This month Ashley Pyle — CXO of GadellNet and a Copilot trainer since 2024 — showed how to make the Copilot sitting inside your Microsoft 365 actually pull its weight. The framing that stuck: think of it less as a chatbot and more as an executive assistant, keep an open mindset about tools, and always stay the human in the room.

Welcome & housekeeping

Ivy opened with a couple of reminders. A topic-suggestion form is pinned in the chat and lives on the website — the group genuinely uses it to decide what to teach each month, so fill it out ("we're begging you, so we're not giving you content you don't want"). Next month's registration link is pinned too and will be posted to the site shortly; everyone on the list gets an email. Sessions are recorded and posted, and Sarah Gudeman (co-founder) and Melody (leadership group) helped moderate the chat. As always, this is a volunteer-run resource — member feedback drives the agenda.

This month's speaker: Ashley Pyle

Ashley is Chief Experience Officer at GadellNet Consulting Services, an IT consulting firm with AI consulting, cybersecurity, and help-desk divisions. She's been in tech for 20 years, 11 of them at GadellNet, and calls herself a champion of accessible technology. Her "why" is personal: prepping for a women-in-construction leadership conference in 2024, she found women were adopting AI at notably lower rates than men — a gap that risks widening the wage gap — so teaching this technology, especially to women, became a mission. She usually spends half a day on this material; today was a fast, one-hour tour with a full resource pack to follow. Connect with Ashley on LinkedIn.

Start with mindset, not tools

The people Ashley sees get the furthest with AI don't have a fixed mindset — as the technology changes, they experiment, keep an open mind, test, and apply judgment. Her standing encouragement: compare Copilot against the tools you already know and love, and keep what wins.

She also grounded the session in some adoption research she shared with the group: chatbot use has jumped and the gender gap has nearly closed — roughly 49% of US adults now use chatbots (up from about 33%), with women's use climbing from about 28% to 48% in two years. Her caveat, which is worth keeping: go beyond the headline, because some of that is personal use, not work. The original research tied the early gap to fear and uncertainty — move people from fear to fascination and adoption follows, especially when leaders work with their teams on real workload change rather than just swapping prompts.

These figures are the ones Ashley cited from adoption research she distributed to the group. Treat them as approximate and verify against the primary source before quoting them.

Every job is now three kinds of tasks

A framing to sort your own work by:

• Tasks AI can fully handle — work AI does quickly and consistently with little direction: predicting, recommending, analyzing, and summarizing (e.g., a summary or basic analysis off a Teams transcript). Build a saved-prompt library so a project-management call or a sales call returns the exact format you need, not a generic one.
• Tasks you do with AI — work shaped through back-and-forth, where the outcome improves as you guide, adjust, and apply your expertise. This is iteration and refinement, and it was most of today's focus.
• Tasks only you can do — judgment, decisions, communication, and accountability. Her example: GadellNet automated month-end close to four days, but there's a serious human review before the P&L is published.

Copilot is the racetrack, not the race car

A mental model worth stealing: the models are the cars; Copilot is the track. When people say they "use Copilot," they're usually not naming a specific model. Inside Copilot you get access to the OpenAI family and — via the Microsoft–Anthropic partnership — Anthropic's models, and Microsoft now ships new models same-day rather than lagging behind. Copilot auto-selects a model, but you can pick (Ashley reaches for Anthropic models when she's working with data in Excel). She also flagged Model Council, where models run side by side so you can compare outputs — she uses it for deep research rather than everyday tasks.

Basic vs. Premium — and what you can safely put in

• Basic — free with most Microsoft 365 plans. Web-grounded chat with enterprise data protection (look for the little green shield): your data stays in your tenant and does not train the foundational models, and there's no consumption limit. It doesn't see your files, emails, or meetings on its own. General guidance she gives when helping orgs write an AI-use policy: financial data is fine — it's no different than a spreadsheet in your SharePoint — but always check your own organization's rules first.
• Premium — work-grounded across your emails, files, chats, and meetings ("Work IQ"). It knows who you report to, your workstreams, and how you communicate. Adds meeting recaps, Copilot Pages, custom agents, and Copilot Cowork for delegating and scheduling tasks. Cowork is consumption-based and does cost money, so prove the ROI.

Know your own guardrails. Ashley can't see your org's licensing, approved tools, or rules on client/regulated data — and if leadership can't answer "when can and can't we use AI?", teams can't either. Learning the tool and learning your organization's policy go together.

Treat it like an executive assistant

Rather than "a brilliant intern," Ashley thinks of Copilot as her chief of staff — hand off the busy work (drafts, inbox summaries, prep). A real example from the day before: after a transcribed Teams call with a partner, she pointed Copilot at the transcript plus a research doc and said "run the most important parts of this research against our conversation and prepare an email." It knew her tone, dropped a draft into her Outlook with everyone's addresses already on it, and the whole follow-up took about a minute after a quick human review.

Power-user mode

From an (older but still relevant) Microsoft study, the most successful users tend to: start the day with Copilot and use it to get ready for tomorrow; re-prompt instead of stopping at a bad answer (you can edit a prompt directly and regenerate); experiment with new prompts and ways of working; share prompts with coworkers; make learning Copilot a deliberate priority; and constantly ask "could Copilot help me with this?"

The live demos

She also reviews whole decks for consistency, brand, voice, and objectives; generates images (the Flux image model is now available); and runs meeting recaps by reasoning over transcripts. One more trick worth stealing: from a single incoming email, ask for three versions for three audiences — a client reply, an internal team brief, and a note to legal — and send each out.

That's Ashley's advice for getting off zero: it may take longer than copy-pasting the first time, but that's the point — you're learning the workflow, not just finishing the task.

Ivy's closing reminders

• Use a paid model — even the $20/month tier — to better protect your data from legal exposure.
• Be the human in the room. Automate the repetitive stuff, but keep the oversight. "It's still a machine."
• Drop your topics on the website so the group can teach what you actually need.

Coming up next

Next month, Michelle teaches everyone how to build your own chief of staff — using Copilot, Claude, and skills in Claude. Registration is pinned in the chat and posts to the site within about a week. And by popular demand, Ashley is already invited back for a part two.

Resources

Ashley built a full follow-up kit for the group — the complete slide deck, a prompt pack / library, a demo kit (with separate sheets for basic and premium licenses), quick-reference cards, and an AI-opportunity worksheet for spotting where AI can save you time. These are being emailed to members and added to the site. In the meantime, browse the CRE Prompt Library, and help shape future sessions through the member survey.

Recurring logistics: 3rd Friday of every month, 10:00 AM ET. Email Ivy at imarguerite@eyeqmonitoring.com · Register · Join the meeting.