Latest / How I Work / The Work Edit: The AI skill that saves me 200 minutes a week. Here's how to build yours.
Transcript
- 0:01Speaker 1: Do you open an AI chat, type a request, get
- 0:04Speaker 1: an answer, argue with it a bit, close the tab
- 0:07Speaker 1: and then do it all again tomorrow? AI is supposed
- 0:12Speaker 1: to give us our time back, yet so much of
- 0:16Speaker 1: it goes on back and forth, doing more and more
- 0:19Speaker 1: rework and arguing with the AI to get a half-baked result.
- 0:24Speaker 2: In this work edit, I.
- 0:26Speaker 1: Sit down with Stephen, an organizational transformation advisor with more
- 0:30Speaker 1: than 20 years of experience. He uses Claude every day,
- 0:35Speaker 1: but he wanted to move from spaghetti style AI use
- 0:38Speaker 1: to a deliberate setup without signing up for 12 different tools.
- 0:44Speaker 2: In this chat, I share why.
- 0:46Speaker 1: Very specific projects beat one big one, how to catch
- 0:50Speaker 1: the AI towels that make readers switch off, and the
- 0:53Speaker 1: tools that I rely on every day. Then we get
- 0:57Speaker 1: to the bigger shift, a one-week exercise that shows you
- 1:00Speaker 1: which tasks AI should be taking off your plate for good.
- 1:05Speaker 1: If you use AI daily but suspect that there's more
- 1:08Speaker 1: value you could be getting, then this episode is for you.
- 1:19Speaker 3: Welcome to How I Work, a show about habits, rituals,
- 1:23Speaker 3: and strategies for optimizing your day.
- 1:26Speaker 2: I'm your host, Dr. Amantha Imber.
- 1:32Speaker 1: Stephen, let's start with what's the challenge that we're working
- 1:35Speaker 1: on today?
- 1:36Speaker 4: The challenge is.
- 1:37Speaker 5: How I move from being an avid user of generative
- 1:42Speaker 5: AI for my solopreneur business, which is advisory, support leadership, coaching,
- 1:48Speaker 5: and otherwise to get more deliberate and architect and where
- 1:53Speaker 5: do I use AI capability, especially around automation, agent, or
- 1:58Speaker 5: like partnering with me on particular things where I typically
- 2:01Speaker 5: default to a human, which is not in and of
- 2:04Speaker 5: itself the wrong thing to do, but like, how do I, yeah,
- 2:07Speaker 5: that's the challenge. Like, how do I start making decisions intentionally?
- 2:12Speaker 5: That's one part B is, and doing that in a
- 2:14Speaker 5: way where I'm not signing up for 12 different tools
- 2:18Speaker 5: trying to get them all to talk to each other,
- 2:19Speaker 5: knowing which ones I use when. So just what can
- 2:21Speaker 5: that architecture begin to look like, where to start and
- 2:24Speaker 5: how do I know, how do I fine tune that?
- 2:26Speaker 4: Does that make sense?
- 2:27Speaker 2: It does.
- 2:27Speaker 1: And give me a bit of context as to your
- 2:30Speaker 1: background because you've recently had quite a big change in
- 2:33Speaker 1: what your work looks like.
- 2:35Speaker 5: The quick version is I've spent 20 plus years really
- 2:38Speaker 5: in organizational transformation, change management, leadership development, and some strategy
- 2:44Speaker 5: and innovation works, kind of those have been the four
- 2:47Speaker 5: quadrants working for large corporations on the inside. And where
- 2:50Speaker 5: I find myself now is I'm no longer in the
- 2:53Speaker 5: big corporate and I'm now helping larger corporates as a,
- 2:56Speaker 5: you know, almost a solo or small collective of advisors
- 3:01Speaker 5: doing all transformation. We're really helping them a lot around
- 3:03Speaker 5: executive leadership alignment in the service of either accelerated digital
- 3:09Speaker 5: transformation or some of my clients are thinking through, how
- 3:13Speaker 5: do we make more of a platform out of what
- 3:15Speaker 5: has been traditionally been products and service org? And so
- 3:18Speaker 5: a lot of the work is around the behavior or
- 3:20Speaker 5: bringing leaders along. That's typically the big meta theme of
- 3:25Speaker 5: kind of what I'm doing. So most of that doing solo,
- 3:27Speaker 5: and then I bring in partners to help me with
- 3:30Speaker 5: big deliverables or big facilitation.
- 3:32Speaker 2: And tell me, how are you using GenAI at the moment?
- 3:35Speaker 1: Like what are some of the typical tasks or use
- 3:39Speaker 1: cases where.
- 3:40Speaker 2: You've been finding it helpful?
- 3:42Speaker 1: And then we're going to dig into ideas on where
- 3:45Speaker 1: you could be using it more and getting better gains
- 3:47Speaker 1: with it.
- 3:48Speaker 5: The day-to-day communication and gut checking my comms, especially email comms.
- 3:54Speaker 5: On the client delivery or client research piece, I'm using
- 3:58Speaker 5: Gen AI to do research. So asking it things about
- 4:01Speaker 5: a particular industry or a particular context, or this is
- 4:04Speaker 5: what's going on with my client. It's almost being a
- 4:06Speaker 5: thought partner in the research piece. The other thing on
- 4:09Speaker 5: client delivery is if I've done a series of interviews
- 4:13Speaker 5: and I'm using transcribers, I say, hey, can I use
- 4:16Speaker 5: my recorder when we're having a conversation or a dialogue
- 4:19Speaker 5: around something, getting the transcript back and then using Gen
- 4:23Speaker 5: AI to interpret, synthesize, going back and forth.
