Overlapping app windows on a trading desktop: CRM, AI assistant, analytics, chat, email, Bloomberg data and Excel, arranged on a grid

Trading Desktop Platforms in 2026: How Banks Build Multi-App Trading Workstations

Banks can build a front-office trading desktop without rebuilding their apps: a platform like io.Connect integrates the web, .NET, Java and legacy apps they already have.

Banks don’t build a trading desktop on one product, because none can do all the things a trader needs. They connect all the apps traders already use through a desktop interoperability layer, often based on the open FDC3 standard. interop.io’s io.Connect is built for this, and firms like Deutsche Numis and Stifel use it to make their trading apps share context.

A trader opens a few dozen applications daily and banks maintain hundreds of tools across departments. On a trader’s screen you will see Bloomberg in one corner, an order management system in another, EMS, the bank’s own pricing and risk tools, Excel, chat, a CRM, several communication tools like Symphony, Teams and Outlook. They come from different vendors and different internal teams, and almost none of them were built to talk to each other. Bloomberg identifies a bond one way, the order management system another, the pricer a third. The “platform” question is really about what sits underneath those applications and lets them share information.

Below, we compare the three main ways to build that platform: in-house, around one vendor’s ecosystem, or on a desktop interoperability platform.

Which firms have trading desks, and why do they need so many apps?

Banks (the sell side)

At an investment bank, trading usually lives in a Global Markets or Sales & Trading division, split into Equities and FICC (fixed income, currencies and commodities). FICC alone covers rates, credit, securitized products, FX and commodities, and each has its own desks, systems and clients. A salesperson on one of those desks works with market data, a pricer and the bank’s trading and risk platform. On top of that, client requests keep arriving from electronic venues, chat, voice and the CRM, often about the same client at the same time.

Asset managers and hedge funds (the buy side)

Portfolio managers decide what to buy and sell, and the firm’s traders execute it. Orders move from an order management system (OMS), which handles the order, compliance checks and allocation, to an execution management system (EMS) that works the order in the market. Around them sit research, analytics, transaction cost analysis and plenty of Excel.

Wealth managers

An advisor’s desk looks different but has the same shape: a CRM, portfolio data, financial planning tools, research and email, all about one client at a time.

What’s running on a typical trading desktop

CategoryWhat it’s forExamples
Market data, news and analyticsPrices, charts, news, research and analyticsBloomberg Terminal, LSEG Workspace, FactSet
Order and execution management (OMS/EMS)Managing orders and working them in the marketCharles River IMS, ION Fidessa, FlexTrade, Bloomberg EMSX
Trading and risk platformsPricing, trade capture, positions and riskMurex MX.3, Calypso
Electronic trading venuesClient requests for quote (RFQs) and electronic executionTradeweb, MarketAxess
CommunicationChat and voice with clients and colleaguesBloomberg IB Chat, Symphony, voice trading turrets
CRMClient coverage, interactions and follow-upsSalesforce
SpreadsheetsQuick pricing, analysis, and lists of orders and positionsExcel, often with data‑vendor add‑ins
In-house toolsPricers, blotters, axe lists and risk views built by the bank’s own teamsBuilt by quants, strats and desk developers

Many banks also run a long tail of older legacy applications, some written decades ago in programming languages few people still use. They also have big teams of developers that ship custom internal applications for better flexibility, ownership or cost cutting. You might expect the total number of apps to stay manageable, but at interop.io, we’ve worked with banks running close to 1,000 applications across teams. An interop platform should be able to work with all of them.

To add even more complexity, we now have AI systems you need to connect with. And we see that teams are shipping new apps or completely rewriting older systems much faster and cheaper.

Why is it hard to connect everything?

Each application has its own way of identifying a client, an instrument or an order. Nothing tells the CRM that the bond you just clicked in the terminal is the one the client asked about. A single desk can run web apps, .NET and Java applications, Excel with VBA, and older apps delivered through Citrix. Connecting a web app to a Java app on the same screen is not a standard web-development task. Some even come from vendors or teams that no longer exist, and some have no APIs at all.

An integration platform as a service (iPaaS), an enterprise service bus (ESB) or a set of APIs can move data between back-end systems. IBM describes iPaaS as self-service, cloud-based tools that integrate applications, systems and data sources across IT environments. A trade booked in the OMS flows to settlement. A client in the CRM can be synced with the ERP, or a document can be sent to the right department. But it can’t connect the applications on one user’s screen, in real time, while the user works.

