I just finished reading LSEG's new report, The Blueprint for Growth: AI's Role in Wealth Management. LSEG sponsored it, and ThoughtLab ran the survey late last year across 500 wealth firms in North America, Europe and Asia Pacific. I went in expecting a standard vendor white paper. It's better than that. There's real data in it, and I agree with most of what it says.

But I kept waiting for the one chapter that never showed up. That's what I want to talk about today.

What stood out first

The number everyone will quote is that 73% of wealth and asset management executives say AI is critical to the future of their business. That didn't surprise me. What did surprise me was how little most firms are getting back. More than two-thirds report small or moderate returns. Twelve percent either see nothing or lose money. The average payback period is about two years. The firms LSEG calls leaders average a 4.7% ROI, compared with 3.1% for everyone else.

So the industry is convinced AI matters, and most firms haven't figured out how to make it pay. LSEG blames fragmented data, siloed systems, cultural inertia and regulatory uncertainty. I think that's right. I would have listed the same things.

I also liked how the report frames the future. It says wealth management won't be advisor-led or self-directed. It will be "advice-powered." That's close to how we talk about it at Datalign. The advisor stops being the person who gathers information and becomes the person who makes sense of it for the client. LSEG calls that an insight orchestrator. I'd probably just call it the job most advisors got into this business to do.

Where I think it falls short

The title says growth. But when you look at where the report points AI, it's mostly inside the firm: portfolio construction, regulatory monitoring, document summaries, code automation, and deeper analysis of existing clients. Almost all of it is about clients the firm already has.

There's very little about how you win new ones. If you're a big wirehouse with a national brand, maybe that's fine. If you run an independent RIA, getting in front of the right new households is usually the hardest problem you have. The report itself says the leaders focus on use cases with measurable impact, and I can't think of anything more measurable than a new client relationship that's a good fit.

That gap is a big part of why Datalign exists. So let me walk through how we've approached it, piece by piece.

1. AI lead generation

Thirty percent of the firms LSEG surveyed say unclear ROI is a major challenge, and 28% point to the cost of implementation. I understand why. A lot of AI's benefit shows up as a few minutes saved here and there, and nobody can put that on a board slide.

Client acquisition is different, because you can count it. Our lead generation uses machine learning models trained on a knowledge graph of more than 200 million Americans to find consumers who are actually looking for advice. Then we match on fit: the advisor's expertise, geography, specialization and communication style. So far we've referred more than $80 billion in assets to over 13,000 vetted fiduciary advisors and served more than 100,000 consumers in all 50 states.

LSEG argues that ROI should include client retention and engagement, not just cost savings. I agree. I'd just add that retention starts with whether you matched the right client with the right advisor in the first place.

2. RelationshipAI

My favorite section of the report describes advisors getting useful context at the exact moment they're reviewing a household or preparing for a meeting. But if you read it closely, it assumes the household is already a client, with years of CRM notes and emails for the AI to work through.

The first meeting is where advisors have the least context, and it's where context matters most. That's why we built RelationshipAI. It gives advisors a briefing on each prospect that they can read in about 30 seconds, and it saves them more than 10 hours of research a week. The number I care about most is that early adopters were over 16% more likely to turn a first conversation into a qualified opportunity.

Put simply, the software does the homework so the advisor can spend the meeting listening. According to LSEG, that kind of listening is what clients value most anyway.

3. GEOsAI

LSEG's first recommendation is to make AI part of business strategy. I suspect most firms will read that as "go write an AI roadmap." I read it more literally. Use AI to help make the big strategic decisions.

Opening a new office is one of the biggest bets an advisory firm makes, and a lot of those decisions still come down to gut feel and where a partner happens to live. GEOsAI looks for underserved markets where demand for advice is high and competition is low, and it gives RIAs a data-driven plan for where to open next. It runs on the same knowledge graph. It's a narrow tool with a clear payoff, which is what the leaders in LSEG's data seem to prioritize.

4. HaloAI

The report says clients want advice that's more personalized, more proactive and available whenever they need it. It also lists principles for responsible AI: accurate, explainable, accountable, safe and private. Those two lists pull against each other, and I don't think our industry talks about that enough.

I recently wrote about PensionBee's research showing that many people act on a chatbot's financial advice without checking it, including on decisions like when to retire. That's what happens when people want help around the clock and the only thing available is a general-purpose chatbot with nobody accountable behind it.

HaloAI is how we've tried to solve that. It's a consumer wealth platform that coordinates specialized agents across tax, retirement, investing and more, and it connects people with fiduciary guidance, with an advisor overseeing the process. I don't want HaloAI to replace the conversation with an advisor. I want people to walk into that conversation already understanding their own situation better.

5. Constellation

There's a finding in the report I think people will skip past. The leaders are much more likely to rely heavily on SaaS vendors, 43% compared with 28% for other firms, and more likely to build a network of technology partners. Meanwhile only 9% of firms use agentic AI today, a number expected to double in three years, and 31% of technology leaders say regulatory uncertainty is a major hurdle.

When I put those together, I see a problem for independent firms. The largest firms may build their own AI agents. Most RIAs can't afford the engineers or take on the compliance risk to do that alone.

That's what Constellation is for. It lets advisory firms of any size deploy their own branded AI agents, built on the firm's investment philosophy, content and client data, inside a compliance framework designed to meet or exceed SEC requirements. We also made a deliberate choice about where these agents sit. Most AI in our industry points inward, at back-office work and advisor workflows. Constellation agents face the client and carry the firm's approach into every interaction. If a client is talking to your firm's AI, it should sound like your firm, and it should be compliant from the first day, not after a regulator starts asking questions.

6. The knowledge graph

LSEG saves its strongest point for data, and I'm glad it does. About half the firms surveyed still don't have basic processes to clean, standardize and tag their data or to bring in trusted outside sources. The report's warning is that AI takes bad data and amplifies it, which leads to hallucinations, drift and bias. Its advice is to fix the data first.

I agree completely, and it's the reason our knowledge graph sits underneath most of what we build. Our lead models are trained on it, and GEOsAI uses it too. A knowledge graph stores people, places and financial circumstances as connected records instead of rows scattered across separate systems. That means a household means the same thing whether we're matching it with an advisor, sizing up a market or preparing a briefing.

The line I'd underline in the whole report is that getting the data right matters more than chasing the newest model. Models change every few months. A good data foundation keeps paying off for years. LSEG also says firms should be able to trace data back to its source and explain how any AI recommendation was produced. That's the right standard, and it should apply to vendors like us as much as anyone.

Where this leaves us

LSEG says firms have about 12 months to move from experimenting with AI to actually using it. Here's how I'd think about that choice. You can use AI to make your current work a little faster, and you'll probably end up with the two-thirds of firms seeing modest returns. Or you can point it at the things that decide whether your practice grows: finding the right clients, being ready for every first meeting, picking the right markets, and meeting higher client expectations without giving up fiduciary responsibility.

The report is right that the future of this industry is advice-powered. I'd just add that advice only helps if it reaches the right person. Much of the industry is adding AI to the back office, and that buys efficiency. But back-office automation has a ceiling—you can only save what you already spend on operations. Growth has no ceiling. Every funded relationship brings new assets and new fees that compound. Back-office AI automates the cost of running a firm. At Datalign, we automate what makes a firm worth more. For most independent firms, that's the part of the blueprint that still needs to be written.