Best of LinkedIn: Strategy & Consulting CW 40/ 41

00:00:00: provided by Thomas Allgaier and Frennus, based on the most relevant LinkedIn posts about strategy and consulting in CW-Forty and Forty-One.

00:00:08: Frennus specializes in B to B market research for strategy and consulting teams with a focus on tech and ICT.

00:00:15: So today we're diving deep into the strategy and consulting trends that have really been bubbling up across LinkedIn these last few weeks.

00:00:22: Yeah, and if you look closely at what people are sharing, the big message seems pretty clear.

00:00:25: For you, the listener, it looks like the days of just, you know, playing around with AI pilots, they might be ending.

00:00:31: Strategy leaders, they're really doubling down now on operating discipline, getting AI working at scale, and crucially responsible governance.

00:00:40: Right, so less experimenting, more actual doing.

00:00:42: Exactly, we're shifting away from those small proof of concept things, those pilots, towards... practical full scale deployment and needing measurable ROI plus real ecosystem collaboration.

00:00:54: So what we want to explore today are the concrete steps needed, you know, across tech, governance and leadership to actually make this shift happen.

00:01:01: Okay, that's a perfect setup.

00:01:02: Let's start with that biggest theme then.

00:01:04: AI at scale.

00:01:07: It feels like, I don't know, the whole consulting world sort of woke up and realized building a cool AI demo is one thing, but making it actually generate value.

00:01:17: That's the hard part.

00:01:18: You nailed it.

00:01:19: Todd Lorne made this point really well.

00:01:20: He said, the conversation isn't if anymore.

00:01:22: It's how.

00:01:24: How do we get AI adopted at scale to drive real business value?

00:01:28: you can point to?

00:01:29: Because the initial buzz wears off fast if there's no substance behind it.

00:01:33: Right.

00:01:33: And people start asking, OK, where's the return?

00:01:35: But then you see data like Gus Cavalcanti shared.

00:01:37: And honestly, it's sobering.

00:01:39: How

00:01:39: sobering are we talking?

00:01:40: He's showing that a massive, eighty percent of organizations are getting zero ROI.

00:01:45: Zero.

00:01:46: from their AI investments right now.

00:01:48: Wow,

00:01:49: eighty percent.

00:01:50: That should set off some alarm bells in boardrooms.

00:01:52: Definitely.

00:01:53: And Anthony Jardim suggests a big reason for this is often just strategic misalignment.

00:01:58: You know, the excitement for the tech runs way ahead of the actual structural changes needed to make it work.

00:02:04: So you end up with fancy tools, but no real change in how the business operates.

00:02:08: Exactly.

00:02:09: And what happens then?

00:02:10: The AI just gets tangled up in the existing messy processes.

00:02:14: Yawkin Thomas Moore from PWC had a really strong take on this.

00:02:18: He basically warned, look, AI is not going to magically fix your broken R&D.

00:02:23: To scale AI effectively, you first need a solid foundation.

00:02:27: Clean,

00:02:28: consistent data, structured systems underneath it all.

00:02:30: He mentioned specific systems, right?

00:02:32: PLM, ERP, MBSC.

00:02:35: Can you unpack those a bit?

00:02:36: Why are they so fundamental?

00:02:37: Sure.

00:02:38: Think of them as the... the digital plumbing of product development and operations.

00:02:43: ERP handles your core business stuff, finance, HR, supply chain.

00:02:47: PLM manages everything about the product itself, designs, changes, documentation.

00:02:52: And MBSC, Model Based Systems Engineering, focuses on using models for requirements, design, analysis, making sure everything's consistent.

00:02:59: Got it.

00:03:00: So they're the backbone.

00:03:01: They are the backbone.

00:03:02: And if that backbone is, well, messy or disconnected, if the data flowing through it is bad, then AI just churns out, as you put it, noise, not value.

00:03:10: It's

00:03:10: like trying to build a skyscraper on sand.

00:03:12: Precisely.

00:03:13: It's the digital version of, you know, painting over rest.

00:03:16: You have to fix the underlying structure first.

00:03:19: That really reframes the AI scaling challenge.

00:03:21: It's a data and process architecture problem before it's an algorithm problem.

