Regardless of the selected solution, there are some straightforward actions to take as you prepare for implementation and go live to facilitate widespread adoption and usage.
Implementing a new piece of software or a novel tool generally means users must alter their current process. While you selected your solution because of its features and functionality, how it compares to the way things work today is the end-user measuring stick. This requires some insights into the status quo.
How are end users completing the task today? How long does it take? Does something prompt them to do the task? By benchmarking your current state, users and the implementation team have a common understanding of what success might look like and how to measure it.
Next, take all the things you learned in the previous step and go about addressing them in your implementation plan. This includes the user experience (UX), such as the number of clicks to complete a task or how many times the user must navigate to a new application. Too much context switching is where workflows often begin to falter. If a task is more efficient to carry out in the new solution, adoption is much easier.
harmon.ie addresses this potential pain point by enabling end users to save emails and/or their attachments to Microsoft Teams or SharePoint without ever having to leave Outlook, which is where email interactions already occur. Instead of needing to switch applications, it all happens with a quick drag-and-drop action and a prompt to add metadata tags. Additionally, harmon.ie cuts down on user actions in this area as well by automatically adding some metadata tags for the email based on the sender, the recipient, the sender’s domain, and other important pieces of information.
By making this a smooth and simple way to complete that task, the odds of each relevant email getting shared to Teams or stored in SharePoint increase.
Deciding what to retain, for how long, and in what location and format depends on multiple factors and must thread the needle between keeping extra information around “just in case” and mitigating compliance and security risks. To ensure all the bases are covered, don’t move forward without representatives from security, privacy, identity management/provisioning, risk management, compliance, and legal. Hopefully, you are already familiar with these colleagues from collecting their ECM requirements during the vendor selection phase. If any of these stakeholders aren’t involved yet, now is the time to invite them to the party. Each organization faces unique sets of factors and decision points shaped by internal business needs and external regulatory requirements. This cross-functional team makes sure you haven’t missed anything.
Entropy is the principle that—absent any outside interference—any system will gradually decline into disorder. Entropy creeps in everywhere, from your pantry to your sock drawer, and SharePoint and Teams are no different. So, before implementing an ECM that uses metadata for search and retrieval or depends on information being stored in particular locations in a hierarchy, define a taxonomy and determine the rules of the road so your information-sharing sites stay as organized as an Instagrammer’s fridge.
For example, if you are a construction firm, you likely use addresses to classify and organize projects. If the structural engineer uses the tag “Baker221B” while the HVAC crew leader uses “221bbakerst,” they can’t easily find each other’s work and no one else is sure which is the “right” tag.
Or if each relationship manager sets up client folders using a different organization strategy (or worse, no strategy at all), the associate responsible for finding the contract and setting the client up in the system might waste many minutes per customer trying to figure out where the right information is. Assembling stakeholders into a metadata advisory council or a data governance council to create some guidelines and guardrails ensures sure chaos does not reign.
Most people’s AI experiences to date have likely been with a large language model (LLM) like Microsoft Compass or Google Gemini, or perhaps they have used DALL-e to create images or artwork. While extremely impressive, these tools ingest mind-boggling amounts of content from the entire internet. Pretty cool, if the goal is to get a summary of Catcher in the Rye or writing a poem about shoelaces in the style of Shel Silverstein.
But when the goal is to make people and processes within your organization more efficient, using input from the entire digital universe is less attractive. Conversely, sharing all your information with an AI beyond your tenant that will use that information outside your organization probably makes your security department squirm.
AI is popping up in more and more applications and every solution has AI somewhere in its roadmap. Ask questions now about your providers’ approach to AI, where the training data comes from, and how your content is mixed (or not) with the content of others.
For example, instead of scarfing up troves of emails, files, and metadata stored in Outlook and SharePoint to train a model in the cloud, harmon.ie 10 keeps your organization’s content in your Microsoft 365 instance and never combines your content with any other organization’s content. Since the model is only trained in each customer’s Microsoft 365 tenant, your data is safe and stays put.
Using this federated learning model, harmon.ie only sends changes to the model itself (which contains none of the training data). This gives users the best of both worlds—AI trained on your own data AND the benefit of a model optimized using large amounts of information—all without losing control.
All the software has been installed and provisioned, the APIs have been connected, information is flowing, sounds like the job is complete, right? Not so fast. Just creating the solution is unfortunately not enough. You also must market it internally, communicating to end users what the change means for them and when it will affect them.
Depending on the degree of change to existing processes, you may want to schedule information sessions as the go-live date approaches. Launch with some fanfare so end users know today is the day. Offer training in multiple formats (self-paced, video, step-by-step instructions) to match their learning styles and schedules. Consider creating brief cheat sheets—called job aids—to help people get the hang of new ways of working.
Finally, follow up on feedback. If everyone is complaining that the new button is in the “wrong” place or the connection to another system tends to time out, communicate your ability to address the issue and then follow through on those you can. Make sure you are offering training on an ongoing basis as people join the company or change jobs within the organization.
Organizations implementing or using an ECM to comply with regulatory expectations around record retention—where the stakes of noncompliance are high—should consider whether policies or penalties should be put in place for those not using the process appropriately. Other, less heavy-handed options include training with or without an assessment to measure learning, spot audits to track usage, and incentives and recognition for champions and power users. The importance of adoption should be used to calibrate the size of both carrots and sticks.
The ECM world is expansive and diverse, but with a little planning and collaboration with colleagues, you can select a solution that works for your organization, without reinventing the wheel or breaking the bank.