AI in Operations: Focus on Augmenting Talent Not Replacing It

The TIME story about the SMB that cut from 48 to 30 employees using AI is making the rounds. Founders are forwarding it. Boards are asking about it. Operators are quietly worried about it. Here's what no one writes the follow-up on: that company's quality, retention and rework costs 12 months from now. Cutting headcount is the easy part. Holding the line on what customers actually receive is the part that breaks teams. A few things I'd want to see before I believed a number like 48 → 30: - Who owns the AI output before it reaches a customer? A tool without a reviewer is a draft, not a deliverable. - What happens on the third escalation? AI handles 80% of tickets. The other 20% are usually the ones that decide whether a client stays. - Where did the institutional knowledge go? The people who left took workflows in their heads. AI didn't learn those — it learned the easy parts. - What's the rework rate at month six? "Same revenue with fewer people" is a Q1 headline. The Q3 number tells you if it stuck. If you're a founder being asked "why aren't we doing this," the honest answer isn't "AI doesn't work." It's: show me your QA layer before you cut a single role. The headcount cut makes the press release. The QA layer is what keeps the customer. #AI #SmallBusiness #Operations #Leadership #Founders

75% of AI's economic gains are going to just 20% of companies. PwC just dropped the numbers and they should make every executive uncomfortable. The divide isn't about who has AI. Almost everyone has AI. It's about what they're actually doing with it. The companies winning aren't using AI to do their old jobs faster. They're using it to enter markets they couldn't before. Build revenue streams that didn't exist 18 months ago. Cross industry lines that used to be walls. The losing companies? Running pilots. Saving 20 minutes on email summaries. Automating the PowerPoint nobody wanted to make. That's not transformation. That's table stakes dressed up as strategy. And here's the uncomfortable truth: most AI “transformations” are productivity theatre. Visible enough to satisfy the board. Shallow enough to change nothing. In 12 months, the gap between the top 20% and everyone else won't look like a performance gap. It'll look like different industries entirely. Which side of that line is your company genuinely on? #AI #BusinessStrategy #AIAdoption #FutureOfWork #Leadership

Everyone's talking about which AI model to use. Almost nobody is talking about what you're going to feed it. You can have the most powerful LLM in the world, but if it doesn't know your organisation, it's just a very expensive search engine answering questions with someone else's answers. RAG bridges that gap. Think of it as giving your AI a proper induction, here's how we work, here's what we know, here's our way of doing things. And the tech? Honestly, that's the easy bit. Where it gets uncomfortable is when a new joiner asks the AI a simple question on their first week — and it pulls an onboarding guide that's two restructures out of date. Or surfaces an internal process that the team quietly stopped following months ago but never got around to updating. Or answers a customer query using product information that changed after the last pricing review. The AI isn't lying. It's just working with what you left lying around. So, before you greenlight the AI layer, go one level deeper: Do your teams trust your internal knowledge base enough to bet a customer conversation on it? If the answer is anything other than a clear yes — that's where the real transformation work begins. The AI is ready. The question is whether your organisation is. ♻️ Share this with the leader in your org who's driving the AI agenda. 👇 What's been the biggest blocker in your AI journey — the tech or the data foundation? I'd love to hear. #RAG #GenerativeAI #AITransformation #EnterpriseAI #KnowledgeManagement #LLM #DigitalTransformation #FutureOfWork #TechLeadership #AIStrategy

Shared Language Is a Force Multiplier Early in my career at ADP, I saw a Professional Services bottleneck that was not really a capacity problem. It was a clarity problem. We did not have a shared language for requirements, request data, triage, or assignment readiness. Everyone was working hard, but too much energy was spent interpreting, reworking, and competing for priority. When we aligned on the language of the work, the process changed. Requests moved faster. Leaders had better visibility. Teams made decisions with more confidence. And we eliminated 80%+ of the friction points. That lesson is even more relevant today. In my current transformation work, shared language around customers, products, entitlements, and journeys is creating the foundation for better systems, better service, and better decision-making. It is also changing what becomes possible with AI. Doing the work today to ensure AI spends less tokens tomorrow is a second-level force multiplier while analyzing trends because: The clearer the language, the less explanation required. The more consistent the data, the faster the insight. The stronger the foundation, the more valuable the output. Without shared language, AI has noise. With shared language, AI has context. And context is where transformation starts. Reflection from the work of transformation. #colibrigroup #disruptingforgood