00:03:26: So, okay, let's say an organization has done some of that hard work, their house is relatively clean.

00:03:31: What's the recipe for success then?

00:03:34: Neha Cabra from Quantum Black AI by McKinsey.

00:03:37: She laid out a very clear methodical approach.

00:03:40: It starts naturally with the data structuring it, annotating it properly.

00:03:45: Then critically, you put human experts in the loop.

00:03:47: They act as the arbitrators setting the logic, the guardrails.

00:03:50: So humans are still essential.

00:03:52: Absolutely.

00:03:53: Then you prove the value small, get some wins.

00:03:55: But here's the step people often miss.

00:03:58: You must redesign the actual workflow around the new AI-powered process.

00:04:03: So it's not just adding a tool, it's changing how work gets done.

00:04:07: Exactly.

00:04:08: If people just keep doing things the old way, but now there's an AI tool sitting there, the investment just stalls.

00:04:13: That workflow redesign is non-negotiable for getting out of that eighty percent failure trap.

00:04:18: That blend of tech, data, process, and people feels like the core formula.

00:04:25: And we're seeing this reflected in where the money is going to, right?

00:04:28: The investment landscape is shifting.

00:04:30: Yes, Sarah Martin's gums pointed this out.

00:04:31: There's a definite pivot happening.

00:04:33: Capital seems to be moving away from those really high, maybe speculative valuations.

00:04:38: Right, the hype-driven stuff.

00:04:39: And focusing much more on, let's call it infrastructure-backed investments, things that generate predictable cash flow.

00:04:46: Investors are asking, who actually earns money every single time an AI model runs?

00:04:51: It's almost a return to, like, industrial thinking.

00:04:54: Predictable inputs and outputs.

00:04:56: That huge NVIDIA deal with open AI feels like a prime example, betting on the compute power itself.

00:05:02: Absolutely.

00:05:03: It's about who owns the picks and shovels in this gold rush.

00:05:06: It signals a maturing market focusing on tangible, reliable value streams rather than just, you know, the next cool app idea.

00:05:12: That maturity is showing up in the tech itself too, pushing beyond current limitations.

00:05:17: Andy Howe was talking about Neurosymbolic AI, NSAI.

00:05:21: Sounds intriguing.

00:05:22: It

00:05:22: is quite fascinating.

00:05:23: It's basically trying to combine the best of both AI worlds.

00:05:27: You have traditional neural networks.

00:05:29: They're great at pattern recognition, fast decisions, the score, as he put it.

00:05:33: But they can be black boxes.

00:05:35: Right.

00:05:36: Hard to know why they decided something.

00:05:37: Exactly.

00:05:37: So NSAI pairs that with symbolic reasoning.

00:05:41: Think of this as the coach.

00:05:42: It uses explicit rules, logic, it's more strategic, and importantly, auditable.

00:05:49: So you get speed and transparency.

00:05:50: That's the goal.

00:05:51: Smarter, more reliable, safer systems.

00:05:54: Especially important for complex enterprise uses where you need to trust the output.

00:05:59: And systems like that could completely change what consultants deliver.

00:06:02: Hugo Reimakers had a great term for this, agentic consulting.

00:06:06: Yeah,

00:06:06: that concept really resonated.

00:06:07: The idea is the deliverable shifts.

00:06:09: It's not just the PowerPoint deck anymore.

00:06:11: Thank goodness.

00:06:12: Right.

00:06:13: Instead, it's moving towards delivering autonomous systems, agents designed for specific tasks, discovery, synthesis, scenario modeling, agents that actually do things for the client and improve over time.

00:06:25: So the consultant becomes more of an architect, a conductor of these agents.

00:06:28: Precisely, supervising them, refining them, it moves the consultant's value way up the chain, away from just data crunching and slide building.

00:06:36: That's

00:06:36: a huge shift.

00:06:37: And speaking of making AI more practical, there was also progress mentioned on tackling the massive cost of training models.

00:06:45: Pascal Bees highlighted something called the training free GRPO method.

00:06:49: Yeah, this came from the U-two agent team.

00:06:51: It's a clever approach.

00:06:52: Instead of that super expensive fine tuning process, which can run tens of thousands of dollars just to tweak parameters, they built this dynamic library.