- 4:26Speaker 4: This is what I heard. Here are my notes.
- 4:29Speaker 5: So just using it as another kind of the thing
- 4:31Speaker 5: that can crunch through everything and my brain maybe forgotten something.
- 4:35Speaker 5: So just using it to synthesize, analyze.
- 4:37Speaker 4: And then.
- 4:38Speaker 5: usually start me off on a deliverable. Oftentimes, it's either
- 4:42Speaker 5: a PowerPoint deck to help me get the shell of this,
- 4:45Speaker 5: or this is what I'm thinking, and what else should
- 4:47Speaker 5: I include? And more recently, I've found with Claude, can
- 4:51Speaker 5: actually build out a decent first pass, and then I
- 4:54Speaker 5: take it and put it into my format or reword
- 4:57Speaker 5: it so that, you know, because Claude has a style
- 5:00Speaker 5: in which it does it, and it's like, put it
- 5:01Speaker 5: in my style. And then written, sort of like Word
- 5:04Speaker 5: document style memo, hey, here's a three-pager on findings or
- 5:10Speaker 5: what you should look at next, or here's your thought
- 5:12Speaker 5: prompt for the town hall. You're going to do, I'm
- 5:15Speaker 5: going to help you prepare it. And so those types
- 5:17Speaker 5: of outputs. And then on the backend, even in just
- 5:20Speaker 5: some of the business operational things are asking it questions around.
- 5:24Speaker 4: I'm an LLC in the U S I've got this
- 5:26Speaker 4: set up.
- 5:27Speaker 5: And before I go and speak to my, either my
- 5:28Speaker 5: CPA or tax person, I know you're not a tax person,
- 5:31Speaker 5: but like, what are the five things I can kind
- 5:33Speaker 5: of discern in media? So just a lot of back office.
- 5:36Speaker 5: just admin type stuff not legal but just i'm using
- 5:40Speaker 5: quickbooks for invoicing what are the other three tools out
- 5:43Speaker 5: there i wanted to be able to do this this
- 5:44Speaker 5: and this so just kind of learning without going out
- 5:47Speaker 5: and googling and that's why i feel like i'm just
- 5:50Speaker 5: kind of doing it in a sporadic kind of spaghetti
- 5:53Speaker 5: kind of manner and i haven't got an ai underlying
- 5:56Speaker 5: architecture but those are some of the common things that
- 5:59Speaker 5: one i didn't mention is around sort of thought leadership
- 6:02Speaker 5: and thinking about what content I want to develop, whether
- 6:05Speaker 5: that's for posting, whether that's just for my own repertoire
- 6:07Speaker 5: of reports, just using it as a thought partner. Honestly,
- 6:11Speaker 5: I could be doing this in a very forward or
- 6:14Speaker 5: backward manner. I have very little frame of reference for
- 6:16Speaker 5: the best way to do some of this outside of
- 6:19Speaker 5: consuming tons of YouTube videos.
- 6:21Speaker 1: I think what you've described there, Stephen, is really typical
- 6:27Speaker 1: in terms of how I see a lot of knowledge
- 6:30Speaker 1: workers and solopreneurs using AI where it's very task-based. It's
- 6:36Speaker 1: like putting some input, like write a prompt, give it
- 6:39Speaker 1: a bit of context, get some output back. I'd say
- 6:42Speaker 1: that output, like what I see, like, you know, particularly
- 6:45Speaker 1: when I might be sitting with a client and looking
- 6:48Speaker 1: at a thread like in Claude or Copilot or whatever,
- 6:52Speaker 1: is that there's a lot of.
- 6:53Speaker 2: Back and forth. There's a lot of rework.
- 6:57Speaker 1: There's a bit of arguing with the AI where it's
- 6:59Speaker 1: not giving you the output that you want first time.
- 7:03Speaker 2: And in terms of.
- 7:04Speaker 1: The use cases that you've described there, so there's quite
- 7:06Speaker 1: a bit that fits under comms in terms of email,
- 7:10Speaker 1: thought leadership pieces, reports to clients. There's a bit of
- 7:15Speaker 1: research and a little bit of, I guess, expert support
- 7:19Speaker 1: behind the scenes. So, I think I've kind of captured
- 7:23Speaker 1: all the themes there. And we can go into each
- 7:26Speaker 1: of those because there's probably some advice that I can
- 7:29Speaker 1: give you there to do that better. And then we
- 7:32Speaker 1: can look at sort of the bigger picture in terms
- 7:36Speaker 1: of How I'm seeing more advanced people using AI, how
- 7:41Speaker 1: we're using it at Inventium to really get some more
- 7:45Speaker 1: massive gains as opposed to just task by task, like
- 7:49Speaker 1: help write this email, help write this executive summary and
- 7:52Speaker 1: so on. So let's talk first about written communication, which
- 7:58Speaker 1: is a really, really big area for so many people
- 8:00Speaker 1: in terms of how they're using Gen AI. And I
- 8:04Speaker 1: would ask, do you have, and because we're in the
- 8:08Speaker 1: Claude ecosystem, but I'll try to use terms that are
- 8:11Speaker 1: referring to all the major ecosystems, because we're probably going
- 8:16Speaker 1: to have a lot of listeners that are on Claude.
- 8:17Speaker 1: I'm Personally, Claude first, although I'm across all of them
- 8:22Speaker 1: because we teach across all of them at Inventium.
- 8:25Speaker 2: But do you.