The question I hear most from operations leaders right now is: how do we get started with AI? It is the wrong question. The right question is: if AI could see everything happening in our operation today, would it have what it needs to help us? Trusted data. Defined metrics. A clear picture of what good looks like and what exception looks like. Most operations cannot answer yes. Not because they have not invested in technology. Because the foundation underneath it was never built to support that kind of question. The good news is that building that foundation does not have to take as long as it used to. AI itself is changing what is possible in terms of cleaning data, identifying inconsistencies, and surfacing the gaps that matter most. The work that once felt slow is starting to move faster. But the intent still has to come first. The organizations getting this right are the ones that made a deliberate decision to build on solid ground before asking AI to run on it. If your answer to the question is not yet, that is where the work begins. #Operations #AIReadiness #DataStrategy

Most companies are approaching AI backwards. They're asking "how many roles can we eliminate?" instead of "how much creative capacity can we unlock?" The conversation is dominated by efficiency gains, cost reduction, and headcount optimisation, using AI to replace people doing administrative work rather than using AI to free those same people from administrative work so they can do something more valuable. It's a fundamentally different question, and the one you choose shapes whether AI becomes a tool for growth or just another round of doing more with less. Here's what's being missed: The person currently spending 60% of their week on status updates, data entry, and coordinating between systems isn't just an admin cost to eliminate. They're someone who understands your operation, knows your customers, sees the patterns, and has ideas about what could work better, but never has time to act on any of it because they're buried in process work. Give them AI to handle the administrative burden and you don't just save time. You unlock the strategic thinking, creativity, and problem-solving capability that was always there but never had space to surface. That's a completely different value proposition than "we automated this role." The companies treating AI as a headcount reduction tool will get short-term cost savings and long-term capability loss. The ones treating it as a way to let people work at the top of their potential will build something more valuable teams who can think, adapt, and innovate because they're not spending half their energy on tasks a machine can handle better anyway. The technology is the same. The outcome depends entirely on what question you're asking it to solve. Are you using AI to eliminate roles, or to eliminate the work that's stopping people from doing their best work? #AIinBusiness #FutureOfWork #Leadership #EmployeeExperience #WorkplaceInnovation

"We can't afford AI right now." You already are. 2.5 hours a day. I hear this on every second discovery call. Here's the math they're avoiding. The average knowledge worker loses 2-3 hours a day to repetitive manual work. I measured my own last week. It was 2.5. Copying data between systems. Formatting reports. Rewriting emails. Rebuilding the same spreadsheet for the fifth time. At 20 people, that's 40 to 60 hours of wasted labor per day, which at $75 an hour fully loaded means you're writing a check for $3,000 to $4,500 every single day you're not automating anything. That's $750K to $1.1M a year. You're not "not affording AI." You're paying for not having it. Here's what most leaders miss. "Can't afford" assumes AI is a cost. It's not. It's a cost offset. You spend $1 to stop spending $5. If the math doesn't work on that, you don't have an AI problem. You have a business problem. "Can't afford it" is usually code for "I don't know where to start." That's a different conversation. And it's a cheaper one. --- What's the task you'd automate first if the budget appeared tomorrow? ♻️ Repost to help someone automate what's slowing them down. 🔔 Follow Jean-Luc Bryar for more on AI and operations. #AI #operations #professionalservices

