Artificial Intelligence (AI) has introduced a new category of business technology that is shaking up traditional ways of working. It also means that many organisations are struggling with a fundamental question: where should they invest their time, resources, and budget in the age of AI?
With new AI tools launching almost weekly, businesses can quickly become overwhelmed by the number of options available. Should they invest in an AI assistant, a content platform, an automation tool, a chatbot, or a custom AI application?
The answer is rarely to choose a single tool. The biggest opportunities today come from building connected AI systems that integrate with existing business processes, data, and technology infrastructure. This includes areas such as customer insights, marketing operations, reporting, and internal workflows where disconnected systems can often create inefficiencies.
In addition, businesses achieving the strongest results are not simply using AI more often. They are the ones designing intelligent workflows where AI automates repetitive tasks, improves decision-making, and supports employees in meaningful ways.
The future of AI is not about having more tools. It’s about creating smarter systems that allow different technologies to work effectively together.
The future of AI is not about having more tools; it’s about building better systems
Many businesses currently approach AI as a collection of standalone applications.
They might use one platform for content generation, another for automation, and another for analytics. While these tools offer individual value, the greatest opportunities arise from connecting them. In this way, the future of AI is about creating intelligent systems where different technologies connect and work seamlessly.
AI becomes significantly more powerful when it’s integrated into the wider business ecosystem. This allows systems to exchange information, trigger actions, and automate complex workflows.
For example, a business could create a workflow where customer data is collected from a CRM, AI analyses customer behaviour, and a workflow automation platform triggers action. Insights are then delivered directly into internal systems. Supporting more personalised customer journeys, campaign optimisation, and faster access to audience insights.
Instead of employees manually transferring information between platforms, AI can orchestrate the entire process. This shifts AI from a productivity tool to an integral part of a business’s operational infrastructure. And it represents a fundamental shift: AI is moving from being a productivity assistant to becoming part of a business’s overall operational infrastructure.
The foundations of effective AI systems
When selecting AI technology, businesses should focus on capabilities that enable them to build flexible and scalable solutions. The most valuable AI systems typically combine three key components:
-
Workflow orchestration: connecting AI to business operations
Workflow orchestration is a significantly underrated technology. Platforms such as n8n enable businesses to create customised automation pipelines that connect applications, APIs, databases, and AI models.
This approach offers greater flexibility. Rather than forcing businesses to adapt their processes around a tool, workflow orchestration enables organisations to build automation around the way they already operate. For example, a company could automate processes by pulling data from multiple platforms. They could use AI to analyse information, appl business logic, produce structured outputs, and update existing systems automatically. For marketing teams, this could include automating campaign reporting, consolidating performance data, or identifying customer trends without relying on manual analysis across multiple platforms.
The value here is not simply automating one task; it’s about creating repeatable systems that reduce operational friction across the entire organisation.
-
AI APIs: embedding intelligence into existing systems
Large Language Models (LLMs) such as OpenAI and Anthropic provide the intelligence layer behind many modern AI applications. However, the greatest business opportunities don’t stem from using these models through a simple chat interface.
The real value of LLMs comes from integrating AI directly into existing business systems. This includes everything from websites to customer relationship platforms, reporting dashboards, and internal applications.
Through APIs, businesses can create AI-powered features that are tailored to their specific requirements. For example, instead of requiring an employee to analyse customer information manually, an AI workflow could automatically process data, identify patterns, and provide structured recommendations. Helping teams make faster, more informed decisions across areas such as customer engagement, content performance, and business strategy.
Overall, this approach allows businesses to move from asking AI questions to building sophisticated systems where AI actively supports business operations and workflows.
-
Turning AI into a business expert with RAG
One limitation of general AI models is that they don’t automatically understand a company’s internal knowledge, processes, or historical information. Retrieval-Augmented Generation (RAG) helps solve this challenge by connecting AI models with relevant business data.
Rather than relying on generic information, RAG allows AI systems to retrieve and use company-specific knowledge when generating responses. As a result, AI systems can access internal documentation, product information, company processes, customer history, project knowledge, and historical data.
Vector databases such as Pinecone, Redis, and pgvector support this approach by allowing information to be stored and retrieved in a way AI models can understand. The result is an AI system that becomes a context-aware internal expert rather than a generic assistant. This benefits the organisation in the long-term by unlocking more in-depth knowledge, data, and value that’s specific to your business model and needs.
Underrated AI opportunities: open-source and private AI systems
Many businesses today are overlooking the potential of open-source AI infrastructure. Often, organisations send sensitive information to third-party AI platforms without considering more private alternatives. For organisations managing confidential information such as client, customer, and operational data – as well as privacy and control mechanisms – these data considerations are key when designing new AI systems.
By combining workflow tools such as n8n with open-weight models through platforms like Ollama, businesses can create secure, private AI agents. This approach provides greater control over sensitive information, reduces dependency on external vendors, increases flexibility, and lowers long-term costs. For organisations handling confidential client data, this is a particularly important consideration.
Choosing the right AI model for each task
The future of AI adoption will be about finding a balance between enhancing capability and building systems businesses can trust. A common misconception is that businesses need to select one AI model and use it for everything. Different models have different strengths.
The most effective AI systems going forward will be those that combine multiple models, depending on the requirements of each workflow. For example, Claude performs strongly in areas such as coding and complex reasoning, while OpenAI’s API can provide a cost-effective solution for high-volume, routine tasks.
