Latest / Best of LinkedIn: Digital Transformation & Tech / Best of LinkedIn: Digital Products & Services CW 26/ 27

Transcript

Best of LinkedIn: Digital Products & Services CW 26/ 27 · Best of LinkedIn: Digital Transformation & Tech

• 0:00
  Speaker 0: This deep dive is provided by Thomas Allgeier and Frennis, based on the most relevant LinkedIn posts about digital products and services in calendar weeks twenty-six and twenty seven.
• 0:10
  Speaker 0: Frinnis is a B-to-B market research company that supports enterprise product teams with building figure by feature competitive intelligence.
• 0:17
  Speaker 0: That shows exactly how their products stacks up against the competition.
• 0:21
  Speaker 0: you can find more info in the description.
• 0:23
  Speaker 1: so true
• 0:24
  Speaker 0: right.
• 0:24
  Speaker 0: So welcome everyone.
• 0:26
  Speaker 0: I want you to imagine just for a second, but your upgrading or factory too A fully automated like state of the art assembly line.
• 0:33
  Speaker 0: Oh
• 0:33
  Speaker 1: nice sounds expensive
• 0:35
  Speaker 0: Very.
• 0:35
  Speaker 0: And the machines, they are incredible right?
• 0:37
  Speaker 0: The conveyor belts are moving at absolute warp speed.
• 0:40
  Speaker 0: but there is one tiny problem.
• 0:42
  Speaker 0: you forgot to double check the product blueprint.
• 0:44
  Speaker 1: Oh no!
• 0:45
  Speaker 1: Yeah.
• 0:46
  Speaker 0: so your brand new highly efficient factory it doesn't just make a single mistake.
• 0:50
  Speaker 0: It efficiently mass produces thousands of completely defective parts before You even realize There's a problem on the floor.
• 0:58
  Speaker 1: That Is A Terrifying Thought.
• 0:59
  Speaker 0: It really is.
• 1:01
  Speaker 0: And I bring it up because that's exactly what the tech industry is doing right now, we are taking a stack of insights curated from top digital products and services conversations across LinkedIn over the last two weeks... ...and we're uncovering this startling reality about building software today!
• 1:18
  Speaker 1: Yeah.. The overarching theme here….
• 1:21
  Speaker 1: well its massive irony actually Because AI has made building software cheaper and faster than at any point in human history, right?
• 1:30
  Speaker 1: Absolutely.
• 1:31
  Speaker 1: Faster than ever!
• 1:31
  Speaker 0: Look this hyper acceleration it hasn't made product management easier.
• 1:35
  Speaker 0: I mean, it is actually pushing the whole discipline into this crisis of purpose.
• 1:40
  Speaker 0: Right?
• 1:40
  Speaker 0: Because the bottleneck in software development... It's no longer code generation!
• 1:44
  Speaker 0: The tech is just accelerating which means your margin for strategic error is basically shrinking to zero.
• 1:50
  Speaker 0: We are seeing a complete shift on what it takes to build a successful product.
• 1:54
  Speaker 1: Okay let us unpack that because if you were listening right now or maybe prepping for road map meeting You're probably intimately familiar with this tension.
• 2:03
  Speaker 0: Oh they definitely Right, like you probably have an executive breathing down your neck asking where the new AI features.
• 2:10
  Speaker 0: But you know that your core users are just begging for a reliable bug-free export button.
• 2:16
  Speaker 1: Just give them the buttons
• 2:17
  Speaker 0: Exactly!
• 2:18
  Speaker 0: The daily reality of anyone working in tech is changing so rapidly and it starts with this fundamental inset from Joy Momin.
• 2:26
  Speaker 0: She pointed out that AI has entirely shifted the development bottleneck.
• 2:30
  Speaker 0: Yeah,
• 2:31
  Speaker 1: I read her post.
• 2:31
  Speaker 1: it is so spot-on
• 2:33
  Speaker 0: It really is!