00:07:01: Think of it as curated experiential knowledge.

00:07:04: And they inject this relevant knowledge directly into the prompts given to the AI model.

00:07:10: So, like, giving the AI really good cheat sheets instead of making it relearn everything.

00:07:14: That's a great analogy.

00:07:15: And the results were pretty stunning.

00:07:17: They got high accuracy on complex math problems for, like, eighteen dollars in API calls.

00:07:22: Compare that to ten thousand dollars for traditional tuning.

00:07:26: Eighteen dollars versus ten thousand.

00:07:27: That changes the economics completely.

00:07:29: It really does.

00:07:30: It potentially opens up scaling specialized AI for organizations that just couldn't stomach those huge training costs before.

00:07:37: Incredible innovation on the cost side.

00:07:40: Okay, so we've covered the tech scaling, the operational side, but all this power, these autonomous agents, it naturally leaves to theme two, governance and responsible AI.

00:07:52: This feels like where leadership really gets put to the test.

00:07:55: Absolutely, because once you unleash these powerful tools, leadership can't just look the other way.

00:08:00: Corolla Schonkis and Matias Bussart from KPMG, they were crystal clear on this.

00:08:05: What was their main point?

00:08:06: That effective AI governance is squarely the responsibility of the board and senior management, not IT, not some compliance team buried somewhere.

00:08:13: Leadership owns it.

00:08:15: And crucially, you need to sort out the governance, the ownership, the rules before these tools are widely deployed, not after something goes wrong.

00:08:23: And if you don't have that clarity, you get that shadow AI problem they mentioned.

00:08:27: People just using tools under the radar.

00:08:29: Exactly.

00:08:30: Which is a huge risk.

00:08:32: Proactively defining the ethical boundaries, the usage policies.

00:08:37: That's how you turn AI risk from a threat into something you manage strategically.

00:08:42: It becomes an advantage if you handle it well.

00:08:44: Vanessa Gaskin from EY added some structure to this, outlining key pillars for responsible AI.

00:08:50: What were those?

00:08:51: Yeah, she had three main points.

00:08:53: First, governance can't be a separate thing tacked on the side.

00:08:56: It needs to be woven into your existing structures.

00:08:58: Makes sense, right?

00:08:59: Yeah,

00:08:59: integrated.

00:08:59: Second, organizations have got to close the AI fluency gap.

00:09:03: And that's not just for the tech teams.

00:09:05: It's for everyone.

00:09:06: Staff and executives need to understand this stuff.

00:09:09: That AI transformation has to be driven by the business strategy.

00:09:12: It's a core business imperative, not just an IT project or a compliance checkbox exercise.

00:09:19: And the level of oversight needed seems to be getting incredibly specific.

00:09:23: Evan Benjamin was talking about boards needing a ninety day plan just for AI agent risk.

00:09:28: Yes.

00:09:29: And a key part of his suggestion was building specific agent inventories, not just your usual list of software or hardware, knowing exactly which AI agents are operating where and then tracking key risk indicators, KRIs.

00:09:43: Like, how many times did your system block a malicious prompt?

00:09:47: That level of granularity is way beyond traditional IT risk management.

00:09:51: It implies a different kind of threat.

00:09:53: And those threats are real.

00:09:55: Gina Primo from Deloitte put a number on the fraud risk.

00:09:58: A

00:09:59: huge number.

00:10:00: She warned that these sophisticated, agentic tools could potentially drive something like forty billion dollars in fraud losses in the U.S.

00:10:06: alone by twenty twenty seven.

00:10:08: Forty billion, just...

00:10:09: Wow.

00:10:10: That number should make every board sit up straight.

00:10:12: Her point was, companies need to operate with a mindset that fraud will happen.

00:10:16: It's not an if, it's a win.

00:10:18: So you have to build monitoring, detecting right into the systems from day one.

00:10:21: You can't just bolt it on later.

00:10:23: It really feels like many of these governance issues circle back to, well, the people and the culture.

00:10:28: Mark Byershooter argued that AI maturity isn't stalling because of the tech isn't good enough.

00:10:33: Right.

00:10:33: He said it's stalling because of organizational drift.

00:10:36: Leaders are over-complicating things.

