Alan Gurung
Advisory AIJake
TimelineIn this session, Alan Gurung from Advisory AI joined us to explore what the next 12 to 16 months of AI in financial advice really looks like. The demo-led webinar focused on Atlas, Advisory AI's conversational AI layer that sits across your entire tech stack and transforms how advice firms handle everything from client prep to compliance. If you missed it, the full recording is below.
What we covered
The session covered the one concept Alan believes defines the next era of AI in advice: context. From there, a live demo walked through Atlas in practice, before an open Q&A tackled data security, integrations, pricing, and the most common mistakes advice firms make when adopting AI.
Why context is everything
Why most AI tools miss the biggest opportunity, and why context is the real competitive edge for advice firms.
Atlas in action
A live walkthrough from generating a daily to-do list to drafting suitability letters and running compliance checks.
Three common AI mistakes
The mistakes advice firms most often make when adopting AI tools, and how to avoid them from day one.
Data security answered
Where client data is stored, how it is protected, and what certifications to ask any AI provider for.
Integrations and legacy systems
Which back office systems Atlas connects to today, and an honest take on the legacy provider bottleneck.
The 75% time saving
Real case study data, including a firm that doubled adviser client capacity within 12 months.
The one word that defines where AI is heading
Alan opened with a provocation: the AI race is not being won by the fastest model, the cheapest model, or the most accurate one. It is being won by whoever has the most context.
The analogy he used landed well with the audience. Imagine hiring the smartest financial planner in the world, but every single morning they wake up with no memory of your firm, your clients, your investment philosophy, or how you like to communicate. That is what most AI tools are today. You have to re-explain everything, every time.
"The future of AI is not won by those who are quicker, those who are more accurate, or those who are cheaper. It is won by those that have the most amount of context."
The data that advice firms already hold across their CRM, cash flow tools, platforms, meeting notes and compliance portals is largely unstructured. You can push it and pull it between systems, but you cannot interrogate it. Atlas is built to change that by bringing all of those data sources into a single conversational interface.
What the Atlas demo showed
The live demo walked through a full client journey, from the start of the working day through to post-meeting compliance, without leaving a single interface. Here is what that looked like in practice.
Morning to-do list. Atlas scanned the entire client base, cross-referenced meeting history and CRM data, and produced a prioritised list of actions split by high, medium and low importance. The adviser did not have to prompt it with a client name or a task. Atlas already knew what needed doing next.
Personalised outreach. For a client flagged as overdue for a review, Atlas drafted a personalised email that referenced previous meeting details, including a personal note picked up from the last conversation. The resulting email did not read like a template.
Pre-meeting pack. Once the client accepted the meeting invitation, Atlas automatically generated a preparation document covering agenda, talking points and relevant client data. No prompt required.
Post-meeting workflow. After the meeting recording was processed, Atlas logged the session in the CRM, identified required next steps, drafted a combined suitability letter, and passed it to a compliance AI model called Colin, built against FCA handbook and Consumer Duty guidelines. Compliance flagged a missing item, and Atlas revised the document automatically.
"It now becomes a whole end-to-end workflow for your whole company. When the letter is created, it notifies the paraplanner. When it goes to compliance, it notifies the compliance officer. And if it passes, it moves to the next person."
Workflow automation. In a brief demonstration of the workflow builder, Alan created a trigger in plain English: when a client reaches age 55, assign a pension review and notify the adviser. Atlas produced the complete automated workflow from a single sentence.
Three mistakes advice firms make when adopting AI
Asked for the most common pitfalls, Alan gave a clear and direct answer.
Starting from the solution, not the problem
Shiny object syndrome is real. Firms that start with "we want to use AI" rather than "we want to solve this specific problem" find it much harder to demonstrate value. Start with a bottleneck. Work backwards from what is limiting your growth, then ask whether AI can address it.
Not checking integrations carefully enough
It is not enough for an AI tool to pull data from your systems. It also needs to push data back. If your AI cannot write back into your CRM, context is lost every time. Without that two-way flow, the AI has no memory of your firm and the value drops significantly.
Making one person more efficient instead of the whole firm
Meeting note AI was a good start, but for many firms it just created more work downstream for paraplanners and compliance officers. The problem moved rather than being solved. The goal has to be whole-firm productivity, not just one role becoming faster.
Data security: what to ask every AI provider
Several attendees raised data security questions, which Alan addressed directly. Client data within Advisory AI is stored on AWS infrastructure, within an S3 bucket held in the UK, with backup storage in the UK and Northern Ireland. This keeps the data within the required GDPR framework.
"Look at their certifications. Things like ICO registration, ISO 27001, Cyber Essentials, and pen testing. These are done by third parties. If a firm cannot show you independent certifications, that is a red flag."
Integrations: where Atlas is today and where it is heading
Atlas currently integrates with Intelligent Office (IO), Planner X, Xplan, and Curo, covering the most common back office systems in UK advice firms. Around 100 data fields are accessible through the IO integration, though Alan was candid that IO itself holds 400 to 500 fields, meaning some data remains out of reach for now.
If back office systems do not improve their API access at pace, firms may increasingly choose to hold client data directly within Advisory AI rather than a legacy CRM. The long-term answer, Alan argued, is deeper integration partnerships.
Looking ahead, Advisory AI is in conversations with FE Analytics to enable direct fund data queries through Atlas, projected around September to October. An integration with Mabel Insights, focused on model portfolio data, is expected around July.
The numbers from firms already using Atlas
For those asking what the return actually looks like, Alan pointed to two case studies from current clients.
In a study with Brooks Macdonald, advisers reported a 75% reduction in time spent on weekly administration compared to before adopting Advisory AI.
In a second firm, the average number of clients an adviser could manage increased from 100 to 200 over a 12-month period. That is not just efficiency. It is a direct impact on revenue capacity.