That’s where you need a desktop interoperability platform – when a salesperson selects a client in the CRM, the blotter, the chat window and the pricer will show that client. When a trader clicks an order, the chart, the risk view and the analytics follow. The apps share context (the selected client, instrument or order) and actions (like “show me a chart for this”) at the moment the user acts. It also keeps the user’s windows and workspaces in order and gives the bank’s platform team one place to manage it all.

And a very tough problem – who governs everything? At banks, every change, permission and connection is regulated and audited.

What it looks like in practice: real-world examples

1. Sell-side equities: the OMS and market data follow the same stock

A sales trader works in Fidessa and checks the same stock in a market data app and in Bloomberg. The io.Connect Fidessa adapter maps Fidessa’s tracking groups to io.Connect or FDC3 channels, so the stock selected in Fidessa is shared, and the market data app and the Bloomberg Launchpad group (linked to the same channel) can follow it. Nobody types the ticker twice.

Deutsche Numis linked Fidessa and Infront this way, “delivered in just five days.” The same adapter works for order data: Stifel merged its European order book from six separate Fidessa instances into one view that sales traders use to create IOIs.

2. From a chat message to a prefilled ticket

Requests for quotes still arrive in chat. With the Bloomberg IB Chat adapter, a trader right-clicks a message to send it to internal systems, for example, to prefill an order ticket or start an RFQ. Firms can add a review step before anything is submitted and capture only happens when the user asks for it. When the trade is booked, a notification can carry a button that opens it in the blotter, because io.Connect notifications can call any registered interop method.

3. The advisor desktop: select a client once

Wealth managers connect CRM (for example Salesforce Financial Services Cloud), portfolio management and planning tools into one advisor workstation. The advisor searches for a client, a client workspace opens, and the portfolio, activity and CRM views follow. The io.Connect Salesforce adapter can sync the selected Contact and its Account (in beta), and the Outlook adapter can draft the follow-up email.

Which framework or platform should you use to build a multi-app trading desktop?

Banks usually choose among three approaches: build in-house, center the desk on one vendor’s ecosystem, or adopt a desktop interoperability platform such as io.Connect. The right choice depends on how many apps, technologies and vendors the desk has to connect.

1. Build it in-house

Some teams keep their apps in the browser, and others build natively. If you build a desktop yourself, you’ll at least want a framework such as Electron, WebView2 or Tauri. Electron packages a full Chromium browser with every app, while WebView2 (on Windows) and Tauri use the web view the operating system already provides. These are generic frameworks that were not designed as a trader’s desktop, so they cover only a small part of the job and add their own overhead.

Things get complicated once a desk runs more than one application: there’s no built-in way for your app to share a selected client with Bloomberg, a Java order management system or Excel. So you need a message bus between apps, layouts that restore across monitors, an app catalog with entitlements, configuration per desk, logging and remote support. Each of these has to be maintained and updated every time a new app arrives, and that becomes a burden over time.

You can also build it native with Qt or WPF/.NET. Qt (C++) is used where raw rendering speed matters, and it has built-in docking for multi-panel layouts. If most of the software a bank uses is WPF and .NET, this may work initially. But you also lock each app to its technology, and still leave the question of how a Qt pricer, a WPF blotter and a web-based CRM work together on the same screen.

2. Build around one vendor’s ecosystem

Another route is to make one large vendor platform the hub, typically the market data terminal or the order and execution management system, and connect other tools through that vendor’s APIs and integrations. It’s possible if most of the desk’s work happens in that one product and the number of other apps is limited. They also must have APIs or other ways to connect. It gets harder when the desk depends on systems from several vendors plus in-house or legacy apps, which most do, and it ties the desk’s workflow to that vendor’s roadmap.

3. Use a desktop interoperability platform

You can also stop choosing between technologies. Keep the applications the desk already relies on, whether Electron, Qt, WPF, Java, C#, Excel or vendor software, and put a platform underneath that connects them. Each app connects once, through the open FDC3 standard or a client library, and from then on, it shares context and actions with every other app on the desk. The platform also handles the windowing around the apps: workspaces and layouts, an app catalog with entitlements, configuration, monitoring and support. That’s what io.Connect does.