00:10:38: There's a lack of clarity.

00:10:39: So what's his solution?

00:10:40: Leadership

00:10:41: discipline, simplify, clarify the roadmap, and his big emphasis, rebuild governance and people first.

00:10:48: get the human organization ready, then focus on the platforms.

00:10:51: The tech only works if the organization can actually absolve it.

00:10:54: That makes perfect sense.

00:10:55: So given all this complexity, the black box nature of some AI, how do you actually build trust in what these systems produce?

00:11:03: Luca Rampini offered a framework for that.

00:11:05: Yeah, he suggested a three-layer approach.

00:11:08: First layer, the people.

00:11:10: up seal them, critical thinking, understanding bias, spotting potential AI hallucinations.

00:11:16: If your team can't evaluate the output critically, you're vulnerable.

00:11:19: Okay, people first.

00:11:20: What's layer two?

00:11:21: Layer two is the process.

00:11:22: Standardize it.

00:11:23: Have clear criteria for evaluating AI output.

00:11:26: He mentioned groundness as it based on facts and relevance.

00:11:29: Does it actually apply to the situation?

00:11:31: Big

00:11:31: sense.

00:11:31: Check the facts, check the fit and the third layer.

00:11:34: That's the technical grounding.

00:11:36: Use techniques like RRAG Retrieval Augmented Generation to connect the AI to reliable, authoritative data sources.

00:11:44: And crucially, demand citations.

00:11:46: If the AI can't tell you where it got the information, you shouldn't blindly trust its conclusion.

00:11:52: Traceability is key.

00:11:53: Absolutely.

00:11:54: And just briefly touching on another governance point, Bernadette Westdorp noted how the role of the data protection officer, the DPO, is evolving because of all this regulatory pressure.

00:12:03: the EU AI Act, DORA, AMLA.

00:12:06: Right, it's a complex landscape.

00:12:08: So

00:12:08: DPOs are moving beyond just compliance.

00:12:10: They're becoming strategic advisors to the board, really pushing for privacy and design right from the start of any AI project.

00:12:17: It's becoming a strategic necessity, not just a legal one.

00:12:20: Okay, so that covers the huge governance piece.

00:12:23: Let's shift gears slightly for our last theme.

00:12:25: Moving beyond just AI, let's talk corporate strategy, operating models, and well, the human side needed to actually execute on all this transformation because strategy on paper is one thing.

00:12:35: And making it happen is another entirely ages.

00:12:39: Right.

00:12:39: Yeshika Kapoor had a nice, concise summary of corporate strategy.

00:12:43: It's the CEO's game plan, fundamentally.

00:12:46: And success hinges on four key things.

00:12:48: Clarity, everyone knows the goal.

00:12:50: Coherence investments or structure, everything lines up.

00:12:53: Capital disciplines, smart spending.

00:12:56: and agility ability to adapt.

00:12:58: Sounds simple, but achieving that coherence seems incredibly hard in practice.

00:13:02: Christian Gruss from BCG pointed out a common failure point here.

00:13:06: Yeah, he highlighted that critical alignment gap.

00:13:08: He said transformations often fail not because the ambition isn't there, but because the key leaders aren't aligned, specifically the CTO, the CFO, and the CSO.

00:13:18: The

00:13:18: tech, finance, and strategy chiefs.

00:13:20: Exactly.

00:13:20: If those three aren't pulling in the same direction, executing the same playbook, the whole enterprise just sort of stalls, gears grind.

00:13:27: That

00:13:27: misalignment is killer, which maybe connects to another concept that came up.

00:13:30: Organizational ambidexterity.

00:13:32: Discussed by Sayi R. and Nicholas Lealakis.

00:13:35: Yes, ambidexterity.

00:13:37: It's this idea that to really win long-term, organizations need to master two different things at once.

00:13:43: On one hand, exploitation efficiently running and optimizing the core business you have today.

00:13:48: Okay, keep the engine running smoothly.

00:13:50: Right.

00:13:51: But at the same time, you need exploration protecting and nurturing those small, agile teams that are exploring the new ideas, the future business models.

00:14:00: You have to do both.

00:14:01: And Leo Leak has mentioned governance is adapting to help this.

00:14:05: Like future committees.