• 2:34
  Speaker 0: I mean writing kind of used to be slow part right?
• 2:36
  Speaker 0: The labor intensive part.
• 2:38
  Speaker 0: Now...the true bottleneck isn't typing out syntax but deciding what's actually worth building in first place.
• 2:45
  Speaker 1: And Gepis Stone takes this bottleneck shift and applies perfectly your factory analogy.
• 2:50
  Speaker 0: Oh how.
• 2:50
  Speaker 0: so
• 2:51
  Speaker 1: Well he warns that AI basically accelerates whatever you put infront.
• 2:55
  Speaker 1: So if your engineering teams can suddenly ship entire backlogs in a fraction of the time vague or poorly researched requirements Well, they become incredibly dangerous because
• 3:05
  Speaker 0: you get there faster
• 3:06
  Speaker 1: exactly.
• 3:07
  Speaker 1: You aren't just building a slightly off-target feature anymore.
• 3:10
  Speaker 1: You are producing these overbuilt irrelevant features at this terrifying velocity Pisco and makes the case that this AI accelerated engineering, it exposes weak product requirements.
• 3:23
  Speaker 0: Wow!
• 3:23
  Speaker 0: So it takes what used to look like an engineering performance problem and reveals it as an organizational design problem?
• 3:29
  Speaker 1: Exactly if you feed a bad strategy into a fast machine... You just get the wrong destination quicker.
• 3:36
  Speaker 0: It is like giving a teenager a Ferrari but no GPS Sure.
• 3:40
  Speaker 0: You are going one hundred and fifty miles per hour, but you might just be driving off a cliff faster.
• 3:44
  Speaker 1: That is exactly it.
• 3:45
  Speaker 1: And preventing that mass production of defects Is why Michael Wall argues for this strict readiness gate.
• 3:52
  Speaker 0: A readiness gate?
• 3:53
  Speaker 0: Okay
• 3:53
  Speaker 1: Yeah his perspective is This much needed reality check For the industry.
• 3:57
  Speaker 1: He basically says look Your customers do not care about your AI strategy
• 4:00
  Speaker 0: They really don't?
• 4:01
  Speaker 1: they Don't care by your tech stack
• 4:08
  Speaker 0: Right, like they do not care how many PDFs you can cram into a prompt.
• 4:12
  Speaker 1: Exactly what they actually care about are their own bottlenecks.
• 4:15
  Speaker 1: You know there's squeezed margins.
• 4:17
  Speaker 1: They're painful manual processes.
• 4:19
  Speaker 1: wall argues that if the new AI feature doesn't significantly simplify A workflow or you know eliminate a manual step it is an innovation.
• 4:28
  Speaker 0: It just bloat
• 4:29
  Speaker 1: exactly!
• 4:30
  Speaker 1: It has costly bloat disguised as innovation.
• 4:33
  Speaker 0: And we saw a really brutally honest application of that readiness gate from Elena Linova with her product, OneRank.
• 4:40
  Speaker 0: Oh
• 4:41
  Speaker 1: yeah!
• 4:41
  Speaker 1: Her story was fascinating...
• 4:42
  Speaker 0: Right she decided to run her own AI feature bets through this massive pressure test and her screen is just devastatingly simple.
• 4:50
  Speaker 0: She said if you delete the word A.I From The Feature Description what gap Is Actually Left?
• 4:55
  Speaker 1: That's such a good test.
• 4:56
  Speaker 0: It
• 4:57
  Speaker 0: IS!!
• 4:57
  Speaker 0: When did it do its road map?
• 4:59
  Speaker 0: she realized putting chatbot on top of data decision tool didn't actually make the user's decision any better.
• 5:06
  Speaker 1: Right, because it just added this conversational UI to a task that completely did not need one?
• 5:10
  Speaker 0: Exactly!
