Knowledge Management using AI

Imagine starting your day, ready to kill at work, immediately bombarded by emails, Slack notifications, and tickets/ questions demanding immediate attention.

For customer-facing teams, this scenario is all too familiar.

They’re the frontline of the business, dealing with everything from customer queries to product issues, and often, the flood of information they need to manage becomes overwhelming.

Talk to your sales team, or your customer team right now

You’ll mostly hear them stress about:

“Where’s the new product update details, I swear I saw this on Slack 🥹”
“Can someone tell me if that product bug was fixed from last week? John is ANGRYY 😡”

🎵 It’s the end of the world and I feel fine plays in the background 🎵

Why does this happen?

Short answer: Information overload, too many tools, constant app switching, unable to efficiently capture, organize, and utilize knowledge, and unable to focus on work.

Long answer:
The current state of knowledge management is a mess.

I’ve identified 3 sets of issues that customer-facing teams, such as sales, support, and success, face due to this:

1. Information Overload
   1. Your teams are drowning in a tsunami of data.
   2. They’re inundated with a constant stream of data from various sources like:
      1. Product updates, customer support chats, market trends, sales questions, internal communications, and more.
   3. According to the tech giant Microsoft, 68% of knowledge workers crib about not getting enough uninterrupted time to focus on their core work
2. Fragmented Knowledge across platforms
   1. On average teams are using upwards of 5-10 tools and 1000+ docs
   2. Each doc or tool houses its own set of data and knowledge
   3. Accessing knowledge across these tools is hard and requires constant app-switching
      1. A senior HR director told me, that Adobe’s HR team uses upwards of 55 tools (think about the total no of tools in the entire org 😭)
   4. Gartner and Forrester found that customer-facing teams at SaaS companies jump between 20+ tools (how is this even productive? 😅)
   5. Teams spend 10+ hours a week just LOOKING for stuff. That’s more time than most of us spend on Netflix (or not maybe)
3. Time spent searching for info
   1. The combination of Point 1 and Point 2 leads to a significant amount of time wasted on searching for relevant info
   2. Downstream effects?
      1. Frustrated teams
      2. Grumpy customers,
      3. And you, pulling your hair out.

I feel this and YOU! (hang in there my friend)

I get it, this sucks, but what can you do?

You could explore if AI can solve this for you. Hold on, let me explain…

(I swear this is not a sales pitch)

1. Generative AI:
   1. You could leverage AI to create new content based on your knowledge. This process over time helps you fill all holes in your current documentation.
   2. Imagine John (your most important customer) asks your support team what’s happening with the bug raised last week.
      1. Inefficient approach: You might panic, ask multiple people, and wait hours for the answer. OR
      2. Efficient approach: You could ask AI to help you find this answer. It will search your Jira tickets, Slack conversations, and existing Zendesk tickets and help you find the exact status of the bug. You could also ask the AI to craft a nifty reply for John and save the day 💪🏻
2. Searching for info using normal human prompts: Wait, what? If you’ve used Slack and when you don’t type the exact keyword, you mostly don’t get what you’re looking for. This is where techniques like Natural Language Processing (NLP) come into play. Without getting too technical, what this means is you ask and you shall receive!

All this sounds cool, but how will it help me?

That’s a fair question, I think onboarding an AI knowledge management tool has 3 advantages:

1. Reduces cognitive load: AI takes over mundane tasks by automating routine tasks, freeing up mental capacity for cool, complex problems.
2. Thought amplifier: AI helps you brainstorm better, finds that needle in the haystack, and figures out the single source of truth so you’re on top of your work.
3. Learning on Steroids: AI personalizes your learning experience, gives tailored feedback on your documentation, and a lot more.

Let’s get deeper into how this will help YOU:

Sales

From “Let me get back to you” to “I have this info right here!”

Sales calls are stressful, your rep only has one agenda: win the customer. But what if your prospect asks a question, and your rep has no idea how to answer it?

This happens more than you think, and those frantic keyword searches to find product updates, testimonials, case studies, or other crucial info take precious minutes that are a lifetime in sales (there goes your customer 🥹).

Here’s how AI can help here:

1. Instant know-it-all mode: What if your reps can have a “know-it-all” earpiece you see in movies? So when a prospect asks a question, your rep has enough context and confidence to answer it – that’s basically what AI can do. With one search your sales team can get what they need without searching across multiple tools, chats, emails, or docs. (not all superheroes wear capes)
2. Newbie to an oldie: Remember how long it took to learn all the product info, ICP/ persona info, sales tooling, and processes? With AI, your new hires can tap into the collective knowledge of your sales department from day one without waiting on anyone.
3. Sales workflows: Sales AI workflows can jot down notes, update the CRM, and even create to-do lists from them. So your reps can focus on winning customers while AI handles the behind-the-scenes grunt work.

Support:

From “Please hold while I check this for you” to “Getting 5 stars on every ticket?!” 🦸‍♀️

The life of a support person is hard and their minds are constantly stimulated due to the high volume of tickets they get daily (yup, this never ends).

