(This column originally appeared in Forbes)
Many studies are saying that small businesses are “adopting” AI. But in actuality most are simply playing with chatbots to help them research topics or craft better emails. But what about real ROI?
If you’re running a small or mid-sized business you’re inevitably wondering how to best leverage AI so that you can increase productivity and profits. But, like most of us, you’re probably not a tech guru and don’t know where to start. I always tell my clients it’s really just two choices: buy or build.
Buying AI For Your Business
When it comes to buying that involves looking at your existing software applications — including Microsoft Office and Google Workspace — and then either upgrading, subscribing (and implementing) new features or getting help from service providers or consultants to take better advantage of Copilot and Gemini.
In addition, you should be considering the same course of action with your existing accounting, CRM (customer relationship management), HR, order entry, estimating, and quoting applications as well as your project, inventory or service management platforms. Talk to your vendors. Find out what they’re rolling out. Take advantage of their AI features. Get training. Get training. Get training.
Or you can build.
Dmytro (or Dima) Negodiuk knows all about that. He’s part of a growing number of “Fractional AI Officers” — tech pros that small and mid-sized companies hire to knock out specific solutions using AI by bringing together their expertise and publicly available tools.
Fractional CMOs, CFOs and CROs are becoming more common in smaller organizations, where they can be used to provide a high level of expertise but on a part time basis. The expertise provided by FAOs like Negodiuk is now quickly being recognized for similar reasons as more business owners need someone that can help implement solutions that bridge the gap between AI hype and practical implementation. Because they’re part-time their cost is much lower and ROI can be more easily determined.
Negodiuk knows what he’s doing. Besides being a consultant, he runs — or should I say his AI bots run — a handful of other businesses that require very little oversight and which — he claims — earn him as much as $15,000 a month. He knows how to use this technology to save time and money.
AI adoption is not limited to a specific sector. Since emigrating from Ukraine less than five years ago Negodiuk has used AI to run content, pricing, reviews and customer service across four e-commerce platforms, and, for another company, created a voice agent that handles more than one hundred daily outbound cold calls and lead qualifications in three different languages without a sales rep. He has built additional solutions that automatically enrolls students at an education services company with automatic parent communications and monitors a logistics company’s online reputation with incident reports
These are real life applications in AI that are not difficult to create and provide enormous productivity benefits.
Where To Start Building AI Solutions
If you want to learn how to implement AI nowadays you don’t a lot of coding skills. Today’s tools — particularly Claude Code or Open AI’s Codex — allows a non-developer luddite to “vibe code” solutions where the user gives these platforms prompts and they write the code for you. But it’s not as easy as it sounds. Negodiuk spent countless hours teaching himself everything he needed to know about AI coding via YouTube videos, online courses and through trial-and-error. He had to familiarize himself with popular AI chatbot platforms and all the other tools required to build large language models, add tags, write descriptions, parse data, ensure security, define governance and other necessary tasks.
But that shouldn’t put off someone who wants to automate with AI. Because the key isn’t coding. It’s defining where AI can help your business. And to do this Negodiuk recommends an AI Readiness “audit,” a service that he routinely performs and which involves discovery calls, detailed questionnaires and oftentimes onsite visits.
“An audit identifies high-impact opportunities,” he says. “It’s a strategic roadmap that outlines potential AI implementations and return-on-investment.”
Negodiuk stressed that an AI “solution” shouldn’t necessarily be stand alone. He said that businesses should focus on integrating AI into their existing systems rather than replacing them, mostly because AI “will transform, but not replace software companies and will augment rather than completely eliminate enterprise software ecosystems.”
Where to focus first? Negodiuk has found success in using AI to improve invoicing and payment processing (“everything connected with money in your company can be automated”), enable quicker financial reporting, complementing customer service humans with better information (“everyone wants to change customer support, but in the future, speaking with a real person who’s augmented by AI will be a huge benefit.”), sales outreach using AI voice agents and notifying businesses of news and events that specifically impact them.
Start Small And Build
After the audit and after identifying key areas for improving, it’s important to take small steps, rather than turning existing systems and processes upside down. Negodiuk advises that companies experiment and scale projects gradually and that it’s not necessary to do a full transformation immediately. Even designing a simple solution that uses AI to help reply to emails or automatically chat with customers on your website or automate certain workflows or generate marketing content is a good first step.
“You just connect everything,” he said. “People continue to work how they worked before. But it’s important now to experiment.”
Easier said than done. AI solutions can be affordably created to specifically solve problems in your business. They just need the right person to help identify the need and then implement.