- 8:26Speaker 1: Have like a project or a skill or a GPT
- 8:30Speaker 1: or an agent? And I use all those terms somewhat interchangeably.
- 8:33Speaker 1: All those skills and projects are quite different in Claude
- 8:36Speaker 1: that you've trained in terms of your voice, your style
- 8:40Speaker 1: of writing.
- 8:40Speaker 5: Yeah. In two of the tools, I'd be like you,
- 8:43Speaker 5: I'm now Claude first, but inside of a project, in fact,
- 8:47Speaker 5: both in my profile and Claude, and then inside of
- 8:49Speaker 5: a particular client project, which is my most meaty one,
- 8:53Speaker 5: I've given it like some input.
- 8:56Speaker 4: This is kind of how we need to think about it.
- 8:57Speaker 5: And then in terms of the voice kind of style
- 9:00Speaker 5: of communication, like also told it like what to avoid.
- 9:03Speaker 5: I don't want to sound like one of the big,
- 9:05Speaker 5: my differentiation is that I'm not a big four consultant
- 9:08Speaker 5: because they do that really, really well.
- 9:10Speaker 4: And what I bring is a bit.
- 9:11Speaker 5: More flexible, adaptable, reflective, kind of co-active approach to working
- 9:16Speaker 5: with the client. And so use language in that manner
- 9:19Speaker 5: versus language that's very directive, very say McKinsey-esque or otherwise.
- 9:25Speaker 4: So there's some of that.
- 9:26Speaker 5: And then I've uploaded contextual documents from like different phases
- 9:30Speaker 5: of the project to kind of give a sense on
- 9:33Speaker 5: like they've It kind of picks up on my interaction
- 9:36Speaker 5: between myself and the client and some of the advice
- 9:38Speaker 5: I've given. So that's been ongoing and progressive in one
- 9:40Speaker 5: of the big projects. I have no agents at this point.
- 9:43Speaker 5: So it's just chats, projects, and inside the projects, context uploaded.
- 9:49Speaker 1: So one thing with projects, and this is a general rule,
- 9:53Speaker 1: or GPTs in ChatGPT or Gems and Gemini or agents
- 9:57Speaker 1: in Copilot is what we recommended in Ventium is making
- 10:02Speaker 1: them very, very specific. So for example, in my setup
- 10:07Speaker 1: back when I was mostly ChatGPT, I would have a
- 10:10Speaker 1: GPT that is a script editor for show notes for
- 10:17Speaker 1: how I work, that specific, as opposed to how I
- 10:21Speaker 1: work GPT overall. That's one thing that I'd say is
- 10:26Speaker 1: have really specific projects for really specific uses. Personally, if
- 10:32Speaker 1: I'm thinking about how a report works in consulting I
- 10:37Speaker 1: would have a project for one that helps with qualitative
- 10:42Speaker 1: data analysis, one that is a report editor, one that
- 10:48Speaker 1: perhaps is an executive summary creator. So I would be
- 10:52Speaker 1: breaking it down into very specific projects because that is
- 10:55Speaker 1: going to help the AI be more vigilant because you've
- 10:57Speaker 1: probably had that experience if you've got a really broad,
- 11:01Speaker 1: big project that's sometimes vigilance in terms of how well
- 11:05Speaker 1: the AI listens to, say, the system instructions within a project,
- 11:10Speaker 1: how well it sticks to that or adheres to that.
- 11:13Speaker 1: So the more specific you can be with a project
- 11:15Speaker 1: or a GPT or an agent, generally the better.
- 11:20Speaker 2: I would also say just with.
- 11:22Speaker 1: Your voice and creating your voice, I'm not sure how
- 11:25Speaker 1: you created it, but did you perhaps find examples of
- 11:29Speaker 1: your best work that you're like, yes, this really represents
- 11:33Speaker 1: who Stephen Clarke is and say, you know, describe this,
- 11:37Speaker 1: describe the tone, the style, the structure. Did you go
- 11:41Speaker 1: through a process like that or did you just more input,
- 11:43Speaker 1: this is how I want to sound?
- 11:44Speaker 4: Sometimes I'm shaking my head for a reason. No, I
- 11:47Speaker 4: didn't do it.
- 11:47Speaker 5: I didn't supply it or provide it with this is
- 11:50Speaker 5: some of my best examples of whether that's all different
- 11:53Speaker 5: types of communication. And so no, that's the first step.
- 11:56Speaker 5: first question i'm going back to what you said prior
- 11:59Speaker 5: now i can see something because i have broad projects
- 12:01Speaker 5: that are aligned by the client type or a couple
- 12:04Speaker 5: of like you know giving it as much as i
- 12:07Speaker 5: can within the realms of like data privacy etc but
- 12:10Speaker 5: those are quite broad and there are lots of tasks
- 12:12Speaker 5: inside of those and now just listening to you talk
- 12:15Speaker 5: about no create specific very specific things for discovery analysis
- 12:20Speaker 5: report editor or gut check or re-evaluation or um workshop,
- 12:25Speaker 5: you know, help me define or design a workshop around
- 12:28Speaker 5: these three themes versus doing that inside of the main
- 12:31Speaker 5: broad project.
