Same licenses.
Different work.

Dana feels confident.

Josh: Introduce Dana as Parkside’s CEO. Show the adoption report, internal assistant launch and promising Sales example.

“Dana feels good about AI at Parkside. Everyone has Copilot. IT launched an assistant. Sales showed her something impressive. She uses Claude to challenge the board deck.”

“If we spent a morning sitting beside her employees, what would we actually see?”

Available licenses and a good demo are useful signals. They do not show how every task gets done.

The research frames the question.

Josh: Spend thirty seconds on the figures. Keep the research general and introduce the employees on the next page.

“Miro found 83% agreeing that AI makes individuals more productive, while 69% said most AI work happens in private sessions. Atlassian found 85% using AI at work, but 29% with it embedded in their workflow.”

“Individual benefit is real. The question is how that benefit becomes shared work.”

Separate, self-reported vendor surveys support the problem framing. They are not measured Nine-67 results.

Meet Parkside.

Josh: Orient the audience using the large employee cards. Open one workspace to establish the interaction. Keep the fixed card positions as the reference for today and the proposed improvements.

“Here are the people. Everyone has a company Copilot license. They have different jobs, skills, habits and local tools. Click a person and you can see the work, not just the tool name.”

“We can also look at one proposed improvement at a time. The people stay in place so you can see what would change for them.”

The proposed map is a simulated design. It connects authorized business outputs and review steps, not everyone’s private AI conversations. There is no all-seeing central AI hub.

Start with what already works.

Josh: Open Ben, Priya and Nina. Show each useful output and the next place it could be useful.

1. Ben · Sales“Copilot in Excel compares four weekly HubSpot snapshots. Ben checks repeated close-date slips and missing buyer activity, then coaches three reps. The method works, but largely stays inside Sales.”
2. Priya · Finance“With approved Claude and synthetic example rows, Priya built a Power Query transformation. Refresh normalizes three branch files and flags duplicates and unmatched rows for her review. Another office still does that preparation manually.”
3. Nina · Operations“Six people use Nina’s Teams, Power Automate and SharePoint-list workflow to group reopened cases for their morning huddle. It is useful. IT has not seen how the team is using it.”

Suuchi: “Discovery should find these successes. Sometimes the best next step is to make a good local method reusable.”

Priya’s process prepares data and flags exceptions for human judgment. Nina’s team reviews the proposed groups. Neither is automatic approval of a business conclusion.

The same access does not mean the same help.

Josh: Open Ava, Elena and Marcus. Keep the contrast practical rather than judging their proficiency.

1. Ava · Customer Service“Copilot helps her rewrite tone twenty times a day. To find the facts for one reply, she still opens six tickets. The next opportunity is the context she has to collect.”
2. Elena · Project Coordination“‘Help with my projects’ produced a generic answer. She still copies Thursday’s tracker manually. A guided review of her actual notes and tracker would give her a practical way to use AI.”
3. Marcus · Account Management“He has a company ChatGPT seat and still rehearses generic objection calls on Grok on his own phone, because that is where he learned before the rollout. No names or customer records. The technique works and his team never sees it. We ask him how he practices and rebuild it as a company Call Prep Coach. His personal chats stay his.”

Suuchi: “These are three different needs: better context, task-specific support, and clear guidance. We should not prescribe the same fix to all three.”

Leah does the assembly. Dana gets the summary.

Josh: First show Leah’s manual Friday reconciliation. Then show the summary Dana receives and uses in Claude.

Josh, quoting Leah: “Everyone gets their update to me faster now. I still spend Friday figuring out which updates are talking about the same thing.”

“In this fictional Friday, Leah spends two hours reconciling updates before Copilot tightens the deck. Dana receives that deck. Claude can challenge what is in it, but someone still had to build the company-wide picture.”

The two hours are a scenario detail, not measured ROI. Do not imply Dana’s assistant can access private conversations or unshared source records.

Three findings from looking across the company.

Josh: Walk the findings. Each should lead to a different kind of recommendation.

1. Reuse a success“Priya has a repeatable preparation method. The other branch still does it by hand. Validate and share the method with an owner and review rules.”
2. Match the support“Equal access produced different outcomes. Elena needs help with Thursday’s actual tracker, while Ava needs a reliable way to get source context.”
3. Carry useful information“Nina’s reviewed findings, Marcus’s account observations and Leah’s weekly brief need an agreed path between the records.”

Suuchi: “We start company-wide. Then we prioritize what to teach, what to reuse and what to build. The first implementation stays focused.”

Show what Nine-67 could build.

Josh: Visit the three build examples. Show the changed work, not a long architecture explanation.

1. Priya · Reusable branch workflow“Another branch can use the reviewed preparation method. The process normalizes the inputs, preserves the checks and sends exceptions to a person. We have made a local success repeatable.”
2. Elena · Guided tracker task“Start from the actual Thursday tracker and the update she needs. Guide the task, prepare the draft and keep her review visible. This is specific help with her job.”
3. Nina · Reviewed finding that travels“Inspect the source and confirm the finding. Assign the owner and publish it. Show the same reviewed issue, owner and due date in Operations, Marcus’s account view and Leah’s weekly review.”

Suuchi: “The improvement is not that three people received another summary. They are working from the same reviewed finding and can see who owns the next step.”

The live service example reviews CASE-201, confirms its finding, and publishes the assigned follow-up. Keep the wider case list optional; do not claim all candidates have been reviewed. The finding is still an issue to investigate, not a proven cause or a resolved problem.

All builds are simulated designs on company tools. No personal account, private chat or live customer record is accessed.