Conversation Design: How journalism can shape UX

This is Part 1 of a series of interviews with the people behind Klüg: the designers, engineers, and product specialists that make our bots the way they are.

Here, Conversation Designer Diana Lee, has a contributed post on how she started writing human-machine interactions after designing human-human interviews.


Like many writers, I stumbled into UX.

I never learned to code, or formally picked up visual design tools (though I can dabble in a couple).

But I always loved conversations. I love interviews where I get to know the other person and tell their story. That’s why I chose to get trained in journalism.

Now, I get to work with conversations every single day, namely designing interactions between machine and humans, specifically bots.

Here are the values I learned that inform my work:

  1. The pursuit of perfection will do more harm than good.

When you have specific milestones, you learn that there is no such thing as perfection.

Working in journalism taught me that being stubborn about perfection leads to a bad cycle of constant missed deadlines and less time for iteration.

The latter is especially detrimental to launching chatbots because you must allow enough time to test with users so you can make data-driven decisions in the final copy.

You do the best in the time you have, and if you notice the quality is not up to par, you tweak a couple things here and there to figure out how to be more time-efficient. Spend more time working on flows you identify as time-suckers. Color-code certain flows to make it easier to come back to iterate.


In this scenario, I wasn’t quite sure about how to handle the bot handing off the conversation to a human.

In this scenario, I wasn’t quite sure about how to handle the bot handing off the conversation to a human.

In this scenario, I wasn’t quite sure about how to handle the bot handing off the conversation to a human.

2. Working with a team, while owning your part.

Any technology product development is teamwork. Anyone who’s worked in news will tell you that it takes an entire newsroom to deliver a story. From the producers to reporters to editors to the interns answering calls, a piece of story passed over the hands of many people before we decided it could be viewed by the public eye.


What must be done before a story gets published (could be multiple rounds).

What must be done before a story gets published (could be multiple rounds).

I’ve met quite a few UX writers from a journalism background. We can bond over editors breathing down our necks as the deadline approaches. When we’re trained in an environment where time is of utmost importance, and being 2.5 seconds late slows down everyone, we learn discipline. Working on breaking news stories will be some of the most stressful times you’ll have, but it becomes muscle memory that serves you in any job function.

3. Agile is a mindset, not a one-time application.

We all talk about specific steps to agile but I can’t think of anything more agile than the flexibility required in getting a story on air, paper, or online.

Now, in the tech environment, I don’t take those couple seconds like life and death. But I do still hold myself to a standard to completing things on time, so that there is time for iteration.

Agile is about adapting to the lack of predictability in technology development. Agile sometimes seems to be used as an excuse to keep pushing deadlines. On the contrary, this is the worst thing to do because we’re exacerbating the already heightened room for errors.

Much like the process of designing and developing bots, nothing is streamlined in producing news.

If you thought you would have 3 days to work on the 3,000-word feature story due in 2 days, well… wait until you get told to turn around 2 articles by the end of day. End of day also means 5 p.m., not 5:02 p.m.

If you also thought you would have 3 days to design 12 different user flows (paths that a user can take to achieve a goal, like setting a calendar reminder), you may discover that there are 2 additional functions that must be built into half of these flows to have them working properly — such as authentication to confirm the calendar owner’s identity.

A snippet of how a user can confirm their identity.

A snippet of how a user can confirm their identity.

4. Prioritize to deliver on time, then iterate.

Trying to be good at everything made me good at nothing. When I’m working on 3 to 4 chatbots, I must learn to prioritize by taking one project, looking at the big picture then refining later. Not only does this help you make changes faster, it allows you to do so with less bias because you’re coming back to your work with a fresh mind. If I explain my methodology early on and demonstrate how this allows us to iterate and deploy faster, then my team members understand.

Incremental, iterative work is what leads to a polished product in the end.

It’s the same for journalism. If you only have 10 minutes to proofread an article before publication, you would prioritize:

  • Someone’s name. Misspelling someone’s name in a journalism class would automatically lead to an F, as if you never did the assignment at all. This seems harsh, especially considering these were assignments often not ready for publication, but it sets you up for reality. Imagine writing about someone who’s accused of sexual harassment and you misspell their name to another public figure. That article could do irreparable damage to their reputation. Case in point: the 8.8 million pound of financial damage from a typo.

The extra “s” that took a company out of business (Credit: Media Wales Ltd)

The extra “s” that took a company out of business (Credit: Media Wales Ltd)

  • Numbers. Imagine reporting on stock prices and saying a company’s share price is $28, not $280. Neither the reader nor the company’s PR department would appreciate this error.

Sitting in a newsroom during a breaking news story trains you to be laser-focused to deliver. It’s better to deliver something on time, no matter how poor it is, than to deliver something perfect at a time that does not allow for improvement. In the first scenario, you may have extra time to make changes. In the second, you’ve already used up the extra time and you cannot improve.

There’s always room for iteration on a poor product, but you can’t iterate on something that you never deliver.

With that said, here’s how I applied journalism training to UX writing.


How Google incorporates UX writing within UX (Credit:  Guy Ligertwood )

How Google incorporates UX writing within UX (Credit: Guy Ligertwood)

Whether or not you come from a writing background, the first step to get started with UX is:

Research. Writing words that enable a good user experience means you know the user. In reporting, this means knowing who your audience is. Writing for a 25-year-old artist living in an urban city will require different language than a 55-year-old just-retired person looking for financial advice.

In journalism, you must know your reader, watcher. In UX, you must know your user, customer.

How is this done in journalism?

Know your publication audience. If you’re joining an organization with a solid readership base, you’ll learn more about your audience through the job — you’ll know depending on how many times your pitches get rejected, anyways! Or maybe you’re creating an entirely new publication based on an observation that there isn’t content suited for a specific audience, like the guys at The Hustle. Their company was founded specifically to address the the early- 20’s to 30’s looking for finance news. Just based on their employee base, you know they were solving a problem they were experiencing themselves. The email chain idea itself isn’t a revolutionary idea, of course, but as one of their readers, I can say they know their readers. And for the areas that they don’t know about, they ask for suggestions:

Feedback request at the end of their email

Feedback request at the end of their email

How is this done in UX?

Know your users. Beyond their demographics and daily habits, you can interview them to understand their decision-making process (how do they decide what phone to buy?) and their pain points in how they currently make that decision (too many review sites to sort through).

So what do journalists and UX writers have in common?

They know who they’re writing for. They have empathy for them, and they fight for words that they need to see or hear to make a decision.


This piece was originally published on the UX Planet magazine.