Beyond email drafts and chatbots
Artificial intelligence has moved beyond writing emails, summarising documents and answering questions. The latest generation of tools can work across documents, spreadsheets, presentations, emails, project systems and company knowledge. They can research, analyse, create deliverables and complete multi-step tasks under human supervision.
For SMEs, this creates a significant opportunity. Work that previously required several disconnected systems, repeated administration and hours of manual effort can increasingly be completed through a single AI-supported workflow.
However, buying access to ChatGPT or Claude does not create an AI-enabled business. Successful AI implementation requires clear priorities, the right tools, secure access to information, well-designed workflows, team training and measurable accountability.
Most businesses do not need more AI tools
They need a clear operating model for using them. Many businesses are already experimenting with artificial intelligence. Employees may be using ChatGPT, Claude, Microsoft Copilot or other tools individually, but that activity is often fragmented.
Common problems include:
- Different teams using different tools without agreed standards
- Employees manually copying information between AI and business systems
- Sensitive information being entered into unsuitable personal accounts
- AI outputs being accepted without adequate review
- Repetitive processes remaining largely manual
- No measurement of time saved, cost reduced or quality improved
- Useful experiments never becoming reliable business processes
The objective should not be to introduce AI everywhere. It should be to identify the workflows where AI can produce a meaningful and measurable improvement, implement them properly and then expand from a proven foundation.
From AI assistant to AI coworker
Traditional AI tools primarily responded to individual prompts. Newer tools such as ChatGPT Work and Claude Cowork can take a business objective, gather relevant information, work through multiple stages and produce a completed deliverable for review.
Instead of asking AI to write one paragraph or analyse one document, a team can increasingly ask it to:
- Review information from several business systems
- Analyse a spreadsheet or operational dataset
- Identify risks, trends and actions
- Produce a management report or presentation
- Update supporting documents
- Prepare follow-up communications
- Repeat the workflow on an agreed schedule
The business remains responsible for setting the objective, controlling access, reviewing the work and approving important decisions. But the amount of administration between the initial request and the completed output can be substantially reduced.
Using ChatGPT Work in an SME
ChatGPT Work is designed to move beyond conversation and help complete substantial business tasks. It can gather context from files, connected applications and existing workflows before producing completed outputs such as documents, spreadsheets, presentations, analysis and web-based materials.
For an SME, this could include:
- Turning financial data into a monthly management report
- Reviewing CRM, email and meeting information before a sales meeting
- Converting research into a commercial strategy presentation
- Producing a board pack from departmental updates
- Comparing supplier quotations and identifying commercial differences
- Creating recurring reports from operational information
- Reviewing project information and highlighting overdue actions
- Turning customer feedback into structured product recommendations
ChatGPT Business also provides a shared workspace for teams. Shared projects can bring together common instructions, files and working context so that employees are not creating separate versions of the same process. Company knowledge and connectors can also allow ChatGPT to retrieve relevant information from approved business systems, subject to the organisation's permissions and configuration — particularly valuable where information is spread across SharePoint, Google Drive, Slack, GitHub, CRM platforms and project-management tools.
Using Claude Cowork in an SME
Claude Cowork is Anthropic's agentic working environment for general business and knowledge work. Rather than only discussing a task, Claude can work directly with selected files and connected tools to complete a multi-stage assignment.
Examples include:
- Reviewing a folder of contracts and creating an obligations register
- Consolidating several spreadsheets into a management dashboard
- Organising and renaming business documents
- Producing a presentation from research and supporting information
- Preparing customer or account briefings from connected systems
- Comparing policies and identifying gaps or inconsistencies
- Building recurring marketing or management reports
- Turning source documents into polished Word, Excel or PowerPoint files
Claude Cowork can show the steps it is taking, allowing the user to follow its work, provide further direction and review the result. It can also use connectors to work with existing software and information sources — particularly relevant to businesses that need AI to operate within established workflows rather than creating another isolated application.
ChatGPT or Claude: which is right for your business?
There is no universal answer. Both platforms offer powerful reasoning, document creation, research, analysis, connectors and business collaboration capabilities. The correct choice depends on how your business works.
The assessment should consider:
- Where your company information is currently stored
- Which applications your team uses every day
- The documents and outputs you need to create
- The complexity and frequency of the intended workflows
- Data protection and confidentiality requirements
- Required user and administrator controls
- Whether the task needs research, analysis, file editing or system actions
- The experience and working preferences of the team
Some businesses will standardise around one primary platform. Others may use ChatGPT and Claude for different types of work. The important point is to make a deliberate choice and establish an approved business environment. Allowing employees to adopt a growing collection of disconnected consumer tools usually increases risk and reduces the value of implementation.
Where AI can create value in an SME
Management information and reporting. AI can consolidate departmental updates, analyse performance data, identify exceptions and prepare recurring management or board reports — reducing the time spent assembling information while giving leaders more time to discuss decisions, risks and actions.
Finance and commercial administration. Management-account preparation, variance analysis, cash-flow commentary, quotation comparison, invoice processing, contract review and commercial reporting. AI should not replace professional financial judgement, but it can remove a significant amount of the preparation and administration around it.
Sales and customer management. Account briefings, analysis of customer communications, structured meeting notes, follow-up actions, opportunity qualification and consistent CRM information. It can also help commercial teams understand why opportunities are progressing, stalling or being lost.
