Sun Global's AI transformation starts where the revenue is made.
IA generates the cash flow, and IA is where Sun Global's growth is bottlenecked. You scaled the team down because advisors couldn't bring in enough flow. The real constraint isn't the advisors. It isn't client onboarding either. It's the systems around them, pulling Amit and Hitesh away from what they do best: building relationships and structuring the right investments. Get that right, and Amit's relationship capacity scales. More clients, served better. AUM growing at the source.
Prepared for
Dhairya Kapadia
Company
Sun Global Investments
Prepared by
Alex Girardet & Alex Garcia Chicote
Date
17th April 2026
What We Heard
There is no single source of truth for a client. What a client was promised, what they care about, what they've said no to, what's due when. Everything is split between Amit, Hitesh, operations, and a dozen different systems. Follow-ups get missed. Context gets lost between handoffs. Every interaction starts with someone piecing together what happened last time. And the number of relationships Amit and Hitesh can carry at a high standard is capped by what they can personally hold in their heads.
A CRM was supposed to solve this. One place for the full picture of every client relationship. But it came at too high a cost: every call, every email, every WhatsApp message had to be manually logged back into the system. The people who needed the CRM most were the ones least able to maintain it, and the team went back to Excel.
"My biggest issue was the number of follow-ups that I have to do. I need those note-takers to then put it into me and tell me: these are the things, to-dos that I need to do for that particular person. And these are the deadlines. Fix up my diary. Fix up my follow-ups."
— Amit
"We actually send those ideas to everybody. But what would help is that if five out of those 50 would have told that if this kind of idea is there, then let us know. Rather than calling all 50, you can call up only those five."
— Hitesh
The fix isn't just a CRM. It's a shared intelligence layer: a structured, living record of every client relationship, kept alive by AI assistants that absorb the admin, surface the context, and free advisors to focus on what actually generates revenue. The team doesn't change how they work. The system wraps around them and builds the picture as a by-product of the work already happening. The team stops feeding the system. The system starts feeding the team.
Phase 1 builds both: the data foundation in Attio, and a personal AI assistant for Amit that manages it on his behalf. When it's running, Amit walks into every meeting with full context, no follow-up slips, and ideas land with the clients who actually want them. What he can hold in his head stops being the ceiling on how much AUM he can carry.
Phase 1: What You Get
Phase 1 is four deliverables. Each is self-contained, each builds on the one before, and each is invoiced only after it's live and you've signed it off.
01
Cleaning the Data
The dirty but essential work. Before the assistant can do anything useful, it needs something to work from. We consolidate all of Sun Global's scattered relationship data into one place: Attio, the most AI-native CRM on the market.
Excel files, email threads, the old CRM, per-deal project files, prospect lists. All migrated, cleaned, and structured around how your IA team actually works: clients, prospects, interactions, preferences, deal interest.
Attio is ~£29/user/month. It's UK-registered, ICO-registered, SOC 2 Type II and ISO 27011 certified.
Outcome: This is the first step in regaining control of Sun Global's relationship data. Cleaning it, organising it, and storing it properly. The foundation everything else builds on. Amit's assistant has something to work from on day one rather than starting cold.
02
Amit's Personal Assistant
Amit gets a personal AI assistant with its own WhatsApp number, connected to his Outlook email and calendar, and to the CRM. He can talk to it, ask it questions about his clients, his calendar, his emails, and instruct it to update the CRM on his behalf. One place to go instead of juggling five systems.
WhatsAppOutlook (email + calendar)Attio CRM
Outcome: Amit has a personal assistant that knows his calendar, his emails, and his client data. If he wants something in the CRM, he tells the assistant.
03
Building Intelligence Without the Burden
The assistant becomes proactive. Instead of waiting for Amit to ask, it keeps the CRM alive on its own.
Recorded call logging: Client calls are already recorded. The assistant ingests the transcript, extracts the relevant notes, and drafts a CRM entry for Amit to approve.
