Starting again
What to check before you start another business
You've learned plenty from running a business before. Here's how to use that experience while checking what's changed about customers, prices and the tools they already have.

Use your experience and relationships to get started, then check whether people still buy and work in the ways you remember.
If you've run a business before, you probably have a more realistic picture of what starting another one involves. You know there will be customers to find, bills to pay and ordinary problems to sort out long after the excitement of the launch.
That experience can save you a lot of time. It can also make some decisions feel settled before you've really looked at them. Perhaps you know how you'll find customers because a particular approach worked last time. Perhaps you've already chosen a monthly price because that's how the previous business charged.
Before you get too far, it's worth checking which parts of your experience still fit. You don't need to forget what you've learned. You need to find out what the new business will ask you to do differently.
Make a quick list of what you can bring with you
Start with things you can use now. You might have people who trust you, knowledge of an industry, savings you're comfortable putting at risk, or a good understanding of how to run a small team. Perhaps a former colleague would like to work with you again.
Then make a separate list of the things you're assuming. Do buyers still want to book a demonstration? Will search engines send you the right people? Does the job come up often enough for a monthly subscription? Will a simpler, nicer-looking product give customers enough reason to switch?
These lists help you decide what to investigate first. Knowing five shop owners who will take your call is useful. Assuming those owners still buy software the same way they did five years ago needs checking.
Try to be specific about each advantage. “I know small businesses” is hard to do much with. “I can ask five operators to show me how they do this job” gives you a starting point for the week.
See what customers can already do with their existing tools
A feature that once needed a separate product may now come with software the customer already pays for. AI has made some useful things easier to build, including for the larger companies you're hoping to compete with.
Shopify's September 2026 overview of AI in online retail describes uses such as customer support and inventory-related work. It's guidance from a vendor, rather than an independent study of results. For someone starting a business, it's a reminder to look at the tools buyers already have before deciding what to offer them. [1]
Try the relevant features where you can, read the limits, and ask users what they think. A feature might look complete on a sales page but leave out an important part of the customer's day. It might also work perfectly well, leaving less room for your idea than you expected.
For example, an existing tool might draft customer replies but struggle when an unusual request needs approval from several people. A specialist product catalog might be harder to handle than a standard one. These are possibilities you can check with real users.
If you're relying on a better AI model as the reason people will switch, find out whether they notice the difference in their actual work and would pay for it. A difference that's impressive to you may be less important to the buyer.
Ask how people are finding services like yours now
The last business may have found its customers through search, an industry event or a social platform. You have good reason to try what worked before, but check that the same route still reaches the people you want.
Cloudflare reported in September 2026 that daily requests from AI agents across its network had grown by more than 1,700% over the previous year. That measures activity on its network. It doesn't tell you how many people bought something or prove that your customers have changed their habits. [2]
You can learn about your own market by asking recent buyers how they searched. Where did they start? Which other options did they consider? What did they need to know before getting in touch? If they used an AI assistant or a comparison tool, ask where it helped.
You don't need to pursue every new way people might discover you. Clear service descriptions, accurate prices and a straightforward explanation of what's included are useful starting points for human buyers too.
Give yourself a chance to hear from people who don't know you
Former customers and friends can be wonderful first contacts. They may answer your questions honestly and help you get a trial off the ground. They may also agree because they like you and want the next business to go well.
Use those relationships, then try the offer with a few people outside your closest circle. Notice whether their questions are different, whether they take longer to decide and how much more explanation they need.
You could ask a familiar customer whether they'd consider the same offer from a provider they hadn't heard of. Their answer is useful, though an actual conversation with an unfamiliar buyer will tell you more.
Your reputation may genuinely help you sell. Just be careful about expecting every future customer to offer the same trust, especially if someone else will eventually handle the calls and emails.
Look closely at ideas that seem similar to the last business
Suppose your previous company sold scheduling software to gyms. Now you're considering a service that prepares compliance documents for training providers. This is a hypothetical example. At first, both might look like help with regular admin for small businesses.
But a training provider might need a quality manager to approve the purchase. A mistake in a document might have more serious consequences than a scheduling mix-up. The work might pile up before an audit and barely happen the rest of the year.
Those differences would affect the price, the sales conversations and how much checking the work needs. The quick demonstration and monthly subscription from the gym business might be a poor fit.
You still have useful experience explaining services, getting small teams started and supporting busy customers. Use it to explore the differences before committing to the familiar approach.
Decide what you need to see before spending more
If you have money and experience, you can keep improving an idea for quite a while without finding out whether enough people want it. A polished website and a capable team can make the business feel further along than it is.
Before spending heavily, choose a few things that would make you comfortable taking the next step. Depending on the service, that could be a paid trial, repeat use without reminders, a customer outside your network or a job delivered well without you stepping in.
Also consider what would make you pause or change direction. Maybe every buyer needs a completely different service. Maybe their existing software already handles the part they care about. Maybe finding each customer costs more than you're likely to earn from them.
These are planning suggestions, not universal rules about how many sales you need. Writing them down early can make it easier to judge the idea fairly once you've put your time and reputation into it.
It's okay to start small again. A few jobs done partly by hand can help you understand the awkward cases before you hire a team or build a full product. Get specialist help where a mistake would be costly, and delay expenses that mainly make the company look established.
You have the benefit of knowing how much work comes after a promising idea. Give yourself time to check the new market, and you can use that experience where it will help most.
Sources and a few notes
We checked these sources on . The examples and suggested steps are our own unless we say otherwise.
- AI in Ecommerce: 7 Key Use Cases for 2026 (opens in a new tab)
Shopify ·
Vendor guidance on ecommerce uses. Not independent evidence of outcomes.
- The Internet has a second audience (opens in a new tab)
Cloudflare ·
AI-agent request growth on Cloudflare's network. Traffic is not the same as purchases or buyer preference.


