4 interesting ways to use Microsoft’s AI chat search

As Microsoft’s Bing continues its chaotic early access rollout, Google’s preparing its own Bard AI search tool. For now, we can only guess how each tech giant’s AI search product will compare.

Instead, let’s focus on a more practical question: why should I care about AI search at all? For over twenty years, internet searchers have managed just fine without AI. In order for AI search to disrupt conventional web search, particularly in business and research applications, it needs to be truly superior at performing complex web searches.

Today’s AI search tools aren’t widely available, so many people do not have the option to explore the tools for themselves. Fortunately, through early access to Bing’s chat feature, we were able to test a number of its features. Here are four of the most interesting ways we found to use this new tool!

Plan meals, events, and itineraries

As of writing, Bing does not integrate with travel services. You cannot directly book or manage flights or hotels, but you can still use it to support your travel plans.

One user on the Bing subreddit posted an example where they requested help planning an important business dinner. Providing information such as estimated budget and dietary requirements, Bing helped the user select a local restaurant, plan courses to suit each attendee, and even choose optimal wine pairings.

Using Bing to plan day trip itineraries or learn more about local events and destinations is intuitive and powerful. Here’s an example:

We told Bing that we wanted to book a wine tour for a small group of friends, including one vegan. We suggested that our preference was for crisp and refreshing whites.

Recommendations proved to be generally accurate, though the proposed dining spot didn’t have an up-to-date online menu. As an early tech preview, AI search results require some vetting to ensure accuracy. The hope is that by the time the feature becomes widely accessible, results will be more accurate, with more powerful integrations with things like booking apps.

Compare products using detailed criteria

The proliferation of affiliate-link-powered product ranking sites can make it frustrating to search for reviews or find specific products that match your unique criteria. It is hard to trust reviews and rankings when the website stands to gain financially from each included affiliate link.

Bing can’t fix that problem, but it can help. By crawling multiple web pages at once, Bing chat does a good job of referencing products from many articles and reviews. It’s also able to sift through review text to find products that meet your unique criteria.

The best way to leverage Bing for product selection is to play a game of 20 questions. Write a prompt like:

“Help me choose the right pair of bluetooth headphones for me. I am looking for a budget-friendly pair of noise-cancelling over-the-ear headphones. Ask me three questions that will narrow down my selection.”

This prompt resulted in three follow-up questions about specific price, battery life, and audio quality. Once answered, Bing suggested the following product:

Quickly learn basic concepts through dialogue

Whether at work or at home, one of the primary functions of internet search is to learn new things. We may explore new hobbies, look up recipes, or simply seek to understand a new concept for business or personal fulfilment.

With Bing chat, we have a more natural method for learning, provided we understand the tool’s limitations. Bing cannot accurately teach advanced concepts, but it’s great at explaining basics.

Its educational power is intrinsic to the medium. A chat is a dialogue, well-known as a powerful learning tool for millennia. Popularized by Plato, dialogues promote deeper learning. The student can query the teacher, challenge assumptions, and build upon concepts throughout the conversation.

Through Bing chat, we can approximate the benefits of dialogue. For example, I’ve been interested in music theory lately. I love listening to and analyzing music, but I lack the vocabulary and conceptual understanding to express my thoughts fully. I asked Bing to teach me about tonic notes. Here’s the conclusion of our dialogue:

Research and structure information at the same time

There are many business situations that require research. For example, product research and evaluation, market research to support an upcoming campaign, or SEO research to find the perfect keyword for your next article. In many cases, we’re synthesizing information from multiple sources. We might need to combine statistics from several websites, or take an average of several values. Perhaps we want to blend different types of information into a table.

With conventional search, we have to organize our own information, then structure it appropriately. With AI search, we can often structure our info without leaving the chat window.

Here’s a personal example. I’m a big hockey fan. The NHL is a hard cap league, meaning that teams must comply with minimum and maximum salary caps that grow (or contract) over time. This means that players must be evaluated based on their on-ice performance as well as their cap efficiency — what they offer to the team relative to their contract.

As a fan of the Vancouver Canucks, I identified winger Tanner Pearson as an utterly average player: average points production with a salary to match. I was curious to know how other players on the team would be compensated if they were offered the “Tanner Pearson special” — a contract with the exact same points/salary paid ratio as Pearson’s. In theory, this would show who was underpaid and who was overpaid for offensive production, relative to Pearson’s middling performance.

To satisfy my curiosity, I assembled a spreadsheet to calculate this value. I had to manually pull stats for each player from websites like CapFriendly and NHL.com, which was quite time consuming. Can Bing do it better?

I asked it the following question:

Within seconds, I received the following table (edited for brevity):

A table that might have taken an hour to research and assemble was generated in seconds!

There are notable caveats here. Bing chat can inadvertently “lie” if fed bad info, so double-checking records for mission critical research is essential. Although Bing AI can search the live web, it may struggle with up-to-date statistics and brand new or emerging topics, hence the use of last season’s data.

Still, generating structured data within seconds using Bing chat is an exciting new way to incorporate AI into a research workflow.

Conclusion

We know that AI search is on its way, and we’ve seen how rapidly this tech evolves and improves. It’s entirely possible that AI search tools completely supplant conventional web search over the next decade. For now, these are a few of the most useful ways to leverage AI search. By next year, there might be dozens more use cases.

Keep an eye on emerging search technologies. Test them if and when you can. This is a sea change for web search, and early adopters will have a leg up over their competitors once this tech goes mainstream!