- 12:32Speaker 1: Another way to think about it, because if you're in
- 12:35Speaker 1: Claude and also ChatGPT now has skills, although I haven't
- 12:38Speaker 1: had a good play around in ChatGPT, but certainly in Claude,
- 12:42Speaker 1: skills can be a way of helping to have broader
- 12:46Speaker 1: projects where you've got a lot of context or knowledge
- 12:50Speaker 1: uploaded into the project, but then you can bring in
- 12:53Speaker 1: skills within that project that are very specific skills. So
- 12:57Speaker 1: for example, for me, I've got an amount, the tone
- 12:59Speaker 1: of voice skill that I've trained and iterated really over
- 13:03Speaker 1: two years. That's kind of an ongoing thing. And I'm
- 13:06Speaker 1: always refining that and now refining that skill that used
- 13:10Speaker 1: to be a project or a GPT slash agent. And
- 13:14Speaker 1: the other thing that I found really important for writing
- 13:19Speaker 1: is that while I've got in my tone of voice skill,
- 13:23Speaker 1: some of the, I'll call them AI tells, the things
- 13:27Speaker 1: that you read a written piece of communication and you
- 13:30Speaker 1: immediately think, ah, I think that was written by AI.
- 13:33Speaker 1: Like maybe there's an abundance of em dashes, maybe there's
- 13:37Speaker 1: lots of sentences that are following the rule of three.
- 13:41Speaker 1: Maybe there's lots of corrective antithesis. It's not this, it's that.
- 13:45Speaker 1: And I mean, I could name many, many of the
- 13:48Speaker 1: AI tells. And I put that within my tone of
- 13:52Speaker 1: voice skill, but I found that adherence wasn't great because
- 13:56Speaker 1: it was just getting quite lengthy and a little bit complicated.
- 13:59Speaker 1: And so I've since built a second skill that's an
- 14:03Speaker 1: anti-AI writing skill that is specifically about scanning a draft
- 14:09Speaker 1: that I've created or that I've been working on with
- 14:13Speaker 1: AI to tell me where are the things that are
- 14:15Speaker 1: essentially AI tells. And that might sound like someone who's
- 14:21Speaker 1: just like, oh, well, you know, you're just getting AI
- 14:23Speaker 1: to produce all your content and then getting it to
- 14:25Speaker 1: strip out any AI tells.
- 14:26Speaker 2: That sounds very dodgy.
- 14:27Speaker 1: And The answer is no, I'm not doing that because
- 14:31Speaker 1: where I found the problem is, is that I will
- 14:33Speaker 1: always do the first draft myself as a human. I strongly,
- 14:37Speaker 1: just as an ethical stance believer, it needs to be
- 14:39Speaker 1: my thinking. It needs to be my ideas. Otherwise, that
- 14:42Speaker 1: feels really disingenuous to me. the people that are consuming
- 14:46Speaker 1: my content. And then it just becomes like popcorn disposable content,
- 14:50Speaker 1: which is not the sort of stuff that I want
- 14:51Speaker 1: to create. So I'll do the first draft and then
- 14:54Speaker 1: I will work with AI as an editor. And again,
- 14:58Speaker 1: like I've got some projects and now some skills set
- 15:01Speaker 1: up that are editors, but they're different kinds of editors
- 15:04Speaker 1: because it's really different being an editor for like a
- 15:07Speaker 1: social media post versus an editor for a book manuscript,
- 15:11Speaker 1: like completely different skills and completely different context. So what
- 15:16Speaker 1: I can then find when it helps me do an edit,
- 15:20Speaker 1: sometimes I'll take on some of those edits and I'll
- 15:23Speaker 1: reword certain things or restructure certain things and so on.
- 15:27Speaker 1: And then what I want to do is in the
- 15:30Speaker 1: final piece that I'm doing, and I should add, this
- 15:34Speaker 1: is not something I do for, say, short back.
- 15:36Speaker 2: And forth emails.
- 15:37Speaker 1: This is more for medium to higher stakes writing is
- 15:41Speaker 1: I will then run it through the AI tell because
- 15:45Speaker 1: I find as a writer there are certain things that
- 15:48Speaker 1: I have learned to do as someone that has been
- 15:50Speaker 1: a writer for most of their adult life that actually
- 15:54Speaker 1: sound like AI.
- 15:55Speaker 2: Because otherwise people go well did AI.
- 15:57Speaker 1: Just write the book and the answer is no a
- 15:59Speaker 1: human did but a human that uses this writing technique
- 16:02Speaker 1: if you like put it in the book. So I
- 16:05Speaker 1: feel like I've kind of rattled and rambled on about
- 16:08Speaker 1: this AI tell, but I think everyone needs this because yeah,
- 16:12Speaker 1: as soon as someone reads a piece of writing and
- 16:14Speaker 1: they see a couple of AI tells, their brain is
- 16:17Speaker 1: immediately going, did AI write the whole thing? It then
- 16:21Speaker 1: goes to reputational damage. And then it also goes to
- 16:24Speaker 1: people ignoring whatever that comms was. So I very, very
- 16:28Speaker 1: strongly believe that. So that's kind of, that's some thoughts
- 16:31Speaker 1: on writing given, you know, everyone's using AI that is
- 16:34Speaker 1: using AI. Are they using it to improve their writing?
- 16:37Speaker 1: So there are a few thoughts there. I'm curious, like
- 16:39Speaker 1: what resonates most out of all of that?
- 16:41Speaker 5: The AI tells hugely, like one obvious one is that
- 16:45Speaker 5: em dashes, but I'm finding certain phrases, like in the
- 16:48Speaker 5: type of work that I do, Like, I'll give you
- 16:51Speaker 5: a couple, like, you know, the things that make the
- 16:53Speaker 5: project stall or drag. And now, and the thing is,
- 16:55Speaker 5: and I started seeing that everyone, I thought I'm using that.
