Working with SMEs Through Cranfield: Understanding Where AI Actually Fits in Marketing and Sales Workflows
Over the coming weeks, a number of SMEs will be working alongside Cranfield University and the Bettany Centre for Entrepreneurship to explore a practical question that is becoming increasingly relevant across business. Where does AI actually fit within day to day marketing and sales workflows?
There is no shortage of discussion around AI. Most of it focuses on tools, features, and outputs. Content can be generated faster, emails can be written automatically, and processes can be partially automated. From the outside, this suggests rapid progress. Inside most SMEs, the reality is different.
The gap between AI capability and business reality
For many businesses, the challenge is not access to tools. It is how work is structured. Lead follow up is often inconsistent, with delays or missed opportunities. Content production tends to be reactive rather than planned. Customer communication varies depending on who is responsible. Key activities frequently rely on founders or a small number of individuals, creating bottlenecks that limit consistency and scale.
AI is being introduced into this environment, but it is often layered onto existing workflows rather than reshaping them. The result is increased activity without a corresponding improvement in clarity or consistency. This creates a clear gap. What AI appears to offer at a surface level does not always translate into meaningful change within the business.
There is growing discussion around governance, risk, and structured adoption of AI in marketing, particularly in frameworks such as those outlined by Smart Insights (https://www.smartinsights.com/artificial-intelligence/). However, much of this remains at a strategic level, rather than grounded in how SMEs actually operate day to day.
What this work is focused on
Rather than starting with AI use cases, the focus is on understanding how marketing and sales workflows actually operate within SMEs.
This includes looking closely at:
where time is being lost across daily activity
where processes break or rely on individuals
where work is inconsistent or difficult to repeat
where visibility into performance is limited
These are not new problems. What is changing is the context in which they sit. AI has the potential to reduce friction across these areas, but only if it is applied in a way that aligns with how work is structured. Without that alignment, it tends to increase output without improving outcomes.
How the approach differs
The approach being taken through this work is deliberately practical. It starts with direct interaction with SMEs, focusing on how work actually happens rather than which tools are in use. Conversations are structured around workflows, decision points, and points of delay or inconsistency.
From there, patterns are identified across different businesses. Where the same issues appear repeatedly, they are treated as signals rather than isolated cases. These signals are then mapped to specific stages within marketing and sales processes. This is not a process of demonstrating tools or proposing solutions upfront. It is a process of understanding first, before deciding where AI may have a meaningful role.
Building a working model alongside the research
Alongside these conversations, a working system is being developed to explore how signals, analysis, and outputs can be structured in practice. This model ingests external inputs such as articles, updates, and relevant market content. These are then aggregated and processed through an analysis layer before being turned into structured outputs.
The purpose is not to automate content production in isolation, but to test how structure can be introduced into what is often a fragmented process.
For a deeper breakdown of how this system is being built, see: Ai content workflow systems. This provides a way to observe how signals are interpreted, how outputs are formed, and where consistency begins to break down. Even at this stage, a number of limitations are clear. Inputs vary significantly in quality and relevance, making consistent interpretation difficult. Content can be generated reliably, but aligning it to a specific audience or business need is far more complex.
What is already emerging
A number of early patterns are starting to appear. First, signals are rarely as clear as they seem. Without filtering and prioritisation, too many inputs are treated as equally important, which reduces the usefulness of outputs. Second, generating content is relatively straightforward, but generating relevant content is not. Without a clear connection to context, audience, and purpose, content becomes activity rather than contribution. Third, content on its own does not create value. Value emerges from how decisions are made about what to produce, when, and why.
Where this work currently sits
At this stage, what exists is a functioning pipeline that can ingest signals, process them, and generate outputs. However, it does not yet determine what content should exist. There is no fully defined trigger logic. Signals are not consistently prioritised or filtered. Outputs are not always aligned to specific audiences or stages within the customer journey. Governance is still minimal, and human input remains essential at multiple points. This is not a limitation of the tools. It reflects the complexity of the underlying problem.
The role of human judgement
One of the clearest findings so far is that human judgement remains central. AI can process information, structure outputs, and increase speed. What it does not do reliably is determine what matters. It does not prioritise signals, define context, or assess commercial relevance. These decisions remain human.
This connects to a broader point explored here: Ai judgement vs Ai automation. As systems become more capable, the importance of interpretation, validation, and direction increases. The value does not sit in the generation of content, but in the decisions that shape it.
The broader commercial implication
Across many businesses, a similar pattern is emerging. More content is being produced than ever before. Tools are improving rapidly, and the cost of generating output continues to fall. However, this increase in activity is not consistently leading to better results. Engagement remains uneven. Messaging lacks clarity. Activity does not always translate into meaningful outcomes. The issue is not a lack of content. It is a lack of structure behind it. Without clear workflows, defined decision points, and alignment between activity and business objectives, additional output simply increases noise.
What comes next
This work is ongoing. The aim is not to arrive at a single answer, but to build a clearer understanding of how AI fits within real business environments, where it adds value, and where it does not. Further insights from this work will be published here as the project develops.
Summary
SMEs do not lack AI tools, they lack structured workflows
AI increases output, but not necessarily consistency or value
The real opportunity sits in workflow design, not content generation
Human judgement remains critical in prioritisation and relevance
This work focuses on practical observation, not theory
Frequently Asked Questions
What is the main goal of working with SMEs on AI workflows?
The goal is to understand how AI fits into real marketing and sales workflows within SMEs, rather than focusing on tools in isolation. This involves identifying where time is lost, where processes are inconsistent, and where structure is missing, before applying AI to improve those areas.
Why do many SMEs struggle to get value from AI in marketing?
Most SMEs do not struggle with access to AI tools. The challenge is that their workflows are not structured. AI is often layered onto inconsistent processes, which increases output but does not improve clarity, consistency, or performance.
What are the most common marketing workflow problems in SMEs?
Common issues include inconsistent lead follow up, reactive content creation, reliance on founders for key decisions, fragmented customer communication, and limited visibility on performance. These problems reduce efficiency and make scaling difficult.
Can AI fully automate marketing and sales workflows?
AI can automate parts of workflows, particularly repetitive tasks and content generation. However, it cannot fully replace human judgement. Decisions around prioritisation, messaging, and commercial relevance still require human input.
What is the difference between content automation and content strategy?
Content automation focuses on producing output more efficiently. Content strategy focuses on deciding what content should exist, who it is for, and how it supports business goals. Many businesses automate production without defining strategy, which reduces effectiveness.
What role does human judgement play in AI driven marketing?
Human judgement is critical. AI can process and generate information, but it does not reliably determine what matters. Humans are responsible for interpreting signals, prioritising actions, and ensuring outputs are relevant to the business context.
What is meant by a content workflow system?
A content workflow system is a structured process that connects inputs such as market signals and customer needs to outputs such as articles, campaigns, or communications. It includes steps for filtering, analysis, prioritisation, and production.
How does this work with Cranfield help SMEs?
Working with Cranfield provides SMEs with a structured environment to explore their workflows, identify inefficiencies, and test where AI can be applied meaningfully. It combines academic insight with practical, real world business challenges.