Operations and project delivery. AI can bring together information from emails, spreadsheets, meeting notes and task-management systems, identify overdue actions, highlight conflicting priorities and produce concise operational reports for managers.
Marketing and business development. Market research, campaign planning, website content, proposal development, tender responses, case studies and repurposing of content across channels. The greatest value comes when AI is given clear brand, audience and commercial context.
Company knowledge. AI-supported knowledge systems can make information easier to retrieve, compare and apply — improving onboarding, reducing repeated questions and protecting organisational knowledge when employees leave.
Recruitment and people management. Preparing job descriptions, structuring interview information, developing onboarding materials, comparing training requirements and supporting workforce planning. Human review remains essential where the output could influence employment decisions.
Software and internal software tools. AI-assisted development makes it practical for an SME to create lightweight internal systems, dashboards, portals and workflow tools that would previously have been too expensive or too slow to justify.
My approach to AI implementation
I approach AI as an operational and commercial transformation programme, not simply a technology purchase.
1. Understand the business. The starting point is the company's objectives, operating model, systems, constraints and existing pain points — where time is being lost, information becomes fragmented, quality varies and capacity is restricting growth.
2. Identify valuable workflows. Opportunities are assessed against business value, implementation difficulty, data availability and risk. The initial priority is a small number of high-value workflows rather than a long list of speculative ideas.
3. Select the right technology. The solution may use ChatGPT Business, ChatGPT Work, Claude for Work, Claude Cowork, Microsoft tools, existing business software, bespoke automation or a combination. Technology is selected around the workflow — not the other way round.
4. Build a controlled pilot. A clearly defined pilot is implemented with agreed inputs, outputs, responsibilities and success measures — to prove the workflow works in the real business environment, not merely during a demonstration.
5. Introduce governance. Appropriate controls around access, confidential information, personal data, checking requirements, approved tools and accountability. Higher-risk activities have stronger controls and explicit human approval points.
6. Train the team. Employees need more than a list of prompts. They need to understand when to use AI, how to provide suitable context, how to review an output and where the limits of the system sit. Training is based on the work employees actually perform.
7. Measure and expand. Implementation is measured against agreed operational outcomes: hours saved, processing time reduced, cost avoided, error rates, reporting speed, conversion rates, delivery performance, customer response times and employee capacity released. Successful workflows are then standardised, documented and extended across the company.
Responsible AI implementation
AI implementation must be proportionate to the information being used and the decisions being supported. A low-risk marketing brainstorm does not require the same controls as a workflow involving customer records, employee information, contracts or financial decisions.
A responsible implementation should address:
- Approved AI platforms and accounts
- Access permissions
- Personal and confidential information
- Data retention and model-training settings
- Human checking and approval
- Accuracy and source verification
- Intellectual property
- Cybersecurity
- Automated decision-making
- Record keeping and accountability
Business versions of major AI platforms generally provide stronger privacy, administration and data controls than unmanaged personal accounts. However, selecting a business plan is only one part of governance — the company remains responsible for how the technology is configured and used.
Practical executive leadership, not theoretical AI advice
AI implementation often fails because it is treated as a standalone IT initiative. The technology may work, but the business process, management responsibility, employee behaviour or measurement system does not change.
My background combines executive leadership, operational transformation, engineering, software development and practical AI adoption. I work with leadership teams to move from the initial opportunity through to implementation, adoption and measurable business results.
My wider experience includes supporting a portfolio with more than €250 million in revenue and 3,000 employees across software, energy, engineering and high-growth businesses. Previous implementation work has produced efficiency improvements of more than 30%. This means I can assess AI from the perspective of the board, the management team and the employees who will use it every day.
Who I work with
I work with owner-managed, investor-backed and growth-focused SMEs across the UK. Core sector experience includes:
- Software and SaaS
- Energy and offshore wind
- Engineering and manufacturing
- Professional services
- Technical recruitment and workforce businesses
- Investor and portfolio-company operations
- B2B service businesses
I am based in Newcastle upon Tyne and work with companies across the North East, London, the wider UK and Europe.
AI implementation services
AI opportunity review. A focused assessment of your current processes, systems and potential use cases, followed by a prioritised implementation roadmap.
AI workflow pilot. Design and implementation of a specific AI-supported workflow with defined success measures and controls.
ChatGPT and Claude business rollout. Selection, configuration and practical adoption of ChatGPT Business, ChatGPT Work, Claude for Work or Claude Cowork across an SME team.
AI operating model. Development of the policies, responsibilities, governance, training and reporting needed to use AI consistently across the business.
AI automation and internal tools. Creation of AI-supported automations, dashboards, internal applications and workflow systems around specific operational requirements.
Fractional executive support. Ongoing CEO-level leadership for businesses where AI implementation forms part of a wider growth, restructuring or operational transformation programme.
Start with one valuable business problem
An SME does not need to transform every department at once. A better starting point is one recurring problem that consumes time, causes delays, creates errors or restricts growth.
That might be a monthly report that takes several days to assemble. It could be a sales team that lacks consistent account information, a tender process that relies on repeated manual work or an operations team managing projects through disconnected spreadsheets.
Once a valuable workflow has been redesigned and proven, the business has a practical foundation from which to expand. If your company is already experimenting with AI but struggling to turn that activity into a consistent business improvement, I can help you define the opportunity and implement it. If you have not yet started, I can help you identify where AI is most likely to produce a measurable return without creating unnecessary cost, complexity or risk.