Here’s how AI can help here:

1. Quick wins: AI-powered knowledge bases can guide support agents to the most relevant information, learning from context from documents, ticketing systems, product updates, FAQs, etc so it serves up answers faster than you think what to say 💪🏻
2. Conversation contextualization: Do you like when you go on a website, and narrate the entire problem only to do it again when you’re passed on to another agent, sucks I know. AI creates a neat package of previous chats, so the next rep can jump in like they’ve been besties with the customer forever.
3. Customer support workflows: Would it not be nice if every time a product feature is released, the customer support team gets a chat notification of what was fixed, what has been deployed, and which customers complained about this? Yes, this is exactly what AI can do (and this is just the tip of the iceberg).

Customer Success:

From reactive to proactive without losing your sanity

Customer Success teams need to stay on top of their customers. It’s not easy when you have 100’s of them (a good problem to have though).

Retaining them is the biggest challenge so you have consistent revenue coming from them.

Here’s how AI can help here:

1. Upsell opportunities $: To upsell, you need a complete understanding of your customer—from their product requests and bug tickets to their overall adoption journey. An AI trained on your docs and tools can help with the above, and also analyze this data to suggest which customers might churn, where you can upsell, etc.
2. Instant answers: Having the right knowledge about your product is a MUST. So a good knowledge management system will ensure you always have the right thing to say in front of your customers.
3. Customer success workflows: Imagine your prospect asked about integrating with Linear two months ago, but it wasn’t on the product roadmap then. Fast forward 60 days, and your Jira now shows a ticket for this integration that’s been deployed. This signals an upsell opportunity with your existing client. With this workflow, your CS team will receive a notification to contact the client and upsell this.

Hmm, you’ve my attention, how do I approach this?

Good point! It’s easier than you might think.

Implementing AI knowledge management systems usually requires 3 steps:

1. Check your tech pulse aka readiness: The first step is to evaluate your documentation and decide how many tools you want to integrate. Don’t rush this process—AI isn’t going anywhere. From my experience, it depends on what stage you’re at. Reflect on this, so you can pinpoint gaps that the new tool can address.
2. Picking the right tooling
   1. Build vs buy: This is an eternal question only you can answer. Here are a few considerations to think about this though:
      1. Do you have the resources i.e money and time
      2. The right technical team
      3. Understand LLMs, RAG, accuracy, fine-tuning (and all other AI concepts)
      4. Integrations and workflows
      5. Support
   2. Integrations: This is another important consideration factor, the tools you’re exploring or building will integrate with your existing and evolving tech stack? You’ll need an AI system that can communicate with all the other tools in your department. To see ROI, it should be able to read and write your existing tech stack.

What does the implementation look like?

I’ve typically seen orgs get the most value when AI systems are rolled out in phases. (We can agree to disagree 😂)

1. Phase 1: Start small and focus on quick wins
   1. Pick 1-2 teams, and ask them to integrate their tools and documentation
   2. Implement AI-powered QNA and search on this data set
   3. Focus on improving the accuracy of the tools
   4. End goal: Get buy-in from these teams and achieve a faster time to those “aha” moments with the information.
2. Phase 2: Figure out automation and workflows
   1. Identify soul-crushing repetitive tasks that your teams are doing mindlessly
   2. Build smart AI-driven workflows
   3. Monitor and optimize performance based on employee feedback.
3. Phase 3: Tech stack makeover
   1. Once you remove redundant steps in your work, you realize, “Well I don’t need so many tools!”
   2. Replace all tools, and workflows where there are multiple point systems.
   3. Slowly but surely, move to a sleeker, smarter system.

Overcoming the bumps in the road

1. Data quality: Ensuring high-quality, well-organized data is crucial for AI to work effectively. If the data you input in these AI tools is flawed, you won’t see any ROI. As the adage goes “Garbage In, Garbage Out”.
2. Privacy Control: AI systems are deeply integrated with your systems, tools, and docs. You should implement robust security measures and ensure compliance with data protection regulations.
3. Employee adoption: Get your team on board with training. If there is no buy-in, the adoption will be low. Make it clear that AI is meant to aid their efforts, not take over their jobs.
4. Keep tweaking: Like a good recipe, your AI setup needs constant tasting and adjusting. Keep gathering feedback to improve the knowledge management process.

How do you measure success?

AI tools are just that—tools. So, they need to meet strict benchmarks and metrics.

How do you know if your system is pulling its weight? I would keep an eye on these quantitative and qualitative metrics:

1. Productivity improvements and time saved with faster information retrieval and automated workflows
   1. Impact on sales: ↑ win rates, ↓ sales cycle
   2. Impact on CS: ↓ first contact resolution, ↑ CSAT, ↑ NPS
   3. Impact on support: ↓ support ticket backlog, ↓ escalation rate, ↑ ticket resolution rate

My closing thoughts

By now, I hope you get the idea that integrating AI into your current systems is becoming as essential as your morning coffee (and we all know how important THAT is).

It’s your ticket to smarter, faster, and happier customer-facing teams.

The future is here, and it’s def going to be powered by AI (either with you or without you)

Time to plug our AI knowledge tool, obviously 😉

If you think you’re evaluating AI for a customer-facing team, talk to us and we can figure out if Albus can help you:

1. Book a demo with our team
2. Sign up for your free trial

In case if you’ve any questions or need any more reading material please feel free to reach out to me at [email protected]

Catch you in the next one!