Most small/medium size businesses (SMBs) and their leadership teams are past the initial shock of “what is this Ai thing; and what do we do with it?” We are well into the “we need to use Ai or we are going to be left behind. Let’s put a Tiger Team (argh!) together and figure out…” stage.
SMB’s are adopting GenAi in their business processes at different levels and for different purposes.
Some SMBs are just fine using “Ai-like” capabilities that are embedded in many business applications. I’m talking here capabilities like Amazon’s Ai “just-walk-out”; Ai is advertised, but the reality behind is more mundane, if not down right anything but. Still, with these capabilities, the work gets done, and there is some marketing rizz that allows the SMB team to say “we are using Ai”.
Other SMBs are doing great with “legit” Ai capabilities embedded in top-tier business applications like Adobe, salesforce, Wix; you name them. This is a smart play when the efficiencies gained and problems solved are narrow in focus and scope. Like above, the work gets done and the SMB can advertise the use of Ai to employees and customers alike.
Both of these adoption levels are great when the goal is to automate or eliminate repetitive tasks, or modifying them through a higher level of functionality. These are sure winners for SMBs looking for that kind of relief, and initial dabbling into the world of GenAi users.
Beyond this level of GenAi adoption, we get into the world of using “proper” (felt like a Brit for a moment there) GenAi applications.
There is a plethora of them, operating in siloed functionality and purpose. Most SMBs start by using them as part of a “shadow Ai” movement within the business. This is an “ok way” to test the waters and capabilities, but a problem for steady state operation. Just ask your friendly IT department what they think of shadow IT ops… ;). Not just an issue of control, but one of corporate security.
The hungrier or more serious about Ai SMBs are moving further into the GenAi value chain, getting into the area of “connecting” Ai applications and “integrating” them. In this zone, is where SMBs are getting, and will get the highest ROI.
There is value and a big differences between both “connecting” and “integrating” GenAi applications. The first approach (connecting them) is like a function/linear equation or sum. The second is more like an integral, or the area under the curve.
Connecting Ai Applications
As it sounds, this means that GenAi applications are connected via “calls” to the application(s) on the tech stack; 1:1, or perhaps 1:many.
Let’s say and SMB is using a business stack application, and they would like to augment it using Ai. The functionality needed is identified, the application is selected, and the connection to the Ai application is made. For example, you want to use ChatGPT behind the scenes to write a blog within your Kajabi or Wix website. You want to generate an image using DALL-E to populate the main image of your Mailchimp campaign vs. using their native tools.
In both cases, the connection makes the process transparent to the user and prevents the “sneaker net” to work in over time by cutting and pasting from one application to the other.
This level of integration is solid, but it focuses on discreet tasks, that are part of a greater project.
Integrating Ai Applications
This is when GenAi applications are connected via “calls”, but rather than 1:1 or 1:many, they are connected in many:many form. Moreover, both the prompts and results are shared between the applications, producing more coherent and complete projects, rather than individual tasks.
Pulling the thread on the case above, imagine that with one simple prompt you could start a chained reaction, where applications like ChatGPT, Midjourney, Synthesia, HeyGen could all be tasked together, and summoned one by the other, to generate with one action a training course, an exam, a piece of collateral - all from one action.
Beyond efficiency, the integrated nature arguably presents a more coherent and seamless product end-to-end. This is where the greater ROI lives.
Your SMB does not need to start at the integrated level, or quite frankly end there. That said, if your SMB is looking for ways to:
Make it easy to adopt its use
Make it simple and invisible to the users
Seriously compete with their peers or punch above their weight
Then connecting and integrating GenAI capabilities to their processes and workflows is the place to be. This is where you go from dabbling, to truly getting the most out of GenAi in high gear! - when it is embedded in the workflows, rather than “bolted on” as an afterthought.
If you are in an SMB, I’m curious to see where your company is in the adoption process, and what your experience has been so far.