Choose a solution

The right AI solution is one that fits your business context. This means looking beyond features to consider your:

• objectives
• budget
• data
• level of risk you are prepared to manage.

Many organisations begin by using existing tools. These may be free tools, paid subscriptions or AI features already built into the software. In most cases, building a custom solution comes later if needed.

This guide helps you understand the different ways to bring AI into your business and what to think about before you commit.

Before finding a solution, it’s important to define what you want to improve. This means understanding the task, process or workflow where AI could help and being specific about the outcome you’re aiming for.

Learn more about how to prepare your business.

Most businesses use AI in one of 4 main ways. Start with the first option and move to later options if you need more control or capability.

1. AI built into software

AI helps your existing software work a bit smarter.

AI built into software is best for everyday business tasks where setup needs to be quick and risk is low.

Many software products you already use may include AI features.

Examples include:

• accounting software that categorises transactions or flags unusual activity
• document systems that search, summarise or classify files
• customer management systems that help prioritise leads or analyse customer enquiries.

In this case, you aren’t buying AI separately. AI is one feature within a broader system that’s managed by the software provider.

2. Standalone AI tools

AI acts like a smart assistant you log into when you need it.

Standalone AI tools are best for testing AI, low‑risk tasks, or use by a small number of staff.

Some businesses use AI tools that operate outside their core systems. These are usually accessed through a web browser or an app with free or subscription-based options.

Examples include tools used to:

• draft emails, reports or marketing content
• research or summarise information
• analyse data.

These tools are quick to start using and often low cost. 

3. AI integrations

AI connects to the systems you already use.

AI integrations are best for businesses that want more capability without building a fully custom solution.

AI can be linked to your existing systems using integrations, connectors or APIs.

Examples include:

• adding an AI chatbot to a website to answer customer enquiries
• connecting AI to sales or inventory system
• using AI to analyse data pulled from multiple systems.

This approach sits between simple tools and fully custom solutions.

4. Customised AI solutions

AI is built around how your business works.

Customised AI solutions are best for complex, higher impact uses where control, data handling and accountability matter.

AI providers can design and build a customised solution for your business and specific need.

They are often used for:

• forecasting and planning
• supporting decision-making
• optimising complex processes, such as logistics or scheduling.

These solutions may use AI models from providers such as OpenAI or Anthropic, but they are tailored to your workflows, data and governance needs.

AI solutions need ongoing technical support to keep them working properly over time. These services aren’t AI tools themselves, but support how AI is deployed, monitored and maintained.

They are often referred to as machine learning operations (MLOps). MLOps help keep AI systems reliable, accurate and safe once they are live.

It covers tasks such as:

• deploying AI into business systems
• monitoring performance and accuracy
• managing updates as data or needs change
• fixing issues when AI behaves unexpectedly.

MLOps is most relevant for customised or customer facing AI solutions.

Once you have a clear task or process where AI could help, the next step is deciding how to implement AI.

There is no single right solution. The best option depends on:

• how you use AI
• how important the task or process is to your business
• what data is involved
• how much control and oversight you need.

The aim is to choose the simplest option that meets your needs, and to add complexity only where the value justifies it.

Start with the impact of the task or process

A helpful way to choose the right solution is to think about the impact if something goes wrong.

As potential impact increases, so does the need for oversight, accountability and ongoing management, as outlined in our essential AI practices.

Learn more about how to assess business impact and identify opportunities.

Check what you already have

Before exploring new tools or suppliers, review the systems you already use. If they come with AI features, this can reduce cost, complexity and risk.

Balance effort, cost and control

Different AI solutions involve different trade-offs. When comparing options, consider:

• how much setup and configuration is required
• whether staff will need training or ongoing support
• how much control you need over how the AI works
• whether the solution can adapt as your business changes.

In general, more control comes with higher cost and effort. For many tasks, a simple option is enough.

Plan for technical support

Some AI solutions need little ongoing support. Others need active monitoring and maintenance.

As part of your decision, consider:

• how issues are reported and resolved
• whether the solution relies on additional services
• how you would pause, fix or switch off the system if needed.

Tasks with higher potential impact need clearer plans for support, responsibility and ongoing oversight.

Once you are clear about the type of AI solution you need, the next step is to look at what options are available.

A market scan helps build a basic understanding of:

• the types of AI solutions available for your task or process
• the range of suppliers in the market
• typical costs, capabilities and levels of support.

This doesn’t need to be a detailed or time-consuming exercise. The aim is to narrow down your options before engaging with suppliers.

Once you have narrowed down the type of AI solution that fits your needs, there are some practical checks to make before you commit.

These checks apply whether you are:

• subscribing to a standalone AI tool
• buying AI as part of a software product
• building a more customised solution.

Checking these early can help reduce cost, avoid surprises and manage risk.

When AI is added to tools you already use, through integration or updates, these checks become even more critical. Teams may miss the need to run these checks if AI features are implemented quietly within familiar software.

Cost and actual use

AI pricing can increase the more you use it.

Think about:

• how often the AI will be used
• how many people need access
• whether pricing changes based on use, data volume or features.

Starting small can help you test value before scaling.

Terms of service and data use

AI solutions often rely on access to business or customer data.

Make sure you understand: 

• what data the AI will collect or use
• where that data is stored and processed
• whether the data is used to train AI models
• who owns the outputs created by the system. 

Be cautious when using sensitive or confidential information, especially in platform-based or browser-based tools.

Integration with your existing systems

AI delivers the most value when it fits into your existing workflows.

Check whether the AI solution:

• works with your current systems
• needs APIs or connectors
• introduces extra manual steps or duplication.

Poor fit can increase effort rather than reduce it.

Vendor lock-in and flexibility

As AI becomes more embedded in your work, switching providers can become harder.

Consider:

• how easy it would be to export your data
• whether workflows can move to another platform
• how you would respond if pricing or service terms change.

Flexibility matters, especially if your business is growing or changing.

Vendor support and reliability

Not all AI solutions have the same level of support.

Understand: 

• what support is available or included
• how to report issues and expected response times
• how the vendor communicates updates, changes or outages.

Reliable support reduces risk, especially when AI is used in important business tasks.

Risk and responsibility

No matter which solution you choose, your business remains responsible for how AI is used.

Think about:

• What data the AI will use and where it is stored. 
• Whether your data is used to train AI models
• what could go wrong and how you would detect it
• who is responsible for responding to and fixing problems.
• how accountability is shared between your business and any suppliers.

When you’re ready to take the next step, use our guide to support your discussions with AI suppliers.