. If you have a clear brief, whether it has come from a Find session or you already knew what you wanted before contacting us, Fix is the build. We confirm the scope and agree the price before anything starts. As covered in the Build section above, Fix prices vary depending on complexity: straightforward single-workflow builds start from £1,500, and complex multi-system builds with decision logic can reach £10,000 or more. Every Fix build includes a three-month warranty as standard.

The third stage is Keep. Keep is a monthly service that does two things: it continues building as your requirements develop, and it maintains and tunes everything already deployed under the same fee. Keep Foundation starts at £2,000 per month, which is the right starting point for most small businesses new to AI and automation. Keep Advanced at £3,000 per month reflects a larger deployed solution stack, and most clients move there naturally within the first twelve months as more builds complete and the value of the deployed stack grows.

Full pricing detail, including what is covered at each tier, is on the pricing page.

Knowing what AI consultancy costs is a more useful starting point than most of the market will give you. The numbers above are real, the ranges are honest, and the factors that move each number within those ranges are the same ones that will determine what a project costs for your specific business.

The most reliable way to get a number that reflects your situation rather than a market average is to start with a Find session. Two hours of structured diagnostic work produces a detailed picture of where AI and automation could genuinely add value, with sized and priced opportunities you can evaluate on their own terms. If the numbers make sense, you move forwards. If they do not, you leave with a clear picture of your own operations and no obligation to take it further.

Full pricing detail for Find, Fix, and Keep is on the pricing page.