How My Small Business Is Using AI

I’ve always been drawn to the latest developments in tech, often buying gadgets just to see what they can do. The Rasberry Pi captured my imagination for a while and I make a mean model of an orange with a 3D printing machine.

My latest obsession is artificial intelligence. Like many small business owners, I initially found it easy to think that I should leave AI to Cambridge University, or the Facebook and Googles of this world. But actually with tools like Microsoft Azure around, it’s surprisingly easy for small software creators to get started on a journey of artificial intelligence.

However, while I’ve learnt you can get started easily, I can also say it’s not so easy to go the distance. That takes a fine balance of ingredients. We have made some significant strides to automate some elements of bookkeeping already, but there is still much to do.

While we are still getting our recipe right, fundamentally these are the things we’ve found work for our business.

Focus on your homework

It goes without saying that before you start you have to know what AI can do for you and your customers. You may find that AI is all about making your internal processes quicker and easier to do, it could be about improving the customer’s experience.

It could be both. But we discovered it’s very easy to get carried away with everything that AI can do and lose sight of your core business activity.

You must stay focused and establish a list of genuine possibilities and then rank them according to the amount of impact they will have versus cost and effort. If you don’t do this you won’t get any one project off the ground with any success.

"We quickly realised that we needed every employee involved in the AI initiative or we would alienate people"

Skills

AI can’t be built into software without people to make it happen. That means you need highly skilled people with depth of knowledge. We employ brilliant technicians but they aren’t equipped for this sort of initiative.

Hiring an AI specialist is an option, but competing with the likes of Amazon for candidates isn’t something we can do. Plus there are plenty of other technical skills we need aside from AI - so we have to prioritise the versatility of our next hire.

In the end we decided to collaborate with Nottingham University and provide internships to research students. It allows us to test theories with someone at the forefront of research and provides MSc students with a practical environment within which to test their learning.

Making collaborations work

The first thing to note is that good researchers are in high demand so you need to be very clear about what you want from an intern, and be prepared to wait for that person to become available. It may seem obvious but if you don’t get the right person working with you then you won’t get the intellectual capability interns promise.

At the same time, don’t’ get hung up on finding a specialist in your business sector. They really don’t need to be. You can teach them that bit. They just need to know enough to apply their expert knowledge to the scenarios you want to explore.

It pays to wait for the right person

That brings me to my second learning about collaboration, you need a meaty problem they can solve and you have to prepare yourself for failure. That’s important because failure is research.

In our case we need to introduce software that can categorise a receipt into taxable or non-taxable, simply from the receipt image.  Understanding what data will get us to this point, and how much of it we need, is integral to the success of making it happen. It’s not something we can take a stab at, but equally it’s not something we have the time to assess.

The final point to make about collaboration is that while this approach may be ‘free’ make sure you are giving the student something in return. They have to benefit from the experience too otherwise they will switch off.

Funding

Collaborations of this sort will help you finance the research in a reasonably risk free way and build the business case for the next steps. You’ll have a clear view of the operating or capital expenditure to get the programmes up and running.

If you find self-funding is not the best option for your project, then a research based business case will allow you to confidently approach the bank and/or local council initiatives that support innovation. You might even find you qualify for the Knowledge Transfer Partnership, which would fund placements of PhD students for much longer periods of time.

Be prepared to change

As well as helping us prioritise the projects we would run, and how they slotted into the business as usual programmes, the research helped us see which team members we needed to train and on what, and the technology they would need to use to meet our goals.

But more than this we quickly realised that we needed every employee involved in the AI initiative or we would alienate people. If people see AI as a threat to their job, which it isn’t in our case, then it’s game over.

So we’re making sure everyone knows about the potential of AI. Now, with a little knowledge of AI automation, our accountants make suggestions on the small things we can tweak that will make a big difference to business performance.

Make sure people have access to your plans

Test it and test it again

Of course there is no point in investing in the analysis or the development of AI apps if what you produce is not fit for purpose. This is where it gets interesting because that ‘magic’ requirement employees or customers demand is very hard to test. So you need very logical testing programmes, supported by skilled people who can see whether the data you’ve selected is being used properly and producing the desired outcome.

If you don’t have research, development and testing working together then you can’t create a product that solves a person’s problem accurately 90% of the time. This is when AI feels like magic to the person using the app and that’s when an industry like accountancy can get remotely close to becoming ‘sexy’.

Be a pioneer

Embarking on AI has been hard work and tougher than we thought. Certainly the best advice I can give from our experience is to use go for it but use research as a spring board. Our research project moved us from a position of ‘unknown unknowns’ to ‘known unknowns’.

We now know what we need to do and how we should structure the way ahead. But most importantly it’s helped us make a pioneering leap in the right direction, one that has garnered awards, and ultimately proved that AI doesn’t have to be consigned to the billion dollar companies of the world.

James Poyser is co-founder of inniAccounts.