How To Incorporate AI And ML On A Budget: 14 Tips For Businesses
Artificial intelligence and machine learning are two facets of modern technology that can have a massive positive impact on businesses today. However, many innovations within the AI and ML space are among the most cutting-edge technologies currently available, and even some of the more widely used business technologies can break the bank if not implemented with a smart plan in place.
However, thanks to more commercial providers entering the arena, the investment required for adopting either of these solutions is no longer excessive. With careful budgeting and a detailed plan, even small and medium-sized enterprises can implement relevant AI and ML technologies. Here, 14 professionals from Forbes Technology Council offer valuable tips for companies that want to successfully incorporate AI/ML on a budget.
1. Identify The Problem You’re Trying To Solve
I step back and think about what the real problem we’re trying to solve is, and how AI and ML can help. Like any other technology investment, I think it’s important to start small with the end in mind and determine how the learnings will be acted upon. The secret to AI is having processes and workflows in place that act upon the results of AI. – Tammy Hawes, Virsys12
2. Start With Non-Core Functions
There is always a lot of focus on the core functions of the business—the expertise, big budgets and focus always go to them. What often gets neglected are the non-core functions, including some that can bring significant improvement to the quality of work (and in today’s world, quality of life) of an employee. Once the team sees what it does for them, acceptance is easier and you get a longer leash. – Suresh Sambandam, Kissflow
3. Identify How You Will Measure Success
Leaders need to define what success looks like to them. Ask, what am I hoping to get out of these technologies? How will I measure success? Answering these questions will ensure you can track progress and communicate with your team on areas of improvement. Setting goals and benchmarks will guide the implementation of the technology and ensure you only invest in areas that provide the most ROI. – Sanjay Malhotra, Clearbridge Mobile
4. Start Small, Gain Momentum, Grow
One thing to realize is that you must start small. Realize that there is no silver bullet. This is why when you first get a taste of ML or AI that is directed at your business, start small. The natural push of momentum will occur if the business proposition involving AI and ML “has legs.” After that, iterate and grow and take advantage of your pricey investment. – WaiJe Coler, InfoTracer
5. Get Support From Your Team
Integrating AI with your daily operations is not an overnight task and, to succeed, you need as much support from your team as possible. Start by tackling a visible problem that will get a win on your record and get people motivated for AI. Then, focus on increasing the ROI and impact of AI projects. This initially small investment will make it easier to implement and scale AI across the company. – Nacho De Marco, BairesDev
6. Look At Proven Use Cases
We know AI is a trending buzzword and everyone wants to utilize it, but most businesses don’t know where to start. If you’re trying to employ AI without first considering the value of the outcomes, the investment is likely wasted. The easiest way to start is to look at proven use cases where the value has already been shown inside your industry, and then take those cases as lessons to build on. – Kevin Macdonald, Kit Check
7. Develop A Risk Management Strategy
With anywhere from 10 million to 100 billion time-varying signals in the enterprise attack surface, cybersecurity is no longer a human-scale problem. AI can perform complex, multivariate analysis, helping to find a signal in the noise, identifying the most significant risk sources and prioritizing gaps for remediation. This allows a business to maximize risk reduction with minimal resource and effort. – Gaurav Banga, Balbix
8. Ensure You Have Well-Defined Data
Computers don’t think. They process data in a very structured and methodical way very quickly to render the appearance of intelligence or learning. Therefore, if you are thinking of bringing these technologies into a company, start by ensuring you have extremely well-defined data and rules that can be used as the basis to build out an AI or ML system. – Rena Christina Tabata, ShareSmart (Think Tank Innovations Ltd.)
9. Take A Plug-And-Play, Modular Approach
Look for an end-to-end partner that offers modular solutions. A modular framework lets you minimize initial investment, implementation risk and ramp time, while integrating new technology into one or two key business areas. When you’re ready, expand AI and ML into new business areas with services that are designed to work with your existing systems so that the transition is smoother. – Ron Cogburn, Exela Technologies
10. Start With A Low-Risk ML Pilot Study
AI and MI technology can mean a lot of things, depending on your goals. I recommend getting started with a low-risk machine learning pilot study. Use machine learning software for data classification and organization. This type of machine learning is low risk and relatively simple to implement, yet it can still save your analysts lots of time. – Nelson Cicchitto, Avatier Corporation
11. Leverage Open Source Tools
There are a lot of open source products in AI/ML. From creating chatbots to training an engine to classify images and comments—it will take less than a day. Start small by using your existing developers to get acquainted with open source tools and run small experiments. By conducting small experiments, you will be a champ at the end of six months. Don’t get bogged down by all the jargon. – Viplav Valluri, Fastlane Americas, Inc.
12. Map And Understand Existing Workflows
The trick to a machine learning strategy is first pinpointing workflow problem areas that are prime for ML. Start by mapping all workflows in your organization and clearly defining each step in each one. This lets you identify and prioritize “best fits” (i.e. opportunities) for an ML solution, including associations to other workflows that could benefit. A focused strategy is the start of cost savings. – John McDonald, ClearObject
13. Index Opportunities Against Impact
To get the best bang for your AI buck, you’ll want to force-rank—using a spreadsheet—all potential AI/ML opportunities against five to seven criteria, such as acquisition and implementation cost, economic benefit, AI maturity, ability to manage, etc. Identified criteria should be researched and, using both qualitative and quantitative insight, input into the model to help drive a decision. – Michael Gurau, Kaiser Associates, Inc.
14. Don’t Let Your Data Scientists Work In Silos
Organizations that want to get started with machine learning usually set up a team of data scientists, reporting to the business. The data scientists will create great models in their notebooks, but struggle to get it out into production. One piece of advice is to organize data scientists from the beginning in cross-functional mixed teams, together with business, product owners, IT and operations. – Christoph Windheuser, thoughtworks.com
Discover Past Posts