• 5:11
  Speaker 0: I mean if a users core job is analyzing data forcing them to chat with an AI just to see that data It just adds this useless layer of obfuscation...it made the screenshot look like it belonged in a twenty-twenty six pitch
• 5:24
  Speaker 1: deck.
• 5:24
  Speaker 1: Yeah, pitch deck features we've seen all the time.
• 5:26
  Speaker 0: So she killed the chatbots.
• 5:28
  Speaker 0: She killed the summarization wrappers, which let's be honest are basically just basic AI models dressed up with a nice interface because she felt they degraded user trust.
• 5:38
  Speaker 0: She only shipped AI that actively did the core job of her product better... Which
• 5:43
  Speaker 1: takes immense discipline right?
• 5:44
  Speaker 1: Especially when the whole market is just shouting at you to include AI everything and systematizing that discipline is where Suffolk-Erikholz' recap essential.
• 5:56
  Speaker 0: Okay, tell me about this framework?
• 5:58
  Speaker 1: So Corti have proposed four specific quality gates that teams absolutely must use to pressure test any AI feature before it ever hits production.
• 6:07
  Speaker 1: This framework It shifts the conversation completely from you know Do we have the technical capability to build this too?
• 6:13
  Speaker 1: do we have The business justification to ship this?
• 6:16
  Speaker 0: oh I love That.
• 6:17
  Speaker 0: let's break down those for gates because i think they are incredibly practical For you listening To apply immediately.
• 6:23
  Speaker 0: definitely so gate number one is is testing for hallucination and refusal accuracy.
• 6:28
  Speaker 0: You aren't just testing.
• 6:29
  Speaker 0: if the model gets the answer right, you have to test.
• 6:32
  Speaker 0: it's a model.
• 6:33
  Speaker 0: knows how say I don't know...
• 6:34
  Speaker 1: Which is so where?
• 6:35
  Speaker 0: Right!
• 6:36
  Speaker 0: If an AI makes up metric in a B-to-B financial dashboard Just to please user your entire software loses credibility instantly…if can't refuse a prompt that doesn't understand It fails the gate
• 6:48
  Speaker 1: Exactly.
• 6:49
  Speaker 1: And then Gate number two is the trust gap.
• 6:52
  Speaker 1: This one measures verification time.
• 6:54
  Speaker 1: Basically, can the user verify the AI's output in under ten seconds?
• 6:59
  Speaker 0: Ten seconds.
• 6:59
  Speaker 0: Wow!
• 7:00
  Speaker 0: That is
• 7:00
  Speaker 1: stripped... It is.
• 7:01
  Speaker 1: but think about it.
• 7:02
  Speaker 1: if a user asks the AI to summarize a legal document But they have to spend thirty seconds carefully reading this summary against The original text you know.
• 7:10
  Speaker 1: just make sure I didn't miss a crucial clause.
• 7:12
  Speaker 1: Well They could've just skimmed the document themselves In twenty seconds.
• 7:15
  Speaker 1: If verification takes longer than manual execution trust Just never builds and featured option will completely flatline.
• 7:22
  Speaker 0: that makes total sense.
• 7:24
  Speaker 0: Okay, gate number three is the cost of wrong.
• 7:26
  Speaker 0: You have to size the risk per thousand uses.
• 7:29
  Speaker 1: Yeah!
• 7:30
  Speaker 1: Context matters.
• 7:31
  Speaker 0: Exactly.
• 7:32
  Speaker 0: If an AI gives a bad recommendation for coffee shop on travel app That's minor annoyance.
• 7:38
  Speaker 0: But if BtoB AI miscalculates compliance metric For healthcare provider Oh boy
• 7:44
  Speaker 1: Massive lawsuit.
• 7:45
  Speaker 0: The inherent stakes dictate whether features should even exist.
• 7:49
  Speaker 0: And finally, gate number four is the adoption wall.
• 7:52
  Speaker 0: Will the most experienced veteran person on your team actually use it?
• 7:56
  Speaker 1: The power user test!