Email & calendar capture: The assistant silently reads Amit's Outlook and logs client interactions against the right contact in the CRM. Fireflies meeting summaries get pulled in too, instead of sitting unread in his inbox.
Voice note debrief: For what the assistant can't observe (in-person meetings, unrecorded calls, WhatsApp chats) Amit sends a voice note afterwards. The assistant transcribes it, structures it and asks Amit for approval before logging into the CRM.
Compliance audit trail: Whenever any of the above touches a trade, a piece of investment advice, or a recommendation, the assistant drafts the FCA-required confirmation email for Amit to review. The paper trail stays maintained automatically. The compliance pattern doesn't change, it just becomes effortless.
Advanced client profiling: As interactions are logged, the assistant builds a structured profile for each client: risk appetite, return expectations, asset class interests, historical investment decisions, and reasons for saying no to past ideas. No more fragmented Excel sheets or relying on memory. These profiles compound over time and power the idea distribution in Deliverable 04.
Client context at your fingertips: Amit asks "what's the latest with Steve?" and gets the full picture: recent conversations, pending actions, preferences, history. One question, complete context.
Built-in protections
Sensitivity filter. The assistant is built to exclude confidential and personal information (health, family matters, anything not related to investments) from CRM entries.
Approval gate. Before anything is stored in the CRM, the assistant sends Amit a summary for approval. AI is powerful but not infallible. Amit is the final gate, so nothing slips through.
Outcome: The CRM stays current without anyone carrying the burden. Amit walks into every conversation with full context — and when Hitesh joins in Phase 2, the same layer serves both of them. What starts as Amit's personal assistant becomes Sun Global's shared intelligence layer.
04
Taking Action: An Assistant with Agency
Once the CRM has context, the assistant stops being a data store. It starts surfacing the right move to Amit at the right time. The follow-up to make, the client to call, the idea to send.
Follow-up engine: After every interaction, the assistant extracts action items, deadlines, and commitments. It surfaces them on Amit's timeline — ahead of upcoming meetings, at end of day, first thing in the morning. Nothing falls through the cracks.
Tailored idea distribution: When an investment idea is ready, the assistant matches it to the clients whose profiles would be interested. Spray-and-pray stops.
Prospect reminders: Amit's 30+ active prospects live in the assistant. It nudges him when one goes cold, or when something from a previous conversation — a planned trip, a fund closing, a fiscal deadline — makes now the right time to reach out.
Daily market snippets: The 30 minutes Amit spends every morning preparing the daily market summary becomes a 2-minute review. The assistant drafts it overnight, tailoring where possible to client interests, and lands it in his inbox ready to go.
Outcome: These capabilities get smarter over time as the CRM fills with context. The more Amit uses it, the better the matching and suggestions become. It's a flywheel. The value compounds.
Timeline
Week 1
Deliverable 01: Cleaning the Data — Scoping session with Amit, walk through what client data exists and what's worth migrating. Attio CRM setup, data model design, and migration executed in parallel.
Week 2
Deliverable 02: Amit's Personal Assistant — By mid-week, Amit has his assistant live on WhatsApp, connected to Outlook and the CRM.
Week 2 – Week 3
Deliverable 03: Building Intelligence Without the Burden — Recorded call ingestion, email logging, voice note debrief, compliance audit trail, client context on demand. Amit approves what gets stored.
End of Week 3
Deliverable 04: Assistant with Agency — Follow-up engine, idea distribution, prospect reminders, daily market snippets. Calibrated on real data.
Week 4
Testing, iteration, and Phase 2 scoping — Refine based on Amit's real usage, address edge cases, and discuss Phase 2 priorities with Dhairya to maximise ROI.
