CASE STUDY · REAL ESTATE SALES · AHMEDABAD
40+ salespeople, from chat to Claude Cowork in two sessions
I trained 40+ people on Shivalik Group’s Sales team in two sessions, on 25 July and 8 August 2026. The first made ChatGPT, Claude and Gemini useful for pitches, follow-ups and HNI call prep. The second moved them to Claude Cowork, which takes a whole task from prospect research to the follow-up draft.
By Keyur Patel, AI adoption advisor and corporate Claude trainerPublished 4 min read
At a glance
- Client
- Shivalik Group
- Industry
- Real estate sales, Ahmedabad
- My role
- Trainer and programme designer
- Format
- Two sessions, two weeks apart
- When
- 25 July and 8 August 2026
- Teams
- Sales
40+
People trained
2
Sessions
6
AI tools taught
On this page
What problem were they solving?
Shivalik’s IT team had already been through ten sessions, and the group asked for Sales next. A sales floor is a different audience: no code, and work made of pitches, follow-ups, site-visit notes, proposals and client calls.
Most of the team had used ChatGPT occasionally, but nobody had a way of working that held up in front of a client. In real estate that matters more than usual. An AI that confidently invents a price, a possession date or a RERA number is worse than no AI at all.
What did the training cover?
Session 1: AI foundations for sales
What a language model actually does, which of ChatGPT, Claude and Gemini to reach for and when, and a five-part prompt practised on Shivalik’s own projects. The demo turned a client’s past emails and notes into a one-screen HNI call-prep sheet.
Session 2: Claude Cowork and Claude Design
Chat to think, Cowork to produce. The team watched Claude take one real enquiry end to end: research, lead score, call brief, an on-brand one-pager from Claude Design, and a follow-up draft, with a person approving every step.
A lead-scoring model built for Shivalik
Engagement, fit and urgency, each scored 0 to 25, minus a penalty for anyone contacted in the last day. It feeds a daily call-card: who to ring, what they care about, and the one question to ask.
Rules for anything a client will see
Manual approval for client work. AI drafts and a person sends. Every price, RERA number and possession date is checked against source, and missing figures are marked to confirm rather than guessed.
What changed?
- 40+ people on Shivalik Group’s Sales team trained in two sessions, two weeks apart.
- A five-part prompt (role, context, task, constraints, format) practised on Shivalik’s own projects.
- One real enquiry taken end to end in Claude Cowork, from research to an on-brand one-pager and a follow-up draft.
- A Shivalik lead-scoring model and a daily call-card, plus a prompt library and a shared Sales Assistant set-up.
- House rules for client work: manual approval, a person sends, every price and RERA number checked.
| Item | Session 1 | Session 2 |
|---|---|---|
| Date | 25 July 2026 | 8 August 2026 |
| Focus | AI foundations: models, prompting, choosing a tool | Claude Cowork and Claude Design |
| Tools | ChatGPT, Claude, Gemini, NotebookLM | Claude Cowork, the Sales plugin, Claude Design |
| Live demo | An HNI call-prep sheet from a client’s emails and notes | One enquiry end to end, from research to follow-up |
| Rules covered | Verify facts, protect client data, keep a human in the loop | Manual approval for client work, never auto-send, verify every figure |
Swipe the table sideways to see every column.
“He kept the sessions simple, practical, and closely connected to the realities of real estate sales.”
Read the full feedbackHide the full feedback
Everyone is talking about AI today. But knowing about AI and knowing how to use it effectively in day-to-day sales are two very different things. Keyur’s AI training made that difference for our sales team at Shivalik Group. He kept the sessions simple, practical, and closely connected to the realities of real estate sales. Instead of focusing only on AI tools and features, he showed us how AI can support our actual work: from better customer communication and follow-ups to lead engagement, sales preparation, research, and everyday productivity. What I particularly appreciated was his focus on practical application. The sessions helped us identify where AI can save time, improve the way we communicate with prospects, and rethink some of our existing sales processes. The training was relatable, interactive, and focused on real-world situations our team encounters every day. We walked away not just knowing more about AI, but with practical ideas we could start applying in our respective roles.
How does this fit the wider engagement?
This track exists because the IT sessions landed. Shivalik asked for Sales next, which meant someone inside the group had to argue the value again to a team with completely different work.
The Sales track is now complete, and a second Shivalik Group entity is onboarding. The full story, from the first IT session to the group-wide rollout, is in the companion case study.
Read the companion case study: From ten IT sessions to a group-wide AI rollout
Common questions
Can you train a sales team that is not technical?
Yes. Shivalik’s Sales track used no code and no jargon. It was built around pitches, follow-ups, call prep, proposals and site-visit notes, and on the judgement of when an AI draft is good enough to send to a client.
What is Claude Cowork, and why teach it to sales?
Cowork is a mode in the Claude desktop app where you describe a finished outcome and Claude does the steps: researching a prospect, building a brief, drafting the follow-up. For sales it removes the work around the work. I teach it with manual approval, so a person signs off every step.
How do you stop AI inventing prices or approvals?
By grounding it in the team’s own material, with one Claude Project per development holding its price list, floor plans and RERA details, and by a hard rule: every figure is checked against source before a client sees it, and anything missing is marked to confirm.
Which AI tools did the sales team learn?
ChatGPT, Claude, Gemini and NotebookLM in the first session, each matched to a kind of task. Claude Cowork, the Sales plugin and Claude Design in the second, for whole tasks and on-brand collateral.
Programmes behind this work: AI Training for Real Estate, AI Training for Sales Teams, Corporate AI Training for Employees. Training in this city: AI training in Ahmedabad.