• 7:57
  Speaker 0: Right.
• 7:58
  Speaker 0: Kuro Tieva targets a thirty-day retention rate above forty percent.
• 8:01
  Speaker 0: If you're power users won't touch the AI feature... ...the wider rollout is just guaranteed to stall.
• 8:06
  Speaker 1: What's fascinating here Is how these four gates completely redefine the concept of readiness.
• 8:11
  Speaker 1: They aren't about the software being code complete Meaning bug free and functional value complete.
• 8:19
  Speaker 1: Value
• 8:19
  Speaker 0: complete, I like that!
• 8:20
  Speaker 1: Yeah and if you apply these strict value gates to a typical enterprise roadmap today... You're going to kill off a massive percentage of AI ideas Very, very quickly.
• 8:31
  Speaker 0: Well here's where it gets really interesting.
• 8:33
  Speaker 0: if these frameworks are ruthlessly filtering out mediocre AI concepts right before launch It puts an immense amount of pressure on the very beginning of the product life cycle.
• 8:43
  Speaker 0: You have to figure out what the customer actually needs Before you write a single prompt which leads us to our next major theme Which is product discovery returning to its roots in deep customer value
• 8:55
  Speaker 1: and the financial cost of getting that discovery wrong.
• 8:57
  Speaker 1: I mean, it is steeper than ever right now.
• 8:59
  Speaker 1: Joe Caprara shared this incredibly painful but highly instructive story about exactly this.
• 9:04
  Speaker 0: What happened?
• 9:05
  Speaker 1: So he spent nine months building an AI product for Product Feedback Discovery.
• 9:08
  Speaker 1: It was called Epic And by all traditional launch metrics...it was a total hit.
• 9:13
  Speaker 0: Okay sounds good so far
• 9:14
  Speaker 1: Right.
• 9:14
  Speaker 1: The beta users loved the demo They were lining up to use at.
• 9:17
  Speaker 1: the UI was beautiful But after week one Usage just dropped off in absolute cliff.
• 9:23
  Speaker 0: Oh wow Nine months of engineering time, dead almost overnight.
• 9:27
  Speaker 1: Dead over night he had to pull the plug.
• 9:28
  Speaker 1: because you realize that once The magic of talking to an AI wore off... ...the product actually created friction instead of automation.
• 9:36
  Speaker 1: How
• 9:36
  Speaker 0: so?
• 9:37
  Speaker 1: Well, Capra realized that the AI was essentially asking users To become prompt engineers just to log a simple piece Of feedback Instead of saving them Time.. ..the tool required users to learn A whole new workflow of interacting with a chatbot
• 9:52
  Speaker 0: Right, which created significantly more friction than just filling out a standard boring web form.
• 9:57
  Speaker 1: Exactly!
• 9:58
  Speaker 1: To understand the mechanism behind that failure we can actually look at Janet Curham's framework for product discovery.
• 10:04
  Speaker 1: He broke user input down into three distinct categories.
• 10:07
  Speaker 1: Okay
• 10:07
  Speaker 0: what are they?
• 10:08
  Speaker 1: Wants needs and hunches.
• 10:10
  Speaker 1: Capra's failure with Epic is this textbook case of building-for-the-wants.
• 10:15
  Speaker 1: he listened to what users explicitly said They wanted in those early exciting demos
• 10:19
  Speaker 0: But he missed The Needs.
• 10:21
  Speaker 1: Exactly, he missed the needs which you only uncover by observing what users actually do.
• 10:26
  Speaker 1: You know where they click and Where they experience real workflow friction in their daily behavior.
• 10:32
  Speaker 1: But
• 10:32
  Speaker 0: wait hold on.
• 10:33
  Speaker 0: if we only rely on documented proven needs like If we only build things that fix existing workflows doesn't that fundamentally kill innovation?
• 10:42
  Speaker 1: That is a great question.
• 10:43
  Speaker 0: I mean.
