ChatGPT in SMEs: 5 Practical Uses from Email to Finance

Key Takeaways

1ChatGPT Work identified unfavorable conditions in a debt consolidation offer

2The AI recommended suitable UPS replacements after analyzing printer specifications and a supplier's site

3ChatGPT Work synthesized public relations emails, generated a slide from 750,000 public data points in 13 minutes, and extracted marketing features from the source code

Why it matters : These examples illustrate how AI can accelerate complex tasks and enhance operational vigilance in a small business.

In a small organization, ChatGPT Work is being utilized for critical tasks such as analyzing loan offers, sizing equipment, synthesizing emails, generating slides from public data, and preparing an in-app purchasing strategy. The use case documents time savings, errors to anticipate, and a cautious method for connecting to sensitive accounts.

In finance and energy, AI identifies unfavorable clauses and suggests UPS replacements

In response to a debt consolidation offer received a few months ago, the company tasked ChatGPT Work with reviewing the document. The AI confirmed that the proposal was issued by a credible entity but highlighted very unfavorable conditions hidden behind a more engaging main text and detailed in the terms and conditions. This alert came before a potentially costly decision, in a context where the company, indebted to several suppliers, is practicing budget vigilance and seeking to remain financially solid.

On the infrastructure side, the battery backup devices in the Fab Lab were overwhelmed by the increasing energy demands of 3D printers, to the point where the UPS could no longer support the load. After contacting Jackery, which was willing to partner on the project, an assessment was provided. ChatGPT Work extracted the specifications of the printers and consulted the manufacturer's website to recommend in detail how to replace the UPS and why certain devices would be more suitable, based on a prior audit of the circuits and despite uncertainty regarding the total load in operation.

Email sorting: refining queries and limited-time Gmail connection

For email support, ChatGPT Work is used with a Gmail connection activated only during important tasks and then immediately cut off, a precaution taken despite the friction it adds. The challenge is to sift through numerous public relations messages regarding AI, including exchanges and press releases related to OpenAI, as well as citation proposals sent by various PR representatives.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

The effectiveness heavily depends on the formulation of the queries. An initial request for a summary of the last five emails from OpenAI's PR representatives was misinterpreted by the AI, which restricted the analysis to emails from OpenAI's PR staff and missed the essence of the conversations. A revised, more precise instruction corrected the course.

ChatGPT Work transforms an incident log into a presentation in 13 minutes

ChatGPT Work proved effective in converting a public dataset, the San Francisco fire incident log—comprising nearly 750,000 records—into a PowerPoint slide. The goal was to identify the 10 most affected battalions, which the tool accomplished after several permissions, transforming a webpage into an aggregated graphic in 13 minutes.

Some touch-ups were still necessary: incorrect shades of green associated with ZDNET, overly wide internal margins, and formatting adjustments. Despite these aesthetic limitations and the confidentiality of other non-shareable content, the result is deemed useful.

In-app purchases: AI reads code to list marketing features

To prepare the in-app purchasing strategy for several Apple applications intended for the App Store and deployed on all platforms except Vision Pro, the company had ChatGPT Work analyze the source code, with the instruction not to modify anything. The constraints pertain to the timing of offers, their presentation, pricing, and included rights.

This method, reused for other projects, allowed the AI to directly produce marketing-focused feature lists from the code. The process is described as time-efficient and secure to avoid omitting important elements when drafting feature summaries.

Expense tracking: searching for a year of invoices in emails

As part of its verification policy, the company also tasked ChatGPT Work with analyzing its expenses with a given supplier. The instruction was to sift through the last 12 months of emails to identify all related invoices or receipts and produce an expense report. This approach is part of a systematic control strategy before any financial decision.