What is a multidisciplinary conversational AI team (CAI)?
Peter’s multidisciplinary CAI dream team would include conversation designers, NLU designers, response designers, developers, product managers, product owners, and customer support people. They’re all part owners of the platform, even if the conversation designers still lead.
This allows you to source issues and fixes from seven unique angles, allowing for much better coverage. And having everyone be part of these discussions helps you fight product bloat.
“Sometimes I’ll join a team working on a chatbot and find they have over 1,000 intents, and most of the utterances are just made up—they’re based on what the person who was building the model thinks someone would say,” says Peter. “The intents are riddled with unintentional bias and those bots don’t work. There are just conflicts everywhere.”
"Sometimes I’ll join a team working on a chatbot and find they have over 1,000 intents, and most of the utterances are just made up."
An enlarged team is more likely to include people who talk to customers daily, who catch those offbeat intents. Odds are, they’ll also include people who are willing to challenge others’ assumptions. It’s the best defense against stakeholders coming to your team asking for unnecessary content. At least someone in the room is going to reply, “Wait, is this actually a thing that needs to be solved? Are people even asking this stuff?”
Let’s explore those benefits in detail.
The four benefits of a multidisciplinary CAI team
1. Diverse perspectives lead to smarter decisions
One of the fascinating things to come out of studies about diverse teams is that the results are better but work may feel less comfortable in the moment. That’s because you actually have a rare thing—people with differing opinions. You need good dynamics and open dialog to make that work. But if you have a baseline of trust and respect, you start to realize the benefits after adding just one person with a different perspective.
In one study, students were told to solve problems in groups, and then partway through, they’d be joined by an extra team member. If that team member were a stranger, those teams became twice as likely to solve the problem (60% compared to 29%).
Plus, diverse perspectives mean you deal with technical constraints up front, not later, after the thing is built. “You want UX designers arguing for the best customer experience, developers reminding everyone of the technical nuances of voice recognition, and product owners bringing up the company’s goals,” says Peter. “And you want them in one place to agree on how all those tradeoffs result in something users enjoy.”
2. Greater innovation
This is where The Avengers effect takes hold. You don’t just have the power to fly. Or a magic hammer. Or whatever it is that Hawkeye does. You can apply all those powers to experiment, share ideas, and push the boundaries of what CAI can achieve.
“In a customer support chatbot project, maybe you involve the platform team, and rather than more intents and flows, they think up a way to solve the issue by incorporating elements like sentiment analysis, context-aware responses, and proactive assistance,” says Peter. “And then, that gets others thinking, and now you’re on a totally uncharted path toward true innovation.”
Diversity also makes your team more reflective of those you serve, and just having more people allows you to incorporate ideas from more backgrounds, life stages, and geographies.
3. Faster communication and clearer decision-making
And by “faster communication” I mean you have small arguments up front rather than big ones later. In a siloed approach, a project goes through checkpoints and sometimes, somebody has a substantial objection that really ought to have been dealt with earlier. If everyone’s in the room, that objection doesn’t wait.
“Take a healthcare chatbot project,” says Peter. “There may be conflicting opinions about how you’ll prioritize user privacy and data security against adding new features. A multidisciplinary platform team is going to work through that before they can move forward.”
Again, this works best when your team is practiced in speaking the truth and resolving disputes. “Are there any teams that really work in perfect harmony together?” asks Peter. “I don’t think people are generally very good at that. Clear, focused communication is the key. Maybe the stakeholder needs something and can’t budge, but you can budge somewhere else. But if you budge elsewhere, the developer’s going to say you’ll blow out the project because it’ll take longer. You hash that out now, not later.”
4. Balancing user needs with business objectives
This collaborative approach allows the team to identify the best way to address both user requirements and business goals, while ensuring the CAI is functional and user friendly. And rather than a messy compromise, you really do get all three of those things.
“In an ecommerce chatbot project, the platform team might recognize that users frequently ask for personalized recommendations,” says Peter. “Normally this might not generate revenue, and might not get prioritized. But the UX team can explain that it enhances user satisfaction and engagement—and through testing, you find that actually does impact revenue, because you learn you can offer sponsored recommendations.”
Multidisciplinary teams are more creative
I don’t know about you, but the most compelling reason is that multidisciplinary teams are also more enjoyable. If creativity is intelligence having fun, I want to be on a team where people bring something totally different. Where everyone isn’t just sleepwalking through work. Where people care enough to have heated conversations and respect each other enough to work through it.
Those are the sort of CAI teams that can save the world. And that’s a movie I’d want to watch.