00:14:06: Exactly.

00:14:07: New structures are emerging, like these future committees, whose specific job is to shield those exploratory, maybe riskier ventures from the constant pressure of hitting quarterly numbers from the core business.

00:14:18: It's about creating space for innovation.

00:14:20: That organizational complexity really changes the game for consultants too, doesn't it?

00:14:24: Dr.

00:14:25: C.D.S.

00:14:25: Cunning made a fascinating point about top-tier firms.

00:14:28: He really did.

00:14:29: He argued that the real challenge, the core skill at places like MB, isn't just solving the business problem on paper.

00:14:36: It's navigating the incredibly complex people problems within the client organization.

00:14:41: So internal politics, hidden agendas, resistance...

00:14:44: All of it.

00:14:45: He called them the invisible skills, reading the room, influencing people when you have no formal authority, managing those tricky human dynamics.

00:14:53: That's often where the real value lies, more than just the analysis itself.

00:14:57: The consultant as diplomat and psychologist, almost.

00:15:00: In many ways, yes.

00:15:01: And that shift impacts who these firms need to hire and train.

00:15:04: Kate Smage wrote about McKinsey being the CEO factory.

00:15:08: Right,

00:15:08: tons of alumni end up in the C-suite.

00:15:10: Yeah, historically trained in that rigorous... problem structuring, analysis, judgment.

00:15:16: But she argues with AI starting to automate a lot of that junior level analysis, the firm now needs to produce architects of transformation, people who can design and lead change, not just brilliant strategists who hand over a plan.

00:15:30: It's a different skill set.

00:15:31: Jane Frazier, the city CEO and a McKinsey alum herself kind of echoed this, didn't she?

00:15:36: She did.

00:15:37: She said McKinsey gave her courage and frameworks, which is huge.

00:15:41: But she emphasized that the years spent actually operating businesses, making decisions day to day, that was absolutely essential for reaching the top job.

00:15:49: Strategy provides the map.

00:15:51: Operations teaches you how to navigate the terrain.

00:15:54: Which brings us right back to maybe the biggest single challenge facing consulting and professional services firms now.

00:16:00: The skills gap.

00:16:02: James or dad was blunt about this.

00:16:04: Extremely blunt.

00:16:05: He basically said AI transformation is fundamentally a people and skills problem, end of story, not a technology problem.

00:16:12: And he had a statistic about initiative failure.

00:16:14: Yeah, something like ninety-five percent of AI initiatives make no measurable difference.

00:16:18: Why?

00:16:19: Because people don't actually change how they work.

00:16:21: The tech seats there, but behaviors stay the same.

00:16:24: So the firm is making progress.

00:16:25: They're focused on building new ways of working, fostering curiosity, changing the culture.

00:16:29: That's the real work.

00:16:30: It all comes down to people, doesn't it?

00:16:32: Bridging that gap between a smart strategy and actually getting it done.

00:16:36: Which leads perfectly to change management.

00:16:38: Christian Roush had a simple but powerful reminder.

00:16:42: He did.

00:16:42: His point was, people adopt what they understand.

00:16:46: Simple as that.

00:16:48: So communication is key.

00:16:50: It's more

00:16:50: than just communication.

00:16:51: It's storytelling.

00:16:52: That's how you turn a dry strategy document into something people connect with, something they commit to.

00:16:58: Good storytelling explains the why behind the change, addresses the fears, builds focus, and clarifies the what and the how.

00:17:05: Without that compelling narrative, even the smartest AI strategy just ends up well.

00:17:10: stuck on the whiteboard.

00:17:11: Falling into that eighty percent failure rate we started with, if you enjoyed this deep dive, new additions drop every two weeks.

00:17:18: Also check out our other additions covering private equity, venture capital, and M&A.

00:17:22: So the big takeaway here for you, the listener, seems clear.

00:17:25: The strategic game has shifted.

00:17:27: It's less about proving AI can do amazing things and much more about proving your organization has the structure, the governance, and frankly, the human discipline to make it work reliably at scale.

00:17:38: It's not about the algorithm wizardry anymore.

00:17:40: It's about relentless execution and organizational maturity.

00:17:43: Thanks for joining us for this deep dive and make sure to subscribe.