Businesses can design workflows where simpler tasks are handled efficiently, while more complex processes are directed towards models with stronger reasoning capabilities. The future will therefore be about intelligently combining different technologies to solve specific business challenges.
Three AI foundations I would recommend for businesses starting out
For businesses with limited budgets, investing in multiple expensive subscriptions is not recommended. Instead, focus on building strong foundations. Here are some tools I recommend:
-
n8n for workflow automation
n8n offers flexibility for connecting systems, automating tasks, and integrating AI models without the high costs that are often associated with enterprise automation platforms. Businesses can use the community version, self-hosted infrastructure, or cloud options depending on their specific requirements.
-
OpenAI or Anthropic APIs
Rather than paying for multiple AI subscriptions, businesses can access powerful AI capabilities through APIs and integrate them directly into areas where they create greater value. This approach allows organisations to pay based on usage and incorporate AI into existing workflows.
-
A reliable CRM with native webhooks
Without structured and accessible data, even the most advanced AI systems will struggle to deliver meaningful results. Legacy systems and fragmented data remain major barriers to scaling AI, with organisations increasingly needing modernised infrastructure before advanced AI workflows can deliver value. AI automation depends on having clean, accessible data. A reliable CRM provides the foundation for AI systems to access information, process it, and update business processes. Optimising AI outputs: better structure creates better results
Many businesses assume that increased AI usage will lead to better results. However, the quality of AI outputs depends largely on system design. The goal should not be simply generating more AI content. It should be creating reliable, repeatable outputs that can be used immediately within business processes.
One key area is moving away from unrestricted functions. Instead, businesses should focus on strong system prompts, clear instructions, structured outputs, and validation processes. For example, rather than asking AI to create responses that require manual editing, businesses can design workflows where AI produces structured JSON outputs that automatically feed into other systems.
The AI capability businesses are underutilising: agentic workflows
Agentic tool use and function calling are among the most promising areas of AI development. Currently, 40% of enterprise applications are expected to include task-specific AI agents by the end of 2026, up from less than 5% in 2025. This highlights the shift from AI assistants towards systems that can execute business workflows.
However, modern AI systems are capable of much more. Agentic workflows allow AI models to decide when to use external tools, access information, and complete multi-step processes based on user intent.
AI models can now:
- Decide when to trigger external APIs
- Search databases
- Retrieve information
- Execute multi-step workflows
- Complete actions based on user intent
These new features represent a major shift from AI simply providing information to AI helping businesses execute processes in more streamlined ways.
Measuring whether AI is actually creating value
One of the most important questions businesses should ask is: is AI actually improving the way we work?
The answer must be measured through operational outcomes. Research shows that 95% of enterprise generative AI pilots are reported to deliver no measurable business impact. This finding reinforces that successful AI adoption depends on integrating technology into workflows rather than simply deploying new tools. See the questions below to help guide you in your business.
Time-to-completion
Has a workflow become significantly faster? For example, has a reporting process that previously took days been reduced to hours or minutes?
Error reduction
Has AI improved accuracy? Or are employees spending just as much time correcting AI outputs as they would have spent completing the task manually?
Cost per successful execution
Does the value generated outweigh the cost of AI subscriptions, API usage, infrastructure, and maintenance?
Remember that AI should be implemented not for its novelty, but because it creates measurable business improvement. It needs to provide measurable improvements for the investment to be worthwhile and sustainable in the long-term.
How we are applying AI at Manning & Co.
At Manning & Co., we have explored practical applications of AI to improve our operations and create more efficient internal workflows. Our approach has not been to adopt AI simply because it’s new and innovative. Instead, we have identified repetitive processes and explored where AI can help our teams work more effectively and productively.
Two examples of this refined approach are our automated reporting workflows and our internal SEO content-generation dashboard.
| AI application | How we use AI | Business impact |
| Automated annual reporting workflows | By integrating n8n with AI APIs, we can pull data from platforms such as Google Analytics 4 and Google Search Console, structure and analyse insights, convert outputs into JSON, and prepare information for dashboard visualisation. | Removes significant manual engineering effort and operational overhead, reducing time spent collecting and formatting data, while allowing teams to focus more on analysis and strategic insights. |
| AI-powered SEO content-generation dashboard | By connecting structured keyword data through the DataForSEO API with carefully designed AI workflows, we can automate elements of the content research and planning process. The system supports keyword research, search intent analysis, content structures, heading recommendations, FAQs, and initial content generation. | Reduces repetitive research and planning tasks, helping specialists spend more time refining strategy, applying creativity, and ensuring content aligns with audience needs and search intent. |
These examples above demonstrate where AI creates the most value for our business: not by replacing expertise, but by removing repetitive work and enabling our Manning & Co. teams to focus on higher-value activities.
Final thoughts
The future of AI will not be defined by which company creates the biggest model or which tool becomes the most popular. It will be defined by how effectively businesses integrate AI into the way they operate day-to-day.
Organisations that succeed in the future will be those moving beyond basic experimentation with AI tools. Forward-thinking organisations building intelligent systems with AI now are well-placed to future-proof their operations. So, going forward, use AI to design smarter workflows, connect business systems, and create technology that helps people work more effectively. This will help you get ahead in the modern world of marketing and AI.