Most companies aren't ready for AI. Not because the technology is too complex. Not because they don't have the budget. Not because their team can't handle it. They're not ready because of something much simpler: 𝗧𝗵𝗲𝘆 𝗵𝗮𝘃𝗲𝗻'𝘁 𝗱𝗼𝗻𝗲 𝘁𝗵𝗲𝗶𝗿 𝗵𝗼𝗺𝗲𝘄𝗼𝗿𝗸. Here's the test I use: If one of your team members gets sick tomorrow, can someone else step in and do their work — without three days of onboarding and 1,000 questions? If the answer is no, you're not ready for AI. Because here's the truth nobody at the AI conferences tells you: 𝗔𝗜 𝗻𝗲𝗲𝗱𝘀 𝗰𝗼𝗻𝘁𝗲𝘅𝘁. 𝗠𝗮𝘅𝗶𝗺𝗮𝗹 𝗰𝗼𝗻𝘁𝗲𝘅𝘁. About your business. Your processes. Your customers. Your edge cases. If that context doesn't even exist in a clear, written form for your own people — how is an AI supposed to use it? I see this every week through one of our businesses – MoreBit Technologies GmbH a managed IT company. We support businesses from 10 to 1,000+ employees. And the pattern is always the same: → Processes live in people's heads, not in documents → Critical knowledge sits with one person who's been there for 15 years → Tickets get solved the same way again and again, manually → Nobody documents the "why" — only the "what" Then the CEO reads a McKinsey article and asks: "How fast can we implement AI?" The honest answer: 𝗦𝗹𝗼𝘄𝗲𝗿 𝘁𝗵𝗮𝗻 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗸. 𝗔𝗻𝗱 𝗳𝗮𝘀𝘁𝗲𝗿 𝘁𝗵𝗮𝗻 𝘆𝗼𝘂'𝗱 𝗲𝘅𝗽𝗲𝗰𝘁 — 𝗶𝗳 𝘆𝗼𝘂 𝗱𝗼 𝘁𝗵𝗲 𝘄𝗼𝗿𝗸 𝗳𝗶𝗿𝘀𝘁. The companies that win with AI in the next 5 years won't be the ones who buy the most tools. They'll be the ones who finally documented their business properly. Tools come second. Clarity comes first. #ai #leadership #operations #strategy #unternehmertum #business #aiforbusiness Nico Haslberger

Organizations Are Systems Your AI project didn’t fail because of the AI. It failed because you optimized one silo while ignoring the rest of the system. Every organization is interconnected — Marketing feeds Sales, Sales triggers Operations, Operations touches Finance, Finance affects HR decisions, and so on. What this means for AI: → Data you need is owned by a different team → Governance requirements vary by function → Fixing one area can break another → Cross-functional buy-in isn’t optional — it’s the project The consultants who succeed think in systems. The ones who fail think in departments. Which department do you think is most overlooked during AI initiatives? #OrganizationalDesign #AIStrategy #ChangeManagement #SystemsThinking #DigitalTransformation

Most companies are asking AI to speed up work that they should have questioned in the first place. The instinct makes sense. You take the existing process, including the monthly reporting cycle, the consolidation routine, and the variance pack, and look for where AI can save time. It works. You get faster. But you're still producing the same output for the same reasons. Automation does that. Redesign is a harder problem. Redesign starts with asking why the work exists. Which tasks survive only because humans were the only ones doing them? Most teams have never had to answer that question; bandwidth was always the bottleneck. AI removes the excuse. The problem is that answering it well requires understanding both the business process and what AI can reliably do. Most organizations have people who know one. Almost none have people who know both. BCG published research on this a few weeks ago. When AI drives productivity rather than cutting headcount, ROI becomes harder to define and defend. So leaders default to what they can measure. Headcount goes down. The gain disappears into the quarterly numbers. Nobody builds the case for what the team could have become instead. Companies that redesign rather than automate don't just hold onto people. They come out with more capacity than they started with, not just lower costs. I've worked through this in commercial finance. Which parts of the analysis can AI own? Which still needs a human in the loop? Which parts of the process probably shouldn't exist at all? It's a different kind of question than most finance teams are used to asking. Most organizations haven't begun that conversation yet. And in most of them, nobody's job is to start it. #FPandA #Finance #AI #WorkforcePlanning #FutureOfWork