ElectronWebView2 inside .NETNative Qt or WPF/.NETOne vendor’s ecosystemio.Connect
Best forOne web-based app on the desktopAdding web UIs to an existing .NET desktopPerformance-critical native screensDesks that live mostly in one vendor’s productDesks that mix vendor, in-house and legacy apps
Connecting other appsYou build itYou build itYou build itLimited to what that vendor supportsBuilt in: FDC3, client libraries, third-party adapters
TechnologiesWeb.NET plus webC++ or .NETThe vendor’s APIsWeb, .NET, Java, VBA, Delphi, C and more
Workspaces and layouts across appsYou build itYou build itWithin one appWithin the vendor’s productBuilt in
Governance and supportYou build itYou build itYou build itThe vendor’s tools, for its own productCentral app catalog, configuration, telemetry, remote support available
Runs in VDI / CitrixWith some tuningYesYesDepends on the vendorYes, with Citrix app support

io.Connect: the platform to build, run and experience the trading desktop

A trading desktop involves three groups of people. Application engineers build the apps, a platform team manages and governs them across the bank, and traders, salespeople and advisors get the connected experience. io.Connect is the infrastructure for all three, and AI assistants use the same layer, limited to the actions a firm has permitted.

For app developers

  • One interop layer on the user’s machine. Apps share context, link through channels, call each other’s methods, subscribe to streams such as market data, and raise intents. Connect an app once and it can talk to others in different ways.
  • FDC3 built in. io.Connect Desktop and io.Connect Browser fully support FDC3 2.0 and 2.2 respectively. interop.io is an active contributor to the FDC3 standard (Julianna Mealin is an FDC3 maintainer), and io.Connect Browser 5.0 is one of the three desktop agents FINOS lists as conforming to FDC3 3.0.
  • The technologies banks actually run. Client libraries cover JavaScript, React, Angular, Node.js, .NET (including WPF and WinForms), Java (including Swing and JavaFX), VBA, Delphi, Blazor, Python Dash and C exports. Citrix and ClickOnce apps are supported app types.
  • Third-party vendors. Adapters connect Excel, Outlook, Word, Teams, Bloomberg, Bloomberg IB Chat, Fidessa and Salesforce.

For platform teams

io.Manager is the platform team’s control plane:

  • Who gets which app. A central app store marks each app as public or restricted to groups, which can come from the bank’s identity provider.
  • Changes by desk. Configuration is served per platform version, group and user, so a change can go to one desk first. Layouts and workspaces are stored and shared centrally.
  • Support without a desk visit. Support can pull a user’s logs or restart their desktop remotely. Crash reports and user feedback arrive in one place, and an audit log records changes.
  • Least privilege for the platform itself. io.Manager 4.0 has per-resource, per-action permission groups, so a support team can inspect sessions while only the platform team changes apps or configuration.

io.Insights sends OpenTelemetry metrics, traces and logs to the monitoring tools the bank already uses, such as Grafana or Prometheus.

UI and UX

  • Workspaces arrange apps from different teams and vendors in one window, and save and restore them with their context.
  • Layouts restore a trader’s whole multi-monitor desk at startup, and can be shared with colleagues and support.
  • Window groups, tabs and a launcher with search keep dozens of windows manageable and let users find apps, workspaces, clients and instruments in one place.
  • Notifications from every app arrive in one panel, and trigger actions in other apps based on rules.

Governed AI

Once an app is connected, AI assistants can use it through the same layer. io.Intelligence turns the methods and intents that apps already register into tools the assistant can call, by configuration rather than a rewrite, with allow-lists for which apps a tool may reach. It also works with the context across the user’s whole workflow, not just one window: the selected client or instrument, and the apps connected to that workflow, even ones the user doesn’t have open right now. Because every tool the assistant can call is an action an app has registered in io.Connect, you have full control over what AI can do.

See how this works in practice

Deutsche Numis connected Fidessa and Infront so that sales traders could follow the same instrument across applications. Read the case study to see how Deutsche Numis streamlined their trading with io.Connect in just five days.

Building or rebuilding a trading desktop? Talk to our team about connecting the apps your desks already run, from Bloomberg and Fidessa to in-house pricers and Excel, and see io.Connect working with your own workflows.

Related Resources
Want to stay connected?

Sign up for the latest news!

About the author
Toni Petrov
Developer Community Manager
Toni is the Developer Community Manager at interop.io, where he has led the development of a new developer community from the ground up. With more than a decade of experience in web development, community building, and open-source engagement, Toni is passionate about helping developers connect, share knowledge, and build smarter solutions.
The Rise of Desktop Superapps: Transforming the Work Environment
Interoperability Everywhere All at Once
Trading Company Redburn Atlantic Syncs Apps with interop.io