Investment
Implementation (one-off)
01. Cleaning the DataData migration, CRM setup, data model design
£1,000
02. Amit's Personal AssistantWhatsApp, Outlook, CRM integrations and base agent setup
£800
03. Building Intelligence Without the BurdenCall logging, email logging, voice note debrief, compliance audit trail, client context
£2,250
04. The Assistant has AgencyFollow-up engine, idea distribution, prospect reminders, daily market snippets
£1,700
Total
£5,750
Ongoing (monthly)
Attio CRMPlus plan, ~£29/user/month
£29 / month
WhatsApp Business number
£10 / month
Agent serverFixed hosting cost
£12 / month
AI running costsLanguage model API usage
£90 / month
Alavida support & maintenanceUpdates, monitoring, improvements, support for the team
£30 / month
How you pay
Pay per milestone. Nothing upfront. Each of the four deliverables is invoiced only after it's live, you've used it, and you've signed it off. If a milestone doesn't meet your standard, we fix it before we invoice — and we don't move on to the next until you're satisfied with the one in your hands.
You control the pace. At every checkpoint, you decide whether to continue. The risk of this transformation sits with us, not with you.
What We Need From You
01 Scoping session with Amit (60–90 min) to walk through existing client data and what's worth migrating
02 Old CRM export from Hetal — whatever can be extracted from the previous system
03 Outlook access for Amit (email + calendar) — We'll walk you through the setup
04 Access to Amit's call recording system — We'll need to understand what software you use and how to integrate
05 Access to Fireflies account to create the integration with the assistant
06 In-person session to integrate Bloomberg — We'll need access to the terminal to connect market data to the assistant
Phase 2: To Be Discussed
Once Phase 1 is running and Amit is using the assistant daily, we extend the foundation. As you suggested, Dhairya, we'll discuss areas of priority together to maximise ROI.
Hitesh's Personal Assistant — Same pattern. Hitesh gets his own assistant, connected to his channels and the shared CRM. He sees Amit's relationship context; Amit sees Hitesh's recommendations.
Research Note Drafting — Hitesh identifies an investment idea, the assistant drafts the research note from publicly available sources. 2–4 hours of manual work, done in minutes.
Refusal Signal Capture — When a client says no to an idea, the assistant captures why: price, sector, risk appetite, timing. That data sharpens every future recommendation.
Deeper Capabilities for Amit — Portfolio analysis, reporting, and recurring analytical workflows, now that his relationship data is live in the CRM.
Phase 3: The Broader Transformation
With the full relationship layer in place and multiple advisors on the system, we work with you and Mihir to identify what comes next. The right sequencing falls out of what Phases 1 and 2 reveal.
Full Team Rollout — Extend the assistant foundation across the remaining team: operations, middle office, back office.
Content & Ideation Ecosystem — Mihir's vision: more content, more formats, delivered more consistently to clients and prospects.
Compliance & Onboarding — Automate the repetitive compliance workflows and onboarding handoffs, once client volume justifies it.
Client-Facing Agent — An assistant grounded in Sun Global's proprietary data that clients can engage with directly. A continuously-available channel beyond scheduled calls and email blasts.
Your Team
Alavida builds AI systems for real businesses. Before ChatGPT, we were already building intelligent data systems inside real companies — enterprise-grade data platforms, warehouses, reporting layers, ML models. The goal was always the same: extract value from a business's own data so it could run better products, better operations, and better decisions. When ChatGPT landed, that foundation turned out to be the thing most companies trying to deploy AI were missing — you can't bolt intelligence onto chaos. Since then we've been shipping AI into environments where the stakes are real: clinical data, enterprise sales, financial decisions. Not gimmicks, not demos. We know the AI landscape, and we know where it actually moves the needle inside a business. You're not paying us to learn on your problem.
Small team. No account manager, no handoff. If something breaks, the person who built it picks up.
Alex Girardet
AI Architect
Builds AI infrastructure at scale, across enterprises and early-stage companies. Finance degree backing a practitioner's eye for where AI unlocks commercial value, not just what's technically possible.
Builds AI across industries high stakes industries. Brings both sides: the consultant's instinct for delivering what business's need, and the in-house engineer's muscle for building systems that last.
