AI Implementation for the Rest of Us
Improving profitability, employee skills and productivity is possible. Let's get started.
If you work in a Fortune 100 company you have a plethora of consultants, vendors and experts to help you implement AI in your organization. Even if I broaden that to the Fortune 1000 there are many people who will intelligently guide you through the process for a fee or as a part of you purchasing their services. This article isn’t for them.
For the rest of us a superior technology implementation may feel out of reach. So often technologists, myself included, have made promises that we didn’t deliver on. We told you things would be faster, simpler and “better” but somehow you still don’t feel that is true. This is not necessarily a measurable sentiment. But you feel it in your gut so you trust it, as you should.
Thanks for reading! Subscribe for free to receive new posts and support my work.
AI is a hugely important moment for humanity. It is important that you pay attention because, like the personal computer and cloud before it, AI isn’t going anywhere and yes, everyone WILL be using it. The question is will you? When? And can you safely and effectively reap real benefits for your business. The answer is “Yes, but…” or like many IT professionals will respond, “It depends…”
If you run a mid-size business with a dedicated IT department stop here and don’t proceed until you read the previous article, “Want to implement AI? Repair your IT Department first.” It addresses fundamental issues most IT department have that should, dare I say must, be addressed before you venture into this new area. Trust me and do this. You can send me chocolate later.
If, on the other hand, your IT group, either internal or outsourced, is reasonably functional you can proceed with the four keys pillars of an AI implementation.
Security
Staffing
Business Priorities
Workforce Management
You’ll note that I have not yet mentioned anything about the technology itself. Not fancy prompts or capabilities from vendors, including my own (Microsoft). That’s because that’s not the most important thing for the rest of us. I can pretty much guarantee you that your security posture and data estate are not hygienic, meaning that you have holes that could be compromised. Your data is the magic that, beyond the entirety of the web, will make AI truly work for YOU. It is the institutional knowledge that grounds the vendors’ AI experience in a relevant dataset, yours. This improves the reasoning that the large language model does on your behalf, improving the responses you will get. You still must, and I do mean MUST, check them for accuracy however they are much more likely to be relevant when grounded in your own data.
The issue is that you don’t want everyone to have access to all the data. You have confidential and proprietary content in your data estate, no matter the size of your organization. Every field service person should not see every employment contract or deal memo. This is where the foundational security work comes in.
I would be remiss if I didn’t mention that Microsoft’s capabilities attempt to make this much easier for any organization that has Microsoft 365. You can learn about these classification capabilities and tools via this documentation. If you don’t know about Microsoft’s Copilot tools I recommend starting here. For what it’s worth I think they are useful tools when positioned correctly and are backed by Microsoft’s Responsible AI principles at every step. Even if you choose another supplier, I feel the Microsoft methodology should be the benchmark you compare others too. I say that not only as an ardent fan and employee of Microsoft but also as an IT Professional with decades of experience.
Whatever AI toolset you choose, you first must think about your information architecture.
This refers to how you classify and segment your data, in your documents, your CRM system if you have one, in financial datasets and elsewhere. This can be simple or complex depending on your organization, but it must be done.
The primary reason this is not already completed is your staffing plan. WHO is accountable for this work? Do they have help? Have you done and continually enforced basic security at a user level that will protect you from the threats that bad actors wish to visit upon us? Not sure? Then it is time to start. MFA, strong passwords, classification, separation of roles. These are relatively basic concepts for enterprise corporate IT professionals however the further you get away from the Fortune 500 the less they are implemented. Training your staff on security in repeatable ways is essential to having a savvy workforce that does not inadvertently compromise your system.
Now, I want you to hear this part. This isn’t already completed NOT because you are failing, inattentive or lacking in knowledge. It’s because you are juggling a thousand and one priorities already and figuring out how to train your people to save truly confidential files differently, or segmenting your database is difficult and probably not on fire. It won’t be until it is. That’s why you have to stop right now and alter your staffing plan and priorities to address this.
Don’t go it alone. Get help. There are information technology partners that will be happy to help you along this journey.
OK – one more side note. Microsoft has a free program for this call FastTrack. They help with migrations, security, and license implementation for Microsoft 365 and Copilot capabilities which is the Microsoft Generative AI tool. It’s FREE. Use it. Again, don’t go it alone! And if you aren’t using Microsoft does your vendor provide a customer success program when you purchase their software? If not they should. Purchasing a service isn’t enough. Every organization needs to ensure customer success for long term satisfaction.
While this information architecture work is going on and you are working with your partner or vendor of choice on your implementation plan, an essential task is selecting your business scenarios.
Think about your business and decide what will help it grow the most, help you control costs or improve the well-being of your employees.
These scenarios will assist you in prioritizing your AI implementation. You don’t have to do everything at once, and likely shouldn’t. Your perception of this technology will change as you gain experience. Some suggestions are:
· Employee onboarding
· Employee training or process Q&A
· Customer inquiries
· Customer service tickets
· Creation of customer quotes
· Customer onboarding
· Product development
· Product delivery – including manufacturing improvements
· Marketing campaign creation
· Creative ideation and production
· Communications, both internal and external
These are basic business scenarios that every business can benefit from improving. Of course of you are in a high tech or engineering field there are other entry level, but specialized scenarios that you’ll want to consider. Every organization can and should be considering AI improvements and agents for these foundational elements.
What you are creating is a competitive edge. The quality of your AI use will differentiate organizations and beyond that will become a basic customer expectation. This is AI for the rest of us. Implemented in scenarios that provide personalized, actionable experiences and remove the repetitive tasks from our workforce.
Workforce management is the final key pillar. It’s time to change your approach to skilling your workforce. You are probably making training available, even online or with support, doing a bit of onboarding and hoping that your employees will figure it out. You even treat yourself like this. This has been the way due to your own time crunch. It won’t work in this case.
Learning how to interact with AI and use it for your own benefit and that of your team requires building new daily habits and working differently than before. I can personally speak to this as it took me time to get to my own personal “ah ha!” moment with these tools. Now, I constantly ask Microsoft 365 Copilot to summarize and categorize my communications channels (Outlook and Teams) for priority communications, tasks I need to accomplish and drop it in a daily table for me. It summarizes strategy documents, helps me create content, understands meetings I missed including the sentiment of the conversation and so much more. I had to change my daily habits and now I am that much more production. This took a daily effort at first.
Supporting the change in daily habits is why you need community. Well managed communities for your employees are the single most effective way to ensure success of any technical implementation. The key differentiator is bringing together a few people who will ensure that community is vibrant, has new training content and ideas and provides assistance when other employees get stuck. This isn’t the Help Desk. Community managers escalate to a help desk when the problem is a technical one. This is peer-to-peer productivity and operational support for employees.
That is a high level description and I’ve written a good deal about this in other locations as part of my job at Microsoft where my team runs many communities of this type as well as one, the Microsoft 365 Champions program, that is dedicated to those who run communities within their own organizations. I will share my thoughts in this blog as well from agnostic perspective in further articles. Another amazing location to learn about this is the CMX organization who produces the CMX Summit each year, bringing together community managers from around the globe and from different sizes & types of organizations. It’s a fantastic place to learn the art and science of community management at scale.
We’ll that’s enough! I’ve given you to do in one blog article. Stay tuned here and subscribe as I write more about AI implementations, business value and leadership topics for those in IT and beyond. Our world is changing. Now is the time for us to lean in, take the reins and ensure these changes benefit the greatest amount of people. Working together we can make sure the future is meaningful and productive for us all.