- 16:58Speaker 5: And I thought it was not that it was unique
- 17:00Speaker 5: and there are my, they are my words, but it's
- 17:01Speaker 5: becoming a style. And I saw someone sent me something
- 17:03Speaker 5: last week and I was like, Claude. And immediately it
- 17:07Speaker 5: wasn't to discredit. I was just like, oh, I wonder
- 17:09Speaker 5: if that's how it lands when I do it as well.
- 17:12Speaker 1: Most of us use AI to do one task at
- 17:15Speaker 1: a time. Write this email, summarize this document. But the
- 17:18Speaker 1: bigger gains come from a different place, the work you
- 17:22Speaker 1: repeat every single week. Stick around because coming up after
- 17:27Speaker 1: the break, I share the task that was costing me
- 17:30Speaker 1: half a day every single week and what changed once
- 17:34Speaker 1: AI took it over. If you're looking for more tips
- 17:40Speaker 1: to improve the way you work and live, I write
- 17:43Speaker 1: a short weekly newsletter that contains tactics I've discovered that
- 17:46Speaker 1: have helped me personally.
- 17:48Speaker 2: You can sign up for that at amantha.com. That's amantha.com.
- 17:58Speaker 1: I feel like there's some stuff that I can share
- 18:01Speaker 1: about my own AI tool stack, and then I want
- 18:05Speaker 1: to move to the bigger picture in terms of where
- 18:09Speaker 1: to next with your AI use.
- 18:11Speaker 2: So it sounds like.
- 18:13Speaker 1: Claude and Otter are the main tools that you're using.
- 18:16Speaker 1: So Claude is your main AI platform. Otter is what
- 18:19Speaker 1: you're using for meetings and transcriptions. I love that you've
- 18:22Speaker 1: just doubled down on Clawed, and I strongly recommend that.
- 18:25Speaker 2: Just stick with one. It knows you.
- 18:27Speaker 1: The more you use it, the more it gets to
- 18:29Speaker 1: know you, and you can really double down on the
- 18:32Speaker 1: specific features, whether that be skill building in Clawed or
- 18:36Speaker 1: whatever the case may be.
- 18:37Speaker 2: It's great that you've just got Clawed. I love Otter.
- 18:40Speaker 1: But I will say that my pet peeve about Otter
- 18:44Speaker 1: and Fireflies and Fathom and all those AI note takers
- 18:49Speaker 1: that are essentially for meeting note taking, that's their main purpose,
- 18:54Speaker 1: is that they also join the meeting. And often they
- 18:57Speaker 1: will join before the person who's note taker it is.
- 19:00Speaker 2: It'll be like, ah... Fred's Fireflies has joined the meeting.
- 19:04Speaker 2: And I'll be like, but where is Fred? Where is
- 19:06Speaker 2: the human? And I.
- 19:09Speaker 1: Then also think that that is really disconcerting because they
- 19:12Speaker 1: are there as a meeting participant. And that just unconsciously
- 19:17Speaker 1: changes behavior when you know that the AI is just
- 19:20Speaker 1: sitting there doing its thing, that can change behavior. And Obviously,
- 19:24Speaker 1: we always need to ask for permission. Obviously, it goes
- 19:27Speaker 1: without saying. But you also need to ask the question,
- 19:30Speaker 1: what impact will it have if I'm using an AI
- 19:33Speaker 1: note taker or if I'm hitting record in my Teams
- 19:37Speaker 1: meeting or my Zoom call and using one of the
- 19:39Speaker 1: inbuilt AI meeting note takers? And often, it won't make
- 19:44Speaker 1: a difference. It will be fine.
- 19:46Speaker 2: But if there are issues with psychological.
- 19:49Speaker 1: Safety or or confidentiality or anything like that, it's probably
- 19:53Speaker 1: going to make the other person or people in that
- 19:55Speaker 1: meeting behave quite differently.
- 19:57Speaker 2: That's not a good thing.
- 19:58Speaker 1: So, How I solve for that is I use a
- 20:02Speaker 1: tool called Granola, which I've talked about on the show before.
- 20:05Speaker 1: And Granola is a meeting note take up, but it
- 20:09Speaker 1: is purely words. So it is sitting in the background,
- 20:13Speaker 1: just basically writing like a really fast typist, writing a
- 20:18Speaker 1: transcript of what is going on in the meeting. And
- 20:21Speaker 1: it is not joining the meeting overtly, it's just sitting
- 20:24Speaker 1: in the background. So what I like about that is
- 20:28Speaker 1: that it's not really changing people's behavior, but what I
- 20:32Speaker 1: always do in any calendar invite, so if you were
- 20:36Speaker 1: to have like a sales call with me, for example,
- 20:38Speaker 1: there would be a note in that calendar invite that says,
- 20:41Speaker 1: just a heads up, I use Granola AI to transcribe
- 20:44Speaker 1: our conversation so I can stay fully focused on you
- 20:47Speaker 1: rather than furiously scribbling notes. No audio or video recorded.
- 20:51Speaker 1: Think of it as a very, very fast typist running
- 20:54Speaker 1: quietly in the background, sing out if that's an issue.
- 20:57Speaker 1: So that's important. It's not recording any video or audio.
- 21:01Speaker 1: And of course, if that does matter to you, then
- 21:03Speaker 1: granola is not a great solution for that meeting. But
- 21:07Speaker 1: the heads up is there. And I don't think I've
- 21:10Speaker 1: ever had anyone push back on that. And it's great
- 21:13Speaker 1: because I generally don't take too many notes in meetings now.
- 21:16Speaker 1: I can just stay really, really present with the conversation.