• 10:43
  Speaker 0: What about intuition?
• 10:45
  Speaker 0: don't we need leaps of faith to build something truly disruptive?
• 10:48
  Speaker 0: If we just watch what people do today, we'd build faster horses.
• 10:52
  Speaker 1: That is a crucial pushback and it's exactly why Janakirim includes that third category which is hunches.
• 11:00
  Speaker 1: Bring through innovation absolutely starts with intuition.
• 11:03
  Speaker 1: The key differentiator is that a hunch requires rapid validation.
• 11:06
  Speaker 1: You don't spend nine months building a hunch in secret with the full engineering team.
• 11:10
  Speaker 1: Right,
• 11:10
  Speaker 0: you validate it in days
• 11:11
  Speaker 1: Exactly and Scott Persinger shared a brilliant example of this working perfectly.
• 11:15
  Speaker 1: So he is the CTO at biztrip.ai And one of his engineers built a feature that lets users simply drag-and drop a screenshot into an AI agent... ...And instantly starts planning a business trip.
• 11:26
  Speaker 0: Oh!
• 11:27
  Speaker 0: That's so cool.
• 11:27
  Speaker 1: It is highly successful.
• 11:29
  Speaker 1: Users absolutely love it.
• 11:31
  Speaker 1: But as Persinger noted This feature was never on a formal product roadmap Never!
• 11:37
  Speaker 1: The engineer built it purely on a hunch, using their own creativity and they tested immediately.
• 11:43
  Speaker 0: Okay so the goal is to make space for that hunch but aggressively test against reality?
• 11:48
  Speaker 0: Yes
• 11:49
  Speaker 1: Exactly.
• 11:50
  Speaker 0: And to balance that out, Dev Shah shared a highly structured discovery approach called the Problem Stack Canvas.
• 11:56
  Speaker 0: It's basically this two-hour exercise designed to prevent exactly The kind of nine month reworks we saw with
• 12:02
  Speaker 1: Caprara.
• 12:02
  Speaker 1: I love a good canvas.
• 12:03
  Speaker 0: Right it is five layer framework but the absolute most crucial layer for you To apply is Layer Four.
• 12:10
  Speaker 0: and Layer four asks What is the cost of not solving this problem?
• 12:14
  Speaker 1: Oh, that's so powerful.
• 12:15
  Speaker 0: It really is.
• 12:16
  Speaker 0: you have to force the customer to measure that costs in hard money waste-to time or severe risk.
• 12:21
  Speaker 0: Shaw argues That if there is no measurable cost to the user for leaving The problem unsolved then it's Not a real problem worth dedicating your engineering resources too.
• 12:29
  Speaker 1: and that concept Of the cost If not solving it it ties directly into being woos product philosophy.
• 12:37
  Speaker 1: what does
• 12:38
  Speaker 0: who say?
• 12:39
  Speaker 1: Woo suggests that every single software product you sell is actually two products simultaneously.
• 12:44
  Speaker 1: First, do have the deal product?
• 12:45
  Speaker 0: The Deal Product.
• 12:46
  Speaker 0: Yeah
• 12:46
  Speaker 1: this Is the version of the software that demos beautifully on a zoom call.
• 12:50
  Speaker 1: You know it has the flashy AI summarization features.
• 12:53
  Speaker 1: It gets the executives all excited and it closes the initial sale.
• 12:56
  Speaker 0: The sizzle
• 12:57
  Speaker 1: exactly but underneath That you have the renewal product.
• 13:01
  Speaker 1: This is the product that has to run flawlessly at two in the morning on a Sunday when nobody from your customer's success team is around to help the user.
• 13:09
  Speaker 0: Right, so The Deal Product gets the signature on the contract but The Renewal Product is what stops them from churning eleven months later.
• 13:16
  Speaker 1: Exactly and Woo notes too.
• 13:18
  Speaker 1: many of these new AI rappers are purely deal products.
