What's in this article

      You’ve probably heard plenty about what AI can do. Turning those possibilities into something useful for your own business can feel less straightforward.

      Perhaps your team has tried a few tools, but their use is still fairly informal. Or you can see that some tasks take longer than they should, without knowing whether AI would help.

      A useful starting point is to look at where work gets stuck. Which questions do staff answer repeatedly? Where does information take too long to find? What needs reading, sorting or drafting before someone can take the next step?

      Those everyday frustrations can reveal worthwhile opportunities. Here are six areas to explore, along with the questions to ask before investing.

      “Businesses will distinguish themselves by how well they use it.”

       Bill Gates, referring to AI, The Age of AI has begun (2023)

      For most organisations, that starts with identifying one everyday problem worth solving.

      1. Answering routine customer questions

      If your team regularly answers the same questions, AI could help customers get an initial response while freeing staff to deal with enquiries that need more attention.

      Typical questions might cover opening hours, service availability, delivery arrangements or how to get started. An AI assistant can use approved business information to answer suitable enquiries and direct visitors towards the next step.

      For example, a prospective customer might describe what they need in their own words. The assistant could explain the relevant service, point them towards further information and offer a route to speak to your team.

      This is worth exploring when:

      • You receive a substantial number of similar questions.
      • Answers are available in reliable, maintained information.
      • Customers would benefit from help outside office hours.
      • Staff spend time directing people to information that already exists.

      The handover matters as much as the answer. Decide which questions the assistant can handle, when it should say it cannot help and how someone can reach a person.

      To assess the results, look at answer accuracy, whether customers get useful help and how much staff time is needed to manage the system.

      2. Helping staff find internal information

      Business knowledge often ends up spread across shared drives, email threads, documents and individual employees.

      When someone needs an answer, they search several places or interrupt a colleague. That can become a daily source of delay, particularly as the organisation grows.

      An internal AI assistant could help staff find relevant information by asking a question in everyday language. Someone might ask where to find an expenses procedure, how a particular process works or which document explains a service.

      The assistant should help users reach the underlying source so they can check the detail.

      This is worth exploring when:

      • Staff repeatedly ask colleagues the same questions.
      • New starters struggle to find guidance.
      • Important knowledge is documented but difficult to navigate.
      • People spend time searching across several locations.

      Before introducing a tool, review the information itself. Old procedures, conflicting documents and unclear ownership will make reliable answers harder to produce.

      Access also needs thought. Employees should only be able to retrieve information they are authorised to see.

      A useful pilot could focus on one well-maintained set of documents, with a small group of users checking how accurately and quickly they can find answers.

      AI opportunities - 6 places AI can help your business

      3. Sorting and routing incoming enquiries

      A shared inbox can contain sales enquiries, support requests, supplier messages and general questions. Someone has to read each message, work out what it concerns and decide who should deal with it.

      AI could assist with that initial sorting. It might suggest a category, summarise the request or identify information that is missing before a colleague responds.

      For example, an enquiry could arrive with a lengthy explanation of a website problem. AI could prepare a short summary and suggest that it goes to the support team, leaving staff to confirm the next action.

      This is worth exploring when:

      • You receive enough enquiries for sorting to take meaningful time.
      • Messages arrive in varied formats and language.
      • Requests regularly reach the wrong person.
      • Staff need to read lengthy messages before assigning them.

      Start by checking suggested categories against the decisions your team would make. Ambiguous or sensitive messages should have a clear route for human review.

      If routing depends entirely on a form selection or another fixed field, straightforward automation may already do the job.

      4. Preparing first drafts of routine communications

      Your team may write similar emails, proposals, meeting follow-ups or service explanations every week.

      AI can help prepare a first draft using an agreed structure and the information supplied. A member of staff can then check the facts, adjust the tone and complete anything that needs their judgement.

      For example, following a discovery meeting, AI could help organise approved notes into a draft follow-up email covering the client’s challenges, agreed actions and next steps.

      This is worth exploring when:

      • Documents follow a reasonably consistent format.
      • Staff spend time turning notes into polished text.
      • Useful templates and examples already exist.
      • Someone can review the draft before it is used.