- 21:19Speaker 1: And I think that's just a better outcome for everyone.
- 21:22Speaker 1: So I highly recommend granola, but obviously disclose when you're
- 21:26Speaker 1: using granola. Another tool that is in my stack is WhisperFlow.
- 21:32Speaker 1: And this is more where my workflows have very much
- 21:35Speaker 1: changed from predominantly typing to predominantly talking things. And I
- 21:41Speaker 1: can talk at about 140 words per minute. I can
- 21:44Speaker 1: type at probably about 70 or 80 words per minute.
- 21:46Speaker 1: So talking is more efficient. And also with AI, when
- 21:49Speaker 1: we're talking to the AI, the benefit there is that
- 21:52Speaker 1: We're naturally giving it more context when we talk. And
- 21:56Speaker 1: that's big because the more context we can give our AI,
- 21:59Speaker 1: the better the job that it's going to do for us.
- 22:02Speaker 1: So they would be a few. I use superhuman for
- 22:04Speaker 1: email and the AI that's built into superhuman is really fab.
- 22:07Speaker 2: And they are.
- 22:11Speaker 1: The things that I'm using absolutely every day. There are
- 22:13Speaker 1: a few other very specific ones that, you know, like
- 22:15Speaker 1: just for podcasting, for example, but they're the specific ones.
- 22:18Speaker 1: So I would highly recommend thinking about like, is there
- 22:23Speaker 1: maybe one or two more.
- 22:24Speaker 2: Tools that could really help with your workflows?
- 22:27Speaker 4: Thank you for sharing, especially on the granola.
- 22:28Speaker 5: Because what I'd found more recently, especially in the last
- 22:32Speaker 5: 90 to 100 days, I was actually pulling back because
- 22:35Speaker 5: of that behavior change.
- 22:36Speaker 4: I'd started to notice it over the last nine months.
- 22:39Speaker 5: So that was super helpful. And then Whisper flow I'd
- 22:41Speaker 5: heard about, but the way you articulate it was much
- 22:44Speaker 5: more simpler than someone who'd explained it to me before.
- 22:46Speaker 5: I noticed you do use superhuman. And I'd seen that
- 22:49Speaker 5: on one of a couple of your emails and hadn't
- 22:51Speaker 5: explored it. But the fact that you said it's something
- 22:52Speaker 5: in the everyday stack could be pretty.
- 22:54Speaker 4: Pretty good as well.
- 22:55Speaker 1: Now, let's have a look at the bigger picture because
- 22:58Speaker 1: what we're finding is that with the organizations and the
- 23:03Speaker 1: folk that we're working with at Inventium AI, we're still finding, look,
- 23:07Speaker 1: that there's a lot of people that just don't have
- 23:09Speaker 1: those foundational skills in terms of how to prompt effectively,
- 23:12Speaker 1: how to have a good conversation with the AI, how
- 23:14Speaker 1: to give appropriate context, all those sorts of things. But
- 23:17Speaker 1: Where we're now starting to work with our clients is
- 23:20Speaker 1: more around, well, how do you really level up the
- 23:23Speaker 1: productivity gains that you can get and the impact that
- 23:26Speaker 1: you can have with AI? And that typically looks at
- 23:29Speaker 1: reworking workflows and processes. So the best place to start
- 23:36Speaker 1: here is is thinking about what are the tasks.
- 23:39Speaker 2: That you do regularly?
- 23:41Speaker 1: So it might be, what are the tasks that you
- 23:44Speaker 1: generally are doing most days or are generally doing most
- 23:48Speaker 1: weeks or perhaps are generally doing most months? And oftentimes
- 23:54Speaker 1: people will be like, I don't know, I just do
- 23:57Speaker 1: my work because most of us are not trained to
- 24:01Speaker 1: be process engineers or workflow architects, we just don't think
- 24:05Speaker 1: in that way. We do our job and we hopefully
- 24:07Speaker 1: do it pretty well, but we're not really thinking about,
- 24:11Speaker 1: what are those repetitive things? So what I recommend is
- 24:14Speaker 1: just for a week, keeping a note of any time
- 24:17Speaker 1: you are doing a task and the task you are
- 24:20Speaker 1: doing is something that you have done at least once
- 24:24Speaker 1: in the previous week, or perhaps in the previous month,
- 24:26Speaker 1: if you want to lengthen it out. And most people
- 24:28Speaker 1: are pretty good at doing that. Just make a note
- 24:30Speaker 1: of it somewhere. And then once you've got a list
- 24:33Speaker 1: of tasks, you can start to pick off those tasks
- 24:36Speaker 1: and ask yourself the question, which ones could AI help
- 24:39Speaker 1: with in terms of what could AI automate?
- 24:42Speaker 2: What could AI speed up? And this is typically.
- 24:46Speaker 1: through building an agent or a project or a GPT.
- 24:49Speaker 1: So typically through that way, like you can get a
- 24:51Speaker 1: little bit more complex and, you know, bring in things
- 24:53Speaker 1: like Zapier or N8N and some different automations. But even
- 24:58Speaker 1: just working within Claude, you can do a lot in
- 25:00Speaker 1: Claude Cowork or even in Claude Code. But, you know,
- 25:04Speaker 1: I certainly think Claude Cowork is really great for any
- 25:07Speaker 1: knowledge worker that is looking to automate things and just
- 25:10Speaker 1: stay within the Claude ecosystem without having to get into
- 25:14Speaker 1: fancy automation software. So once you've got one process that
- 25:19Speaker 1: you're going to figure out, then the question is, well,
- 25:22Speaker 1: how can AI help? So one of those tasks that
- 25:25Speaker 1: I do multiple times a week is I will have
- 25:31Speaker 1: a sales call with someone I've never met. Like in
- 25:34Speaker 1: a typical week in my diary, I've probably got between,
- 25:37Speaker 1: I reckon, five to 15 sales calls. 15 is a
- 25:40Speaker 1: really busy week. 5 to 10 is more of a
- 25:43Speaker 1: standard week. And these are typically people that I've never
- 25:47Speaker 1: met before. They've reached out to.