• 13:22
  Speaker 1: They look incredible during sales pitch But they fail the Two AM renewal test because aren't reliable They hallucinate or they don't actually solve that core layer for problem.
• 13:32
  Speaker 0: Yeah, you can win a deal once with hype but You have to re-earn the renewal every single week With utility yep
• 13:38
  Speaker 1: hundred percent.
• 13:39
  Speaker 0: so if The deal product versus renewal product framework forces us To have this like flawless customer discovery before we write a single line of AI code, it begs the question who is actually responsible for that discovery now?
• 13:53
  Speaker 1: Right.
• 13:54
  Speaker 1: Because their roles are changing
• 13:55
  Speaker 0: Exactly because people traditionally doing this job they're disappearing.
• 13:59
  Speaker 0: Just look at what Claire Vo just shared about Gusto.
• 14:01
  Speaker 1: Oh!
• 14:01
  Speaker 1: This story blew my mind.
• 14:03
  Speaker 1: It's
• 14:03
  Speaker 0: crazy.
• 14:04
  Speaker 0: Duster recently shipped an entirely new first version in less than ten weeks.
• 14:09
  Speaker 0: but timeline isn't the shocker.
• 14:12
  Speaker 0: its team composition They built it with absolutely zero product managers.
• 14:16
  Speaker 0: Zero?
• 14:17
  Speaker 0: None, It was just a CTO, designer three engineers and Claude code.
• 14:21
  Speaker 1: And when you look at how they did that, Claude essentially acted as an autonomous junior developer Right!
• 14:27
  Speaker 1: He's handling all the boilerplate code generation running basic QA scripts even compiling technical specs based on the CTOs direct vision.
• 14:35
  Speaker 1: The engineers weren't waiting around for PM to write a Jira ticket You know...they were using AI bridge gap between design and deployment, basically instantly.
• 14:44
  Speaker 0: So I have to ask you if engineers can use AI assistance to build features at warp speed?
• 14:48
  Speaker 0: And AI can pull insights directly from user feedback channels.
• 14:52
  Speaker 0: is the product manager role becoming obsolete?
• 14:55
  Speaker 0: like are we witnessing The end of the PM was
• 14:58
  Speaker 1: fascinating here as how Akash Tawari reframes this whole industry panic.
• 15:02
  Speaker 1: he calls the idea Of magic AI running on pure autopilot a dangerous hallucination
• 15:08
  Speaker 0: A dangerous solution, I like that phrasing.
• 15:11
  Speaker 0: Yeah!
• 15:11
  Speaker 1: Yes you can feed an AI-generated requirement doc into autonomous development build but without strict human oversight.
• 15:18
  Speaker 1: You aren't accelerating your product lifecycle... ...you are just fast tracking a catastrophic production outage
• 15:24
  Speaker 0: Right going back to the factory without the blueprint.
• 15:26
  Speaker 1: Exactly.
• 15:27
  Speaker 1: Tuari points out That the PM role isn't dying.
• 15:30
  Speaker 1: It is shifting.
• 15:31
  Speaker 1: The shift moving the PM from being creator To governor
• 15:35
  Speaker 0: From Creator to Governor.
• 15:37
  Speaker 0: That is a massive identity shift for the profession.
• 15:41
  Speaker 0: It means PMs are no longer spending their days just writing endless product requirement documents or curating JIRA backlogs.
• 15:48
  Speaker 0: Instead they are designing the governance systems.
• 15:51
  Speaker 0: Yes, exactly.
• 15:52
  Speaker 0: They're the ones ensuring regulatory compliance Maintaining the overall architecture integrity and keeping the engineering team aligned with the actual business strategy.
• 16:01
  Speaker 0: They are managing the AI output not competing with it for tasks
• 16:05
  Speaker 1: because AI doesn't replace The core product lifecycle right?
• 16:09
  Speaker 1: It just changes where the cognitive load sits.
