Bank AI
Get the app
Investing7 min read

Holding investments at more than one broker is now the norm rather than the exception. UK investors routinely have an ISA at one platform, a SIPP at another, a Trading 212 or Trade Republic account for fractional shares, and crypto on Coinbase or Binance. EU investors stack PEAs, assurance-vie wrappers and Depots in similar combinations. The hard bit is seeing all of it on one screen. Here is the modern toolkit for doing that, and the trade-offs to be honest about.

Why it is hard in the first place

Two reasons. First, brokers do not have an incentive to make their data easy to leave - the longer you stay inside a single platform, the longer you pay them. Second, the data shapes are different across platforms. An ISA, a SIPP and a General Investment Account at the same broker have different tax framings; an assurance-vie at Linxea looks nothing like a Depot at Comdirect; a Bitcoin wallet on Binance is not denominated in the same currency as a fund holding on Saxo.

Pre-2018 the answer was a spreadsheet, fed by manually downloaded CSVs. That works, until it does not. The two failure modes are: (a) you stop updating it because it takes 30 minutes; (b) it lies to you because of stale prices, missed dividends or a broker change you forgot to apply.

Three categories of tool today

In 2026, you have three real options for building a multi-broker dashboard.

  • CSV-fed portfolio trackers. Sharesight, Snowball Analytics, Parqet (DE). You upload statements and the tool reconciles. Strong on tax reports; weak on freshness because the data is only as live as your last upload.
  • Broker-API portfolio trackers. Increasingly, brokers expose APIs that third-party trackers can plug into directly. Faster than CSV; coverage depends on which brokers the tool has integrated.
  • Open-banking aggregators. PSD2-licensed AISPs (like Bank AI's partner Powens) connect to brokers as well as bank accounts. Daily refresh, no CSV uploads. Coverage depends on the AISP's connector inventory.

Choosing the right approach

If your priority is tax reports - capital gains, dividend reports, cost-basis tracking that an accountant can use - the CSV-fed trackers are still the strongest tools in their category. Sharesight in the UK and AU, Parqet in Germany. Pay the subscription, upload monthly, and treat them as your tax-prep workspace.

If your priority is daily visibility - knowing where your money is and what it is doing without spending an hour a month reconciling - open-banking aggregators are now the better fit. The data is live, the maintenance is zero, and they cover bank accounts and crypto in the same view.

If you have both priorities, the honest answer is to use both tools. They overlap on portfolio visibility and diverge on tax-report depth. Bank AI is in the second category; we are clear about not replacing Sharesight or Parqet for tax reports.

What a good dashboard surfaces

Once you have all your data flowing into one place, the dashboard should answer four questions cleanly.

  • Net worth. Total wealth in your home currency, broken down by cash, investments, pensions and crypto. The single most useful number you have - and the one most apps butcher by hiding it behind dashboards.
  • Allocation. By asset class, by geography, by currency. The geography one matters more than people think: a UK ISA, a US-listed Vanguard fund and a Trade Republic ETF can quietly all be 65% US tech.
  • Fees. Aggregated across all platforms. Fund TERs, platform fees, FX spreads. This is where the multi-broker view earns its keep - the fees you pay across three brokers are easy to lose visually if you only ever look at each in isolation.
  • Tax wrappers. ISA contribution remaining, SIPP allowance, PEA cap, assurance-vie 8-year clock, Depot Sparer-Pauschbetrag. The wrappers are the where, not the what; a good dashboard treats them as first-class metadata.

The case for AI on top

A static dashboard answers the four questions above. An AI assistant answers the next layer of questions, the ones that previously required a financial planner: 'how much am I actually paying in fund fees this year?', 'how exposed am I to US tech across all my brokers?', 'where am I bleeding money on FX spreads?', 'when is my next dividend hitting?'

The reason this is interesting now and was not five years ago is that the underlying data layer - PSD2 open banking - is finally good enough that you can ask plain-English questions of your portfolio and get answers grounded in your actual positions, not in generic market commentary. The combination of clean data and a capable language model is what makes 'an AI you can ask' a useful product surface and not a gimmick.

If you are building a multi-broker dashboard in 2026 and you are not at least testing this lane, you are leaving the most valuable layer on the table.

Keep reading