- 25:49Speaker 2: Inventium or Inventium AI to inquire about.
- 25:52Speaker 1: What we do or scope out a project with us.
- 25:56Speaker 1: And I want to go into that meeting knowing about
- 25:58Speaker 1: who are they and what's their company, if I haven't
- 26:02Speaker 1: heard of their company. And what has this organization been
- 26:06Speaker 1: in the news for lately? What do I need to know?
- 26:09Speaker 1: That's about a 20-minute task. And if you multiply 20
- 26:12Speaker 1: minutes by, let's say, 10 meetings, that's 200 minutes. That
- 26:18Speaker 1: is over three hours. That is about half a day's
- 26:20Speaker 1: worth of time. that I'm doing a task that I
- 26:24Speaker 1: could actually give to AI. And so now what I've
- 26:27Speaker 1: done in Claude is I've created a skill called pre-call dossier.
- 26:32Speaker 4: Dossier?
- 26:32Speaker 1: Is that how I pronounce it? I always pronounce that wrong.
- 26:34Speaker 1: But anyway, it's basically my pre-call skill that I can
- 26:38Speaker 1: type in who I'm meeting with, the organization that they're from,
- 26:41Speaker 1: and Claude will now go through all those steps which
- 26:46Speaker 1: I've programmed into a skill and it will produce results.
- 26:49Speaker 1: approximately a two page report that I can quickly read
- 26:53Speaker 1: and get across all the things that I need to know.
- 26:55Speaker 1: And now what I've done is I've given that skill
- 26:57Speaker 1: to my EA who runs that skill. And they just
- 27:00Speaker 1: put a parallel calendar invite in my diary that has
- 27:04Speaker 1: that pre-call dossier. And now all I have to do
- 27:08Speaker 1: is like, if I've got a meeting, then I just
- 27:10Speaker 1: pop in two, three minutes prior to the meeting. I
- 27:12Speaker 1: read that short report and I'm like, great, I'm prepped
- 27:16Speaker 1: for the meeting. So that's just one example of quite
- 27:19Speaker 1: a few in my workflow. And literally every couple of weeks,
- 27:24Speaker 1: I'm doing that for another process because I think it's
- 27:27Speaker 1: really hard to set aside a whole big chunk of
- 27:30Speaker 1: time to do that. But I know that I can
- 27:33Speaker 1: do this in bite-sized chunks. And so long as I've
- 27:35Speaker 1: got that list of what are all the things that
- 27:38Speaker 1: I'm doing repeatedly where AI can play a role, then
- 27:41Speaker 1: I know that I'm going to be focusing on the
- 27:43Speaker 1: right things when I do have AI build time. So
- 27:46Speaker 1: I'm wondering for you, Stephen, is there a process maybe
- 27:50Speaker 1: that's coming to mind or something that you might do
- 27:52Speaker 1: with what I've just described?
- 27:54Speaker 5: That research one or prep before the meeting. Sometimes if
- 27:57Speaker 5: it's someone, if they're not brand new, sometimes I haven't
- 28:00Speaker 5: spoken to them in a while. So I'm in that
- 28:01Speaker 5: mode of like reconnecting with people I haven't spoken to,
- 28:05Speaker 5: maybe it's like been 18 months, some longer. And just
- 28:08Speaker 5: trying to get a quick, before I go in, what
- 28:10Speaker 5: have they been up to? Kind of glanced at LinkedIn
- 28:13Speaker 5: or I've glanced at some of, what other work do
- 28:15Speaker 5: they have out there or what other, where other places
- 28:17Speaker 5: have they spoken?
- 28:17Speaker 4: What other things are going on with them? So that,
- 28:19Speaker 4: that would definitely be one.
- 28:21Speaker 5: A big one for me is I take sometimes input
- 28:24Speaker 5: from a discussion or a meeting and I'm trying to
- 28:27Speaker 5: formulate like meaningful next steps and not just tacticals, like ideas.
- 28:33Speaker 4: How should we be thinking about this? What other prompts
- 28:35Speaker 4: do I have for them?
- 28:35Speaker 5: What other ways should they be thinking at this particular
- 28:37Speaker 5: opportunity or issue to kind of get me and the
- 28:41Speaker 5: person I'm working with or the client in the back
- 28:43Speaker 5: of the world, keeping them thinking, keeping them feeling like
- 28:46Speaker 5: they're getting value more than.
- 28:48Speaker 4: Just, Next steps, complete this by Thursday.
- 28:51Speaker 5: You know, the thing I think about, it's how can
- 28:54Speaker 5: I add value over and above what we've just discussed
- 28:57Speaker 5: or the status check, et cetera. So it's almost like
- 28:59Speaker 5: a value creation repeat thing. And I'm doing it manually,
- 29:02Speaker 5: somewhat sporadically. If I remember, I'm trying to be deliberate.
- 29:06Speaker 5: So that's just another one that probably springs to mind.
