How do we represent that a search feature has "intelligent" capabilities?

Question

My software company is beginning to implement "intelligent" query capabilities to the usual search boxes for our customers and I think that this necessitates the need to evolve from the old standard magnifying glass icon to something else.

For example, consider the search box right here on Stack Exchange:

This makes sense to have the traditional magnifying glass icon because it is a traditional search. However, the same "search" box in some of the applications we currently have in development are "intelligent" or "bot-enabled" which means that it almost becomes more conversational instead of a one-way simple command.

For example, instead of simply typing in "AA Batteries" to see a list of search results, the user could type in "Add AA batteries to my cart, 36 pack" or "Buy 36 pack of AA batteries" and their request is carried out by the smart algorithms.

We are seeing many new requests for this type of thing and the underlying technology is common now and works well (you have probably used a personal assistant today already).

To take this even one step further, we are also seeing requests to enable microphones (speech inputs) to these boxes in addition to text (keyboard). I am not sure if this should be a second icon or if there is some sort of "omni" icon that can serve as a single call to action (in other words, the user recognizes that once they click it they can begin talking OR typing).

In Windows 10, both the mic and the magnifying glass icon are shown when the textbox has focus, but I can't help but feel like there is a better, more unified way to convey the combined capabilities:

It looks like Google is just using a mic and, as Anthony Hobday pointed out below, people seem to know that the Google search capability does much more than just search:

As far as iconography goes, the best I can come up with is either an ear abstraction or a soundwave abstraction to indicate that the system will listen to you:

• 

• 

Problem is, this doesn't necessarily indicate that the system is "intelligent".

Answer 1

Use a small stylised graphic of a persons face / head and shoulders next to the search box - rather than the magnifying glass.

With increasing penetration of the Siris and the Alexia's into the market place people will get used to being able to interact with 'digital assistants' and the line between whether you are actually dealing with a human or a machine will get increasingly blurred.

You might want to add a microphone icon which appears - in the text box - when someone clicks into it to signify that the assistant is now actively listening.

Answer 2

Teach the user how to use your feature. Show them example queries to illustrate what is possible. This can be done in some sort of short onboarding experience for new users, or perhaps when the input is in focus.

A feature that is not discoverable might as well not exist.

I would venture to say that this problem is not one that is solvable by simply choosing the right icon. This conversational style is still a pretty uncommon type of interaction, so a user will likely not assume that these types of queries are possible unless they are shown.

Siri and Cortana are more of a "personal assistant" type of interaction, as you're describing. I know on my Windows 10 computer, it provides me with example queries as to what I can do with Cortana. On my iPhone, Siri shows me suggested queries as well.

As for interacting with these assistants, as far as I can tell, there's no hard-and-fast standard for these "personal assistants"--Microsoft made their own icon (the filled-in circle enclosed by another circle) to trigger their assistant, while Apple chose a gesture (push and hold on the home button).

Answer 3

The actions the user will want to perform take this pattern:

object [action]

Search field:

term [search]

"Search for this term" = "this term" [Search button]

Item list:

item [add to cart]

"Add this item to cart" = this item [Add to cart button]

In your case, users will use the text input the following way:

term action [execute action]

As you will need a button for the field, you might find an icon that could represent this "execute" or "run" action, but I am not aware of any convention so probably it would be better to stay away from icons and use text. @AnthonyHobday proposal of using "Go" as the text for the button is a good solution.

If I understood correctly, your algorithm will detect what action the user wants to perform and execute it. If this is the case you could actually fill the button with the action the user typed. This would need extensive testing to analyze all the possible outcomes, but it might be a direction to start from.

---

About the updated question.

Note that there are two things: The magnifying glass as a hint (that the field/input is meant to search) and the button to search.

Take the example of this UX Stack Exchange site (as it is right now). It doesn't actually have a search button, just the magnifying hint.

In the case of the Google input, the action button is not the microphone but the "Google Search" button. The action the microphone button does is listen and then perform. It also gives a hint of what can be done in the input. My guess is users could associate this hint with "the input can do anything you ask".

In your case the button is not going to listen, it is going to execute the action the user types. So I am not sure it is a good idea to use icons that suggest this kind of actions (talk, listen, etc.).

Answer 4

Users are likely to try to use the search box as a search box, no matter what icon you choose.

My assumption is that if someone sees a text field in the top right corner of a website, for example, they will identify it as a search box.

I think an approach you should test is, as you said in your comment on another answer, having an expanded section which is shown when the person tries to search, letting them know, "Hey, search isn't all you can do!".

This could include examples, and a link to a help article explaining the full capabilities of the feature.

As far as visually identifying these features somehow, I'm not sure you need to. Google, for example, has all sorts of features that are NOT search related. I can type "£10 in dollars" into Google, and it will calculate it for me. It shows search results as well, but hopefully you'll do that as well.

My point is that maybe people are slowly becoming used to the idea that a search field doesn't just submit a query and list the results. Especially since the largest search company in the world is leading the way.

Some things to consider trying, if you did want to make it stand out:

• Change the "Search" button, if there is one, to "Go" or similar. Gets you away from the "Search" concept a little.
• If you need an icon, consider something like a person speaking. This gives the impression that the person using your site is talking to it, which covers all sorts of interactions.

Answer 5

Well it depends on your clients preffered interaction type. Some clients will need to emphasize more on voice search while with others more on keyboard input. Each client will have different needs because of their different users, and contexts.

So you can divide interaction types on:

• speach as primary action
• keyboard typing as primary action
• equal weight between the two

If it primarily relies on voice commands a speach icon and accompaniyng text affording speech should be emphasized (and vice versa).

Different variations of the text and the icon should be tested with users. You can test this by showing different image variations (.png's) to random people and ask how they think it can be interacted with the search bar. Allow users to look at each image variation for 3-5 seconds only, in order to capture their first impressions of the interface.

This experiment is affordable and easy to run so don't make excuses and test it.