Transform Workflows with Custom GPTs Now

Full video transcript

If you’re not already building GPTs for your own personal or internal team or company use, this is a great way to get started, and to think about doing that instead of only thinking about the ones that are public facing, for chat support, or as a lead magnet.

Leandra Nisbet joined to share how she is scaling her business with AI as part of the Scale with AI Summit series. She opened by outlining some tips and actionable steps for creating a custom GPT to boost productivity, aimed at helping attendees leverage information they’re already learning at a rapid pace. She noted that most people have a store of information sitting around, collecting digital dust, that could be put to better use for productivity, and framed the session around getting a better return on investment from courses, training classes, programs, and masterminds.

Nisbet introduced herself as the founder and owner of a professional services company focused on helping small business owners launch, grow, and scale, with a personal interest in tech automation, process improvement, efficiency, and structure — while acknowledging the “shiny object syndrome” that comes with fast-changing tech. Her goal was to show how to leverage information people have compiled over weeks, months, or years but aren’t fully using: notes from summits, trainings, workshops, masterminds, podcasts, and YouTube content, plus paper workbooks and manuals scattered everywhere.

She laid out the process in steps. Step one is physically pulling all of that material together — workbooks emailed from a course, paper copies mailed by an organization, PDF guides shared in event chats, and files shared via Google Drive — into a centralized database or collection. Step two is determining actionable goals based on that material. Her working example was goal setting and planning: using compiled information to build an actionable weekly plan, support quarterly planning, and improve daily time management. She noted the goal could differ for anyone — some people just want a better understanding of information they absorbed too quickly (“drinking from a fire hose”), or want to revisit material that wasn’t immediately applicable and get back up to speed months later.

Once goals are set, the next move is creating a private data store, in this case a custom GPT, to compile those resources. She noted this requires ChatGPT Pro, though other tools (she mentioned one called Team AI, which combines multiple AI models) can serve the same purpose. She then demonstrated live in ChatGPT: creating a new GPT named “time management and planning assistant,” adding an internal description reminding herself of its purpose, and filling in detailed instructions — that the GPT is a professional coach focused on support, accountability, and guidance, approachable, polished, and professional, meant to help users implement what they’ve learned, asking clarifying questions as needed. She skipped conversation starters since this GPT was for her own private chat use rather than external users.

She then uploaded prepared knowledge files — PDFs she built for annual, weekly, and daily planning — into the GPT’s Knowledge section, noting she believed the platform allowed roughly 20 files in a knowledge base at the time, something to plan around in advance. She set the GPT to be usable only by herself, not published to the GPT store and not shareable by link.

In the live demo, she typed “hi, help creating my annual plan” into the GPT. It followed the steps built into her uploaded resource, starting with a reflection on the past year’s wins, and asked clarifying questions when her answers were brief. She then switched topics to daily time management, telling the GPT she was getting distracted by pop-up tasks; it asked about her specific distractors (email, phone calls) and about when she felt most focused versus most drained, working toward protecting her high-focus time from disruptions.

She then showed one of the underlying PDF resources on screen in response to a chat question, explaining how she structures notes as she takes them: writing raw takeaways at the bottom of a document while building a “recipe” at the top — defining who the GPT is, its goal, and a numbered sequence of steps. For the annual/Q1 planning resource specifically, the steps were: brain dump wins from the year, summarize the year into a word or phrase, identify top three priorities for the year, brainstorm what a successful year looks like, narrow to a top three, brainstorm what a successful first month (January) looks like, and narrow that to a top three as well — ending with defined top-three goals for both the year and the quarter.

She offered several closing tips. First, be strategic and organized in note-taking: centralize all files for a given course or program in one cloud-storage folder rather than letting them live in email, Drive, and chat exports; and build “recipe” prompts into notes as you go rather than revisiting raw notes later. Second, think broadly about available assets — not just workbooks and worksheets, but transcripts and chat files too. Third, be cognizant of usage rights: her own GPTs are for personal use only, not published or shared, and most course or workshop intellectual property is licensed for personal use rather than for building external tools. Fourth, extend this same strategy to internal business use — she mentioned her team is currently building a GPT-based training vault for a client, incorporating existing and newly written standard operating procedures and manuals, to support team onboarding and reduce recurring questions.

As a closing note, she suggested building prompts into notes as you go to save time later, and that these knowledge bases can always be refined, updated, or expanded afterward. She then opened the floor for questions, mentioning she could be reached in the GenAI community and on Instagram, before the session handed back to Darby to close out ahead of the next Scale with AI Summit session.