- 29:09Speaker 5: Do you know the other one is actually something I
- 29:12Speaker 5: try and do this when I go back through the
- 29:13Speaker 5: last 30 to 40 days of dialogue. Otto will kick
- 29:17Speaker 5: out at the beginning of a week Based on some
- 29:20Speaker 5: of the meetings I have allowed it to participate in,
- 29:23Speaker 5: here's what's coming up or here's the last, it's just actions.
- 29:27Speaker 5: So somehow bringing forward the context from the last 30 days, activities, actions.
- 29:31Speaker 4: Functions, part of that's for me. Like, what am I learning?
- 29:34Speaker 5: What am I seeing? What patterns am I observing? What
- 29:37Speaker 5: should I be thinking about as an implication? That's probably
- 29:40Speaker 5: the other one, some sort of recap.
- 29:42Speaker 4: What am I learning?
- 29:43Speaker 1: And so what I would recommend there, and one of
- 29:45Speaker 1: the reasons why I think Granola is so great, and
- 29:48Speaker 1: it comes with me to every meeting, and then because
- 29:51Speaker 1: Granola is with me with any conversation I'm having from
- 29:55Speaker 1: my one-on-ones with my team to sales conversations to team
- 29:58Speaker 1: meetings and so on, It holds so much knowledge. And
- 30:02Speaker 1: then if you connect Granola to Claude, which you certainly can,
- 30:08Speaker 1: what you can start to do is if you think
- 30:10Speaker 1: about what are those questions that I have regularly, like
- 30:12Speaker 1: for example, every Friday you might have the question, what
- 30:15Speaker 1: are all the actions that I had to do based
- 30:18Speaker 1: on all the meetings and conversations that I've been in
- 30:21Speaker 1: this week? You could then program that as a recurring task, Typically,
- 30:27Speaker 1: co-work is the easiest place to do that. The AI
- 30:31Speaker 1: can do that for you so that you're not manually
- 30:34Speaker 1: having to prompt the AI. All you need to remember
- 30:37Speaker 1: to do is just pop into the core desktop terminal
- 30:41Speaker 1: and go, oh, I just need to get this scheduled
- 30:45Speaker 1: task and look at the output. I find where there's
- 30:49Speaker 1: something that I have to do every week and the
- 30:53Speaker 1: input is there. I don't have to give the AI
- 30:55Speaker 1: specific input, but if I do, then obviously, I can
- 30:58Speaker 1: create a skill or a task to do that, depending
- 31:01Speaker 1: on what your AI has access to. It just eliminates
- 31:05Speaker 1: a little bit of friction that is in the way
- 31:07Speaker 1: to doing those things that you do every week or
- 31:10Speaker 1: every month. For example, for me, I've got a researcher
- 31:14Speaker 1: that runs every morning at 8am and I've prompted it
- 31:18Speaker 1: in a really specific way so that I can stay
- 31:20Speaker 1: up to the latest news in AI, but in the
- 31:23Speaker 1: context that I care about, which is from an organizational
- 31:26Speaker 1: psychology context and someone that runs an AI training company.
- 31:30Speaker 2: So, Stephen, we've talked about a lot of things.
- 31:33Speaker 1: I want to know what is one thing right now
- 31:37Speaker 1: that's really sticking with you that you can commit to
- 31:41Speaker 1: doing in the next seven days or commit to trying
- 31:44Speaker 1: in the next seven days?
- 31:45Speaker 5: The main one is just that activity sort of the
- 31:48Speaker 5: next seven days, writing down when I've repeated. A particular
- 31:52Speaker 5: task I'm mindful there's probably more gold there than I
- 31:56Speaker 5: realized because I don't naturally inherently do. Before I even
- 31:58Speaker 5: pick another tool, which you've given me some great examples,
- 32:01Speaker 5: some granola yams, which I'm like, yeah, I'm ready to
- 32:03Speaker 5: jump on. But that's the one in the next seven days,
- 32:06Speaker 5: which is accessible. Just requires me to have a notepad
- 32:08Speaker 5: right here.
- 32:09Speaker 4: Or a notion. I'll probably do digital and analog, to
- 32:12Speaker 4: be honest.
- 32:12Speaker 2: Amazing. I love that. And I look forward to hearing
- 32:14Speaker 2: how you go.
- 32:15Speaker 1: Stephen, thank you so much for coming on and just
- 32:17Speaker 1: sharing really honestly about like what's happening for you in
- 32:22Speaker 1: your world of work and AI. I think it's just
- 32:25Speaker 1: so relatable to where so many listeners are at right now.
- 32:29Speaker 2: So thank you so much for coming on and sharing.
- 32:31Speaker 5: Thank you for having me, allowing me to be open, honest,
- 32:34Speaker 5: and like almost a little vulnerable about what I know
- 32:36Speaker 5: and what I don't know and how early in the
- 32:37Speaker 5: curve I am.
- 32:38Speaker 4: Appreciate you.
- 32:39Speaker 1: If you would like more tips on how to get
- 32:42Speaker 1: more value from AI, then I think you will love
- 32:46Speaker 1: How I AI. There is a link to that show
- 32:51Speaker 1: in the show notes where every week my co-host Neo
- 32:54Speaker 1: Applin and I sit down and share practical ways that
- 32:58Speaker 1: you can be extracting a lot more value from whichever
- 33:02Speaker 1: AI tool you are using. If you liked today's show,
- 33:05Speaker 1: make sure you hit follow on your podcast app to
- 33:08Speaker 1: be alerted when new episodes drop.
- 33:11Speaker 3: How I Work was recorded on the traditional land of
- 33:13Speaker 3: the Wurundjeri people, part of the Kulin Nation.