CurrencyTransfer offsite · facilitator's copy · private

CT Offsite Run-Sheets

Two sessions, one page. A short facilitation crash course, then a timed block-by-block plan for Wednesday's selling session and Thursday's AI session. Timings assume ~75 minutes each — every block notes what to cut if you're running short.

Facilitating in six rules

You don't need to be a trained facilitator. You need a structure (below), a visible clock, and the discipline to ask instead of answer.

Session 1 · Wednesday

How to Sell What We Already Have

Facilitated by you · runs right after Mark's Jobs to Be Done core session — build on it directly, don't restart from zero.

The trap in the title: this can't be you presenting a feature list. The goal is that the team builds a shared answer to "what do we sell and why does it matter", in their own words, so they'll actually use it. Your best asset is sequencing — Tuesday produced an agreed ICP, and Wednesday morning produced client jobs. This session converts capabilities into value language for those clients doing those jobs.

00:00 – 00:05Frame it

One minute of setup, no slides. Then the structure: inventory → translate → pitch → capture.

"We're not learning to sell something new. My bet is we already have more to sell than we consistently talk about. By the end of this hour I want one page we all agree on: what we have, which client jobs it does — stealing directly from this morning — and the words we'd use to sell it."
00:05 – 00:20Capability inventory

Silent brainstorm, one capability per sticky: "Everything CT actually has today that a client gets value from — features, service, people, process." Two minutes silent, then stick them up and cluster together as a group. You'll get duplicates; that's data — the duplicated ones are your core story.

It will drift into what we don't have. One sentence — "real, parked, that's tomorrow's Magic Wand session" — and back on track. This session is deliberately about the existing kit.
Running short? Cluster faster — 10 min. Don't skip the silent writing.

Pocket list — seed these if the wall looks thin

00:20 – 00:40The so-what ladder

Pairs. Each pair takes 2–3 capabilities off the wall and climbs the ladder for each: We have X → so the client can Y → which matters because Z — where Z must connect to a job from this morning's JTBD session or to the agreed ICP. Give one worked example, then let them go for ~12 minutes, then each pair reads out their best one.

"Example: 'We have multiple regulated partners' — so what? — 'so we route your payment to whoever prices it best that day' — so what? — 'so you get bank-beating rates without ringing three brokers.' Nobody buys the first sentence. Everybody buys the third. I want the third sentence for everything on this wall."
Running short? Two capabilities per pair, skip the read-out of weaker ones.
00:40 – 01:00Pitch gauntlet

Same pairs, roles: one sells, one plays a specific client from the ICP work — sceptical but fair. 90-second pitch using only ladder language, then two minutes of feedback, then swap. If energy is high, run a second round where the listener throws one real objection ("my bank does this", "what if you disappear", "why not Wise").

Keep feedback structural — "which sentence landed, which lost you" — not performance review. And resist demonstrating a "correct" pitch yourself; the point is their words.
Running short? One round, no swap. This block is the heart — protect it over the inventory.
01:00 – 01:15Capture and commit

Back to plenary. Scribe builds the one-pager live: the 5–7 capabilities that came up most, each with its best ladder sentence. Read it back. Then assign: one owner polishes it into a shared "how we talk about CT" sheet within a week, and it feeds Mark's client-communications work and Ingrid's marketing.

"This page is the deliverable. If it's not in your own words now, it never will be — last chance to object to any line."

Night-before checklist — Session 1

Session 2 · Thursday

AI and Automation: CT in 2026 and 2027

Led by you · runs right before Mark's Objectives Sep–Dec session — your output should hand him candidate objectives with owners.

Different mode from Wednesday: here you lead. You're the credible AI person in the room, so a short demo is expected and earns you the right to run the exercise after it. The structure: show what's possible, then let the team tell you where the pain is, then converge on a 2026 quick-win list and one or two 2027 bets. The output is a ranked backlog Mark can turn into owned objectives an hour later — that's what makes this session land.

00:00 – 00:12Demo, don't lecture

Open with something real and CT-shaped, live if at all possible — e.g. an agent drafting a client follow-up from a meeting note, or the WhatsApp quote-and-book concept if you're ready to show it. One demo done well beats ten slides about AI trends. Then one framing slide only: 2026 = automate what we already do (weeks, not quarters). 2027 = capabilities that change how we serve clients and grow.

"None of this is about doing more with fewer people. It's about deleting the boring 20% of your week so the interesting 80% gets your full attention — and giving clients responses in minutes instead of hours. What I can't know from where I sit is where your boring 20% actually is. That's this session."
The unspoken question in every AI session is "is this about replacing me?" Kill it explicitly in the first five minutes — the line above does it — or you'll get polite, useless answers all hour.
00:12 – 00:27Time audit

Silent, individual: "Write your three most repetitive or annoying tasks — things you do weekly that follow roughly the same steps every time — with a rough hours-per-week guess on each." Three minutes writing, then round-robin: everyone reads theirs out, stickies go on the wall. Cluster obvious duplicates as you go. Duplicated pain across roles is gold — automate once, save everyone.

Running short? Two tasks each. Never cut the round-robin — hearing everyone's pain builds the buy-in.

Pocket list — prompts if the room blanks

00:27 – 00:42Map it: impact × effort

Draw a 2×2 — impact (hours saved or client experience) vertical, effort horizontal. As a group, place the clusters. You're the calibrator on effort: the room will overestimate some ("that's actually a week of work now") and underestimate others ("that needs the dev team and a quarter"). Top-left quadrant — high impact, low effort — is your 2026 list. High impact, high effort is your 2027 capabilities list.

Don't commit dev-team capacity live in the room — Kamen's team has a full plate. Sort candidates into "we can build with AI tools around the edges, no platform work" vs "needs platform work, I'll scope with the dev team" and only promise the first kind on 2026 timelines.
00:42 – 00:55Vote and pick

Dot-vote: three dots each on the top-left quadrant. Take the top three as the 2026 quick-win commitments. Then you nominate one or two 2027 capability bets from the high-effort column — this part is a leadership call, not a vote — e.g. a conversational client channel (the WhatsApp direction) or AI-assisted client communications at scale. Explain your reasoning in two sentences each.

Running short? Skip formal voting — call the top three from the clustering, check for objections.
00:55 – 01:10Commit and hand to Mark

For each of the three quick wins: an owner (the person who feels the pain, not necessarily the builder), a first step, and a "working by" month. Write them in the same format Mark will use in the objectives session. Close by physically handing the list over.

"Three things working by [month], with names on them, and two bigger bets I'll scope for 2027. Mark — these want to be on the Sep–Dec objectives list."

Night-before checklist — Session 2