ChatGPT adds personal finance as Codex goes mobile

Two announcements, one direction

OpenAI shipped two things this week. On 14 May 2026, Codex landed inside the ChatGPT mobile app for iOS and Android. It is free across every plan, including the entry-level tier. On 15 May, a personal finance preview rolled out to ChatGPT Pro users in the United States. They can now connect bank, credit card and investment accounts to ChatGPT through Plaid.

The first move puts a powerful coding agent in your pocket. The second turns ChatGPT into a personal money assistant that can see your full financial picture. Different audiences, same direction: AI is moving out of the specialist bubble and into everyday use.

What ChatGPT can now do with your money

The feature uses Plaid to connect to more than 12,000 financial institutions. The list includes Schwab, Fidelity, Chase, Robinhood, American Express and Capital One. Once accounts are linked, users see a dashboard covering portfolio performance, spending, subscriptions and upcoming payments.

You can also share context: a mortgage, a savings goal, a planned house purchase. ChatGPT saves this as Financial memories and uses it to ground future answers. OpenAI says more than 200 million people already ask financial questions through ChatGPT every month. The new product, built on GPT-5.5, is designed to give those questions proper grounding instead of generic advice.

Two practical caveats. The preview is United States only, on Pro plans only, on web and iOS only. And OpenAI is clear this is not a replacement for professional financial advice. The team behind personal finance startup Hiro, acquired in April, helped build the experience.

Codex on mobile is not about coding on a phone

The Codex update is easy to misread. You are not typing code on a 6-inch screen. The mobile app connects to your laptop, Mac mini or a remote machine where Codex is running. From your phone, you can review outputs, approve commands, switch models, redirect a task or start a new one.

This matters because agents now work on long-running tasks. A refactor might take 20 minutes. A test run might pause for human approval. Before the mobile release, you had to be at the keyboard for those moments. Now you are not. OpenAI says Codex has more than 4 million weekly users, and the mobile release fills the obvious gap.

It is preview, iOS and Android, all plans including free, with macOS desktop required at the other end. Windows support is coming.

Take the AI readiness assessment to see where your team stands.

Why both moves matter for UK businesses

At gecco we work with finance teams to deploy AI across financial literacy, reporting and pulling fragmented data together. We see this work directly with our clients. So the personal finance launch is welcome, even if UK access is some way off.

The deeper point is about adoption. AI succeeds inside organisations when people are comfortable using it. People become comfortable when they use it for things that matter to them: their money, their decisions, their time. Once a colleague has built a useful habit at home, they bring that confidence into work.

This is the 80% of AI adoption that gets ignored. The technology is 20% of the problem. The other 80% is people, culture, habits and trust. When OpenAI puts a finance product in front of consumers, it does some of that adoption work for free. When Codex moves to mobile, it tells developers that AI works around their day, not the other way round.

The accessibility direction in practice

Both moves point to the same shift. AI is becoming something you check on the train, approve from a café or ask in the kitchen. That changes what teams can realistically do with it.

A few examples we see playing out for UK SMEs:

• Finance teams using AI to translate management accounts into plain English for non-finance colleagues.
• Operations leads approving agent tasks between meetings rather than scheduling sit-down time for them.
• Senior leaders running scenario questions about cash flow or pricing on the move, not just at month-end.
• Customer service staff handling first-line queries with AI support while remaining on the floor.

None of this is exotic. It is the same kind of practical use case we help clients deploy through our Assistants and Automations work. The mobile release just removes one more friction point.

The honest caveats

Personal finance AI raises real questions. Data sharing, accuracy of advice, what happens when ChatGPT gets a number wrong. OpenAI has done sensible things. Data is removed within 30 days of disconnection. Private chats do not use financial data. Users can delete Financial memories at any time. None of that removes the need for professional advice on big decisions.

The same applies for business use. AI is a strong first draft. It is rarely the final answer. Teams that get this right pair AI with clear policy and training. They also know the difference between a confident answer and a correct one.

Next step

Two announcements in two days from the same company is not coincidence. OpenAI is pushing AI into the rhythm of daily life and daily work. For UK SMEs, the question is one we ask every client: are your people ready to use this well?