Two Step Approach to Deep Research
Using the response from an initial prompt as a starting point for deep research.
This week, I want to talk about another practical process that I have been using generative AI for, which is combining a basic prompt and using the output for Deep Research. One of the best uses of AI chatbots is for ideation and it's a fantastic tool for brainstorming, exploring different facets of an issue, and finding keywords/angles to research. We also now have more capable Deep Research models which excel at thoroughly exploring an idea. These models take a prompt, go away and 'research' a whole host of resources, compile it's main responses, and then produce a comprehensive report. Given the amount of work the system is doing, this can take around 10-15 minutes, which feels like a lifetime compared to a near instantaneous response with regular models. But, when you consider this type of report might take a researcher weeks, if not a month, to produce, then it's beyond quick.
Let's Look at an Example
Using the basic Flash 2.0 model in Gemini for a quick response, I prompted the following:
"Given the current state of the uk economy, can you highlight the top three challenges facing the hospitality industry. In response to that, can you also suggest the top three solutions businesses have to minimise the issue."
It gave me three solutions as requested, one of which I thought was particularly interesting: "Enhance Value Proposition and Customer Experience". I then used this topic to focus on as part of the prompt for my research.
Now using the Deep Research with 2.5 Pro model in Gemini, I gave it the below prompt:
"given the state of the uk economy, can you research how the hospitality industry can Enhance Value Proposition and Customer Experience to minimise the challenges they currently face."
It took about 10 minutes to produce a 34 page report on that specific topic, citing all of the various sources that it has used. It also provides tables for comparing basic metrics and lays out the information in a very logic and digestible format.
It really is a good way to explore an idea or concept you need to think about. And with anything related to generative AI, don't blindly take all of this and paste it into a piece of work, but take the time to understand the core themes and messages to deepen your own understanding.
Do Your Own Validation
As impressive as this is, don't take everything it produces as 100% fact. Look at the sources yourself to validated some of the numbers and metrics it produces. Now that you have a deeper understanding of your topic, you can undertake further research on your own, now that you know where to look, and what to search for.
And if you're interested, you can see the full report here.
This week's email is brought to you in partnership with Capsule.
Too many CRMs feel like they’re built for massive sales teams. Capsule is different. It’s built for small businesses that want a simple, effective way to track customers and sales, without spending hours figuring it out.
- Easy to use
- Helps you close more deals
- No unnecessary features
Curated News
Visa Empowers AI Agents to Make Purchases on Your Behalf
Visa has announced a new initiative to enable AI agents to make purchases on behalf of users by linking them to its global payments network. Partnering with leading AI developers like OpenAI and Microsoft, Visa aims to facilitate autonomous AI assistants capable of performing shopping tasks, such as managing routine purchases like groceries or plane tickets, based on user-defined budgets and preferences. This development could transform e-commerce and become a competitive response to the dominance of Amazon and Google in digital retail, while easing everyday shopping burdens for consumers.
Microsoft Envisions a Future Workforce Managing AI Employees
Microsoft envisions a future where every worker becomes an "agent boss," managing AI agents to perform tasks and enhance productivity. The company anticipates the rise of "frontier firms"—businesses driven by autonomous AI systems that provide on-demand intelligence to streamline operations. Over the next five years, organizations are expected to evolve through three phases: adoption of AI assistants, integration of digital AI colleagues, and full delegation of tasks to AI agents, with human oversight. However, such advancements raise concerns about job displacement and the erosion of institutional knowledge.
Gartner Identifies Agentic AI as 2025's Top Strategic Tech Trend
Gartner has identified agentic AI as the top strategic technology trend for 2025. These AI systems can autonomously plan and execute actions based on user-defined goals, transforming how businesses operate. By 2028, Gartner predicts that at least 15% of day-to-day work decisions will be made autonomously through agentic AI, up from 0% in 2024. This shift is expected to redefine leadership and accelerate digital transformation across industries.
Upcoming AI Events
Rise of AI Conference
Berlin, Germany, May 7–8
AI in Business Conference
London, May 15
Google I/O
Shoreline Amphitheatre, Mountain View, CA, USA, June 9–12
Thanks for reading, and see you next Friday.
Simon,
Are you looking for a job in the tech sector? You'd do well to check out all the available roles on Wellfound. Wellfound is a platform that connects startups and tech companies with job seekers. It offers over 130,000 remote and local job listings, allowing candidates to apply with a single profile and view salary and equity details upfront.
You can also follow me on social media:
Comments ()