Keith Jones of Sw7, one of the leading independent technology accelerators in Africa, offers practical tips for start-ups in the artificial intelligence space.

If there’s one industry that seems to be the hot topic nowadays, it has to be artificial intelligence (AI). For Keith Jones, co-founder of Sw7, there’s never been a better time to be part of the field.

“We’ve seen more tools available than ever before and a huge proliferation of data,” he says. “In the last two years, we’ve also seen shifting market behaviours that have brought enormous change and an appetite for digitisation.”

Despite the pandemic’s impact on just about every sector, Jones still finds that “a lot of money is flowing in” and that demand for “AI-as-a-service tools” is increasing. So, if you have a business that you want to start and/or scale, here are his seven tips.

1.       Solve in sequence

Different people love different things. For data scientists, it’s building great algorithms. For software developers, it’s creating cool tech. For business people, it’s making sure the outputs are useful to the business as a whole and the clients it serves.

“The first step is to start with the end in mind,” Jones says. “When you’re designing and building a system, ask where the data is going to be consumed inside the business and ask how you’re going to change behaviour or create action that creates value. This will frame your data blueprint and product architecture.”

2.       Seek accessible datasets

What are you selling to the customer? It might sound like a simple question, but many AI businesses get it wrong when they offer a service that will take weeks or months and require a lot of effort on the customer’s part. A far better option is to sell a solution that depends on the data you already have and not on the data that the client will need to find or prepare.

“In the post-Covid market, there’s a low appetite for large, complex projects,” Jones says. “People just aren’t interested. They don’t have the capacity to sign off the budget or mobilise resources. So make sure that you start with an easy process with low friction and a shorter time to value.”

3.       Solve existing pains

The movie Field of Dreams is famous for the line, ‘If you build it, they will come’. But it’s far easier to sell to a customer with an existing problem, especially when they’ve already allocated a budget to solve it. Trying to create a need when one isn’t recognised (or doesn’t exist) will lead to a long and complex sales cycle.

“Your business will scale based on the quality of customer problem statements that you serve,” Jones says. “So if you can say ‘I’m going to do X, and this is the benefit you’re going to receive’, you can get approval a lot quicker.”

4.       Do ‘enough’

Technical teams often get excited by the power of the available tools. This can lead to overdesigned solutions with lots of unnecessary features, when something good enough is, well, good enough.

“More features create more confusion,” Jones says. “For example, building a real-time solution for a client team that meets monthly is a huge expense that doesn’t deliver any value. It would be better to deliver the data only when it needs to be delivered and make it super simple for the customer to understand.” 

5.       Build what you can sell

Similar to the previous point, it’s easy for technical teams to create solutions that apply to many different sectors. But this makes it hard to take those products to market because it’s hard to tell who exactly they’re for and why they’re needed at all.

“You can’t serve many sectors, build a distribution channel, or mobilise resources into multiple channels because there’s a huge effort required,” Jones says. “So you need to choose one sector if you want to scale. It’s about building what you can sell, not selling what you can build.”

6.       Create high customer empathy

By picking a specific sector, you can find and service a niche with the kind of solutions the customer feels are built just for them. Operating in a niche also eliminates competition, makes finding distribution channels easier and allows you to create the kind of trust that can lead to referrals.

“In the current market post-Covid, the value of peer referrals has gone up,” Jones says. “So a niche is going to be one of the best tools that you create in lead generation… The narrower the niche, the higher the customer value is and the easier the sale will be.”

7.       Create value through action

The value of a business comes down to the value the customer receives. But if the data or insights you generate aren’t acted on, there’s no value for the customer. More than that, if the customer doesn’t act on your insights to derive value, it’s not their problem – it’s yours.

“We can produce way more insights than a business can ever act on,” Jones says. “So the main problem is getting people to act on the data. If you focus on making sure the customer takes action and that the actions add value, that’s how your business is going to scale.”

Three steps to AI growth

A business will only scale as fast as it can turn data into action and insight. Jones breaks it down into three steps:

  1. On-ramps – gathering, structuring, cleaning, and aligning data so it’s usable. This takes 40% of the effort and results in products that are easy to sell (focused on features instead of benefits).
  2. Highways – processing and analysing data to generate insights. This takes about 20% of the effort and helps you build the business case.
  3. Off-ramps – turning the data into insights and actions that create value. This takes about 40% of the effort and results in products that are easy to buy (focused on benefits instead of features).

“No matter how easy you make it for the customer, it’s not easy enough,” he says.  “Remember that a business only scales when it’s easy to buy.”

Related

Three Lessons for (Business) Leaders in the Age of AI

Three Lessons for (Business) Leaders in the Age of AI

Clever Co-piloting: Is A.I. a Coach or a Tool?

Clever Co-piloting: Is A.I. a Coach or a Tool?

AI. Not as bad as your current boss?

AI. Not as bad as your current boss?