• 16:12
  Speaker 1: Where does that sit now?
• 16:13
  Speaker 1: The cognitive load moves from generating output like writing specs or code, to making highly complex decisions.
• 16:19
  Speaker 1: Melange Makisha's interview strategy illustrates this shist
• 16:22
  Speaker 0: perfectly.
• 16:23
  Speaker 0: Oh the interview question?
• 16:24
  Speaker 0: Yeah
• 16:25
  Speaker 1: when she interviews product management candidates today She doesn't ask them about AI prompt engineering Or agile frameworks...she asks one simple incredibly revealing Question which is What product decision did you regret most?
• 16:38
  Speaker 0: That was such a difficult interview question.
• 16:42
  Speaker 0: It forces the candidate to talk entirely about trade-offs and consequences.
• 16:46
  Speaker 0: As Makayja points out, AI can help almost anyone generate a roadmap.
• 16:51
  Speaker 0: you know it can summarize fifty hours of user research in seconds And it can draft beautiful spec documents.
• 16:57
  Speaker 0: What AI absolutely cannot do is replace human reasoning required for difficult ambiguous decision where there's just no perfect answer.
• 17:07
  Speaker 0: Judgment is the ultimate PM's skill now.
• 17:09
  Speaker 1: It really does!
• 17:10
  Speaker 1: We're
• 17:10
  Speaker 0: even seeing platforms being built specifically to support this.
• 17:14
  Speaker 0: Look at Skilvana, created by Mitharavi Prakash.
• 17:17
  Speaker 0: it acts as an AI co-founder but its whole purpose isn't generate code.
• 17:21
  Speaker 0: Its designed pressure test a product manager judgment.
• 17:24
  Speaker 1: I think that is brilliant.
• 17:25
  Speaker 1: Yeah,
• 17:26
  Speaker 0: actively challenges assumptions and maps out trade off so a PM can defend their strategic decisions to leadership.
• 17:32
  Speaker 0: And
• 17:32
  Speaker 1: if we connect us with bigger picture The need for that defensible judgment explains so much of the current dysfunction in tech industry right now.
• 17:40
  Speaker 1: Melissa Perry shared some truly alarming survey data regarding organization alignment.
• 17:44
  Speaker 0: What did they say?
• 17:45
  Speaker 1: So, sixty-two percent product managers reported a lack.
• 17:49
  Speaker 1: clear AI strategy is their biggest daily challenge.
• 17:52
  Speaker 1: However when you ask C level executives the exact same question only nineteen percent feel this way
• 17:58
  Speaker 0: Wow!
• 17:59
  Speaker 0: That is a forty-three percent gap between the executives funding the product and people actually building it.
• 18:04
  Speaker 0: Yes,
• 18:05
  Speaker 1: The strategy usually exists at top right?
• 18:07
  Speaker 1: The executives in board.
• 18:09
  Speaker 1: they are having these high level conversations about AI investment market positioning but that strategy has never translated into daily decision making frameworks.
• 18:17
  Speaker 1: PM needs on Monday morning.
• 18:20
  Speaker 0: They're just left guessing
• 18:21
  Speaker 1: Exactly!
• 18:22
  Speaker 1: The PMS don't know rules of engagement.
• 18:24
  Speaker 1: What data should we feed the AI?
• 18:25
  Speaker 1: what user workflows should be?
• 18:27
  Speaker 1: leave alone The translation layer between executive vision and engineering execution is completely missing.
• 18:33
  Speaker 0: Which brings us to the new identity of product leadership.
• 18:36
  Speaker 0: Scott Lowry uses a brilliant analogy to describe how this gap must be closed.
• 18:41
  Speaker 0: He says that the modern VP of Product or Chief Product Officer shouldn't act like a feature curator who just manages A list of customer requests.
• 18:48
  Speaker 1: right no more suggestion boxes
• 18:50
  Speaker 0: Exactly!
• 18:51
  Speaker 0: They need to act like a private financial planner managing high net worth capital portfolio.
• 18:56
  Speaker 0: In a SaaS business, your R&D budget is your investment capital—your products are your portfolio holdings.
• 19:03
  Speaker 1: That is a great way to look at it.
• 19:04
  Speaker 0: Right, the product leader has to sit across from the CEO and balance the portfolio you know.
• 19:09
  Speaker 0: identifying high-risk AI growth bets maintaining stable core features that drive renewals having the ruthless discipline pinpoint legacy features they need to exit entirely.
• 19:22
  Speaker 0: It's pure capital allocation.
• 19:24
  Speaker 1: You cannot allocate capital effectively if you are drowning in tactical feature requests.
• 19:28
  Speaker 1: The AI era requires product leaders to step up and own the strategy, design the governance.
• 19:33
  Speaker 1: And basically take absolute responsibility for the ultimate business
• 19:36
  Speaker 1: outcomes.".
• 19:37
  Speaker 0: So what does this all mean?
• 19:38
  Speaker 0: If you take away one thing from This Deep Dive it's that AI is not a magic shortcut That allows you to skip the hard messy work of understanding your customer.
• 19:46
  Speaker 1: Definitely Not.
• 19:47
  Speaker 0: It Is actually a magnifying glass!
• 19:49
  Speaker 0: It exposes weak strategy...it highlights vague requirements..and it brutally punishes poor organizational design by building the wrong things faster.
• 19:58
  Speaker 0: But for the product teams that adapt, AI rewards deep customer understanding strict value gates and impeccable human judgment.
• 20:06
  Speaker 1: This actually raises an important question for you to consider as we wrap up today.
• 20:10
  Speaker 1: We've spent decades judging this success of technology by its ROI return on investment.
• 20:16
  Speaker 1: Did this software save us money or make us faster?
• 20:18
  Speaker 0: Right, the classic metric.
• 20:20
  Speaker 1: But as we move into an era where AI doesn't just automate data entry but actively recommends actions, coordinates agents and shapes organizational strategy well Vanya Sahi suggests that perhaps ROI is dead.
• 20:31
  Speaker 0: ROI's Dead.
• 20:32
  Speaker 0: wow!
• 20:33
  Speaker 1: The next massive competitive advantage for your company won't come from building or buying a slightly larger language model.
• 20:39
  Speaker 1: it will come from building better decisions.
• 20:41
  Speaker 1: The true metric you should be evaluating moving forward is ROD, Return on Decisions.
• 20:46
  Speaker 1: Exactly!
• 20:47
  Speaker 1: Does the AI you are implementing consistently and measurably improve quality of human judgement in your organization?
• 20:56
  Speaker 0: That's a completely different way to look at road map tomorrow morning... Let's go back to that image of the upgraded factory we started with.
• 21:04
  Speaker 0: The AI gives you the state-of-the art machines and the unprecedented speed, but your product strategy?
• 21:11
  Speaker 0: Your strict value gates in human judgment?
• 21:13
  Speaker 0: well... That is the blueprint.
• 21:14
  Speaker 1: Without it, you are just mass producing defects?
• 21:17
  Speaker 1: Exactly!
• 21:18
  Speaker 0: But with it... You can navigate this new landscape faster and more effectively than ever before.
• 21:23
  Speaker 0: If you enjoyed This Deep Dive New deep dives drop every two weeks.
• 21:27
  Speaker 0: Also check out our other editions on ICT & Tech Artificial Intelligence Cloud Sustainability in Green ICT Defense Tech And HealthTech.
• 21:36
  Speaker 1: Thank-you so much for listening.
• 21:37
  Speaker 0: thank you So Much For Joining Us Learning With us and sharing your time.
• 21:41
  Speaker 0: Don't forget to hit subscribe so never miss an insight.
• 21:43
  Speaker 0: Keep building, keep questioning and we'll see you next time.