      Be clear about what the tool is allowed to infer. Missing prices, promises or project details should be flagged for completion rather than invented.

      Measure the time taken to reach an acceptable final version. A fast first draft offers little benefit if correcting it takes longer than writing the document yourself.

      5. Summarising meetings and lengthy documents

      Reports, meeting notes and long conversations can contain useful information that takes time to extract.

      AI could help prepare an initial summary, highlight questions or organise actions for review. This can give staff a starting point when they need to understand a large amount of material.

      For example, a project manager could use it to turn meeting notes into a draft action list, then confirm the owners and deadlines before sharing it.

      This is worth exploring when:

      • Staff regularly review lengthy material.
      • The source information can be used in an approved tool.
      • Summaries follow a useful, repeatable structure.
      • Someone familiar with the subject can check the result.

      Summaries need checking against the source, particularly where details affect decisions. A tool may omit an important qualification or describe a suggestion as though it were an agreed action.

      For meetings, also establish an agreed approach to recording, transcription and handling the information captured.

      6. Finding themes in customer feedback

      Reviews, surveys, support messages and website comments can tell you where customers are struggling. The challenge is finding time to read them together and spot recurring issues.

      AI could help group comments into themes, such as difficulty finding information, confusion about pricing or delays getting a response.

      Your team can then examine the original feedback and decide what needs attention.

      This is worth exploring when:

      • Feedback is collected across several channels.
      • You have more comments than staff can easily review.
      • You want to compare recurring issues over time.
      • Someone can validate the themes and investigate their causes.

      Check how well the system handles context. A short comment, sarcasm or a message containing both praise and criticism may be categorised poorly.

      Use the themes as prompts for investigation. Keep the original comments available so your team can understand what customers actually said.

      How do you choose where to start?

      You may recognise several opportunities in that list. Choose one manageable task to investigate first.

      Ask your team:

      "What recurring task takes up time each week, and what makes it difficult?"

      Then assess the answers using five questions:

      1. How often does it happen? A frequent task may offer more scope for improvement.
      2. How much time does it take? Record the current effort so you have a baseline.
      3. Is suitable information available? Check whether the necessary documents or data are usable and accessible.
      4. Can someone check the result? Decide who will review outputs and handle exceptions.
      5. Would the benefit justify the cost? Include setup, subscriptions, training, maintenance and staff oversight.

      For example, a task taking three hours a week represents around 156 hours over a full year, assuming it happens every week. If a pilot suggests AI could reduce that effort, compare the likely saving with the total cost of introducing and managing it.

      Treat those estimates as assumptions to test.

      “Of course, with this empowerment comes greater human responsibility.”

      Satya Nadella, The Future of Work with AI (2023)

      When AI may not be the answer

      Some frustrating tasks have simpler solutions.

      Copying details from a website form into a CRM may be resolved through an integration. Sending a reminder on a fixed date may need a standard automation. Difficulty finding documents may first require clearer organisation and ownership.

      AI becomes worth investigating where interpreting language, working with varied information or preparing a useful draft could improve the process.

      Sometimes the best first step is to improve the systems around it. Our digital transformation services take that wider view of your technology and workflows.

      AI Checklist on tablet - where to start using AI

      Start small and measure what changes

      A useful pilot has a defined task, an owner and a clear way to judge success.

      Record current performance, test with a limited group and review the outputs carefully. Look at time saved, quality, running costs and the experience of the people using it.

      At Made Simple Media, developing our own platform, Free Comments, has given us practical experience of AI features such as answering questions from approved information, summarising conversations and supporting human follow-up. That work informs how we assess potential uses with clients.

      Our AI consultancy service follows three stages: Discover, Recommend and Implement. We learn how your organisation works, assess the strongest opportunities and help put the agreed solutions into practice.

      Unsure where AI could help your business? Book a discovery call to discuss your challenges and whether a more detailed review would be useful.

      Article by David Reeder. LinkedIn Profile: https://www.linkedin.com/in/david-e-reeder/

      Related Articles

      Keep up to date

      Subscribe to receive occasional email newsletters from us.