Unlocking Knowledge: Frictionless AI Adoption for Enterprises

In previous blogs from the ‘Unlocking Knowledge’ series, we discussed how AI can boost efficiency and significantly add value to organizations. However, deploying new tools in large enterprises is challenging due to data migration and training costs. We faced these challenges while building Adam, our internal knowledge base assistant. Learn more about Adam in its introduction blog.

In this blog, we’ll dive into the decisions we made to help ensure the adoption of Adam is as frictionless as possible.

Our existing knowledge base management system, SharePoint, does not offer a useful native AI assistant or the ability to integrate an external assistant. Therefore, we considered migrating our knowledge base to a new system. However, replacing a company’s primary knowledge base is a considerable challenge. It is a vital resource that is constantly accessed and/or updated. Any disruption in its availability during migration can hinder the employees’ ability to perform their tasks leading to operational setbacks.

Even after the data migration is complete, the transition to a new knowledge base inevitably disrupts existing user workflows and requires training. This adjustment period can lead to temporary decreases in productivity as employees reorient themselves and adapt to the new system.

Beyond these practical challenges, adopting AI-powered knowledge base system introduces a deeper paradigm shift. Unlike traditional systems, which are structured and tightly controlled by administrators, AI-generated responses are dynamic and less predictable. This reduced control can erode trust if answers deviate from expected or approved content leading organizations to revert to legacy systems.

To address these concerns, we decided to build our own in-house solution. Our goal was to ensure a seamless transition that keeps the existing systems in-place while unlocking the benefits of AI.

Our solutions to these problems were simple but effective. Firstly, we decided to design Adam so that knowledge base migration will not be required. Instead of bringing data to the AI, we brought AI to the data. Adam enables seamless integration with existing knowledge base systems like SharePoint by periodically (or manually) fetching data from knowledge base sources. This synchronization system ensures that the AI always has access to the latest information without requiring any changes to the source knowledge base, which in turn ensures that existing processes and user workflows are not disrupted.

Next, we addressed the challenge of minimizing the need for user training. This proved relatively straightforward, as we realized that most employees were already familiar with publicly available chat-based AI tools such as ChatGPT. To capitalize on this familiarity, we designed Adam to be accessible through a similar chat interface. This approach significantly reduced the need for additional support or training, ensuring a smooth and intuitive user experience.

The final challenge was ensuring that the AI remained aligned with the administrator’s intent and followed the constraints they defined. We needed a simple and effective way for administrators to guide the AI’s responses. It might seem intuitive to directly change how the AI behaves by modifying its underlying model, a process known as fine-tuning which involves retraining the AI on new examples to adjust its behavior. However, the process of fine-tuning requires significant computational resources and it is not well-suited for situations where quick and flexible updates are frequently required. Instead, we developed a more elegant approach:

1. Memories: We introduced a “memories” feature that allows administrators to insert curated information directly into Adam’s knowledge base through its user interface, without changing anything in SharePoint. Using context engineering, these memories are given priority over all other data sources, including SharePoint and the AI’s original training data, when there is conflicting information. This gives administrators precise control to update or correct the AI’s responses in real time. For example, if SharePoint states that there is a weekly all-hands meeting every Friday, but the meeting is cancelled for a particular week, an administrator can create a memory noting the cancellation. Even if some employees miss the email notification about the change, Adam will still provide the correct information when they refer to it.
2. Granular Access Controls: Administrators have fine-grained access control over what information is stored in Adam’s database. This ensures that AI gets access to only the correct and up-to date information, in-turn minimizing the risk of hallucinations caused by irrelevant or outdated documents. We have already discussed fine-grained access control in detail while talking about security in Adam.

By thoughtfully designing Adam to work with existing knowledge systems, adopting a familiar user interface, and enabling administrators to guide AI behavior without technical overhead, we significantly lowered the barriers to enterprise adoption. Our approach of bringing AI to the data and not the other way around ensures that organizations can unlock the benefits of AI without disrupting their workflows. With features like “memories” and granular access controls, Adam provides a flexible, secure, and context-aware assistant that aligns with company needs from day one. This deliberate focus on usability, control, and seamless integration makes AI adoption not just possible but practical for any enterprise.

We hope these solutions offer a starting point or spark inspiration for addressing the unique challenges within your organization. If you’re looking for expert support in building or adapting AI-powered systems, our team is here to help. Please reach out to us at [email protected] to discuss how we can help you to gain AI efficiencies in your organization.

At SMS, we see the rise of artificial intelligence as an opportunity to re-imagine long-standing organizational challenges and develop innovative, streamlined solutions. To that end, we’ve made significant investments in building deep AI expertise — both to enhance our internal capabilities and to better support our clients.

With nearly 50 years of experience, SMS is a trusted provider of professional services, delivering high-impact solutions across both public and private sectors. We operate in highly secure, compliance-driven environments that demand government-grade standards, while also surpassing the agility, efficiency, and speed expected from leading private sector service providers.

If you’re looking to use AI to boost efficiency, cut costs, and maintain top standards in quality, security, and compliance, let’s connect. Our team of experts is ready to help identify opportunities tailored to your unique needs.