Anthony Maggio - Driving AI Value with In-House Citizen Development

And Anthony will also tell us a little bit more about some of Airtable's key products and features which best enable all this. Anthony, welcome to the podcast. Very, very happy to have you here. Anything to add before we get into the meat of our discussion?

Anthony Maggio: Thanks so much for having me, Tim. Looking forward to discussing with you today.

Tim Butara: Okay, awesome. So the first thing that I'm interested in and to kind of set the stage for the whole conversation. Why is it particularly important to talk about citizen development in the context of AI? I mean, maybe we can start with a very, very brief breakdown of what citizen development is, if anybody's not familiar with the concept yet.

Anthony Maggio: Yeah, let me start there. So citizen development is simply the concept of enabling your line of business users, people who come from non technical backgrounds to develop software and applications and at Airtable. Our mission has always been to democratize software creation to enable a wide range of individuals who come from a diverse set of backgrounds to be able to create applications to improve the way that they work and the way that their teams manage recurring processes.

And what we're seeing is that this trend that, you know, we have been 12 years or so at Airtable. It's just continuing to be exacerbated by AI and that the people who are closest to the work, who really understand processes within an organization are really those who are in the best position to define how AI can automate those processes.

So that is at a high level why we believe that no code and citizen development is really going to be a beneficiary from AI. The latest developments around AI and LLMs and some of these really powerful models that have yet to be applied to workflows in most organizations and really profound ways.

Tim Butara: Are there any like obstacles, any challenges to this really being, you know, put into motion and really being implemented properly?

Anthony Maggio: Yeah. I mean, there's a number of challenges. I think, you know, we've been talking about AI for close to two years now, right? It's been about two years since ChatGPT took the world by storm and everyone started to get a preview of what was possible with the latest generation of LLMs.

But I think there's a big disconnect right between what we've all seen to be the capabilities of AI and the capabilities of LLMs and the actual impact that they've had within most companies and with most organizations. And, you know, I think that there's a few reasons for that. One is that. Many companies have been you know, reticent to apply AI into their workflows, have had, you know, concerns over security or data training or model training or whatnot.

I think that, you know, most of those concerns have now largely been addressed by the model providers, as well as by companies like Airtable, who are, you know, taking different approaches to enabling these things. Models for enterprises. For example, we host all of our own models and guarantee no model training on all the models that we provide.

You know, that said, we still see that there's a temperature of caution in terms of where and how to apply AI within the enterprise. The bigger blocker, you know, that I see though, is that for most companies, they have not yet identified. The best way to actually reap the benefits of these LLMs. And I think there's that the biggest disconnect that I see is that most business leaders will cite AI as a top priority.

Most will say that they feel increased urgency to implement AI, to modernize their products, their workflows, the way that their teams are working. You know, if you look at the latest generation of startups in Silicon Valley, almost every single one of them is an AI native company going after some type of incumbent business and trying to disrupt those businesses.

But, you know, despite all of that, AI adoption you know, within most large organizations has been. You know, Sower has been tepid and what, you know, from my point of view, I think that largely that kind of comes down to the approaches that enterprises have taken to adopt AI, what we've seen, you know, by and large is that a lot of companies went out, they bought chatbots, they bought some type of copilot. They hope that this would be, you know, a simple, easy answer in terms of how to bring the benefits of AI to their workforce. It quickly became clear that that approach was not going to, you know, really deliver kind of profound business transformation might've been helpful from a sort of simple productivity standpoint.

Point solutions have, you know, in many cases implemented shallow AI capabilities like summarization or kind of additional efficiency gaining capabilities here and there. But those, you know, also have not really promised kind of larger, you know, business transformation results. And then, you know, what we're seeing is that some companies are trying to go build.

Really kind of custom AI solutions, traditional software development, but you know, those can take a long time to implement. So, you know, we really believe in what we see at Airtable is that no code is the missing link. It's that missing middle to driving, really recurring and compounding results from AI to be able to identify.

Workflows and processes where there's the opportunity for a large degree of AI automation and then to allow, you know, those people who sit within the lines of business and marketing or product development or business affairs or legal, you know, to, who really understand where those opportunities exist to develop their own custom AI solutions for automation and for improving efficiency and for gaining greater insights from their data using the latest AI capabilities.

Tim Butara: This is why we're talking about citizen development. It's the people who typically weren't given the access and the power to really create all of this at a high level are now able to, to do so much more easily. And on a similar level to, you know, like developers are very technical people, exactly because of these no code tools, no code platforms, local platforms, which are only streamlined by AI, but as we just highlighted, only if you do it properly, I guess.

Anthony Maggio: Exactly. That's right. And I think that, you know, it's the, as I mentioned, it sort of compounds the same challenge that we have seen around point solution software in general. I think historically companies, you know, would either build or buy a piece of software and then implement it for a specific workflow and as people start using the software and they quickly identify ways that it should be, you Improved or changed to support their work.

But the challenge historically has been that, you know, with point solutions, you're limited in terms of how much customization you can do. With internal tools, change is costly and you're often bottlenecked by internal developers. And so, you know, in the current age of AI, companies are largely still implementing software in this way, really prohibiting them from unlocking, you know, the full value of customization.

And that's where Airtable has been a partner to so many, you know, organizations. We now work with 80% of the Fortune 100. We work with thousands of organizations, you know, globally to provide this, this