How small businesses can use AI

This article comes from “Making AI Work”, the limited newsletter of MIT Technology Review, which is based on the cross-industry application of LLMs : To receive it in your inbox, register here. From accounting to design to market research and product development, a breathtaking range of skills is required for the management of a company. A large company can hire experts for these tasks, but small businesses cannot always afford this luxury. This is where the AI comes in. Today's AI models work well on these tasks. The trick for small businesses is to understand where AI is good enough and where not. An area where a ‘sufficient good’ AI can already be very valuable for small business operators is the communication of secretarial competencies and the handling of basic administrative matters. Let's take a look at how a private teacher uses it to improve his records and save his time. Case study

Sam Finnegan-Dehn works in fundraising for a charity organization, but also works as a tutor for mathematics and philosophy for university students from his home in London. Through this part-time business, he can use his degrees in philosophy and share his love with customers. But meetings with students are just a fraction of the work you need to be a good tutor. He also plans lessons, finds fresh reading material, creates tasks, sends bills and keeps up on new research – all in addition to his regular work. In view of these requirements, Finnegan-Dehn does not have as much time as he would have liked to extend his follow-up plan. That is why he has turned to AI to help manage the everyday aspects of his company. He says that the AI has taken over a secretariat function in all its digital notebooks, recording memories of its customers' progress and new readings to keep up to date. He describes the use of AI as a kind of second memory that helps him to combine ideas he has written down in different places. While using various tools like Claude and ChatGPT has experimented, he has now landed at Notion AI, because it can be better integrated into his remedies that can be found on his notebook tabs in the Notion app. Finnegan strain does not use AI to create teaching materials, but it allows Notion AI to record meetings with its customers (after obtaining their consent) and then uses their automated summaries to refine its teaching strategy. For example, if he finds, on the basis of the summary of the AI, that it seems that a particular technique did not help a student, he could change his approach to the topic next time. In addition, Notion AI also helps him with the goal, the writing of teaching notes, the accounting and the creation and synchronization of social media contributions. As far as the objective is concerned, Finnegan-Dehn says, for example, that he understands the long-term goals of his company, but does not always know the concrete steps to achieve them. He uses AI to close these gaps. He starts writing up a “North Star” target – for example, having a certain number of customers by the end of the year. Next, he asks his AI to generate the steps he has to take to get there using the profile created in the app. He can then think about the results and decide which tasks he wants to tackle first. The tool

Notion has been an important provider of note software for many years. The KI add-on published at the end of 2023 now has tools that allow it to interact with many other online productivity platforms. There is an email client, calendar integrations and a newly published agent. And although this level of access has led to data protection concerns, it can also become a pretty powerful virtual assistant. Many of the tasks to which Notion AI aims are less creative and more routine: for example, synchronizing information across documents or searching for old scribbles. This makes the tool particularly attractive for small business operators who have limited bandwidth, especially for minor work. Other companies develop tools aimed at certain industries. Grandma’s Quilt Shop in Yuma, Arizona uses Rain, for example, with a handcrafted software suite to create inventory descriptions and prices for its stock of fabric designs. The owners claim that this AI tool shortens the time to list articles by 60 to 80%. However, there are also disadvantages, since Finnegan strain sometimes described some of the peculiarities of Notion AI as “clobite”. And the AI add-on for notion costs $20 a month. As with all new tools, small business operators should carefully consider how the potential gains and problems behave in relation to the cost of doing the job themselves. User tips

Consider these points if you think about whether AI can help you manage a business or make part of your working life a little easier. - Look before you jump. There. LLMsbased on the data you have entered to answer your questions or do tasks, you want to provide them with information that is convenient for you and the model. For many of these notebook AI services, this means, for example, that they use their platform to alert, so that you do not have to enter or upload notes later. For this reason, it is a good idea to carefully weigh your options before choosing an AI-based ecosystem. - Work on your strengths. Consider which skills you lack internally and check if AI can either help with the training or take over these tasks for you. Be aware, however: AI hallucinates and makes mistakes. So think about where accuracy is required and let people there control. - AI is not always the best tool. It's okay to use something from the bar if that's the better choice. For example, it will be safer to use existing payment processing platforms such as Shopify or Square as a vibe code platform using AI. - Consider using local models for confidential information. In our reporting, we have dealt with the risks that online AI models are associated with sharing sensitive data, and there have been many reports about how AI companies collect your data when you ask their chatbots. Even if your company does not process personal data, there may be some things you prefer not to publicly disclose. In these cases, the use of an open source model that locally draws conclusions about your prompts can be used instead of chatGPT or Claude or other proprietary models are a good option. Fortunately, some LLMs can now be run on laptops and small desktops. Learn how to set up and use one. Sign up for “Making AI Work”, the limited newsletter of MIT Technology Review, which deals with the application of LLMs in healthcare, climate technology, education and more. Deep insight

Artificial intelligence

Would you like to understand the current status of the AI? Check out these charts. According to Stanford's KI-Index 2026, the AI is on the rise and we have trouble stopping. 10 things that are currently important in AI

The key overview of the 10 technologies, emerging trends, bold ideas and powerful movements in the AI in 2026. Musk vs. Altman Week 1: Elon Musk says he had been cheated, warns that AI could kill us all, and admits that xAI models of OpenAI distilled

Musk kept a cool head and the lawyer of OpenAI bombarded him with boring questions about his motives to sue the company. Stay connected

Get the latest updates from

MIT Technology Review

Discover special offers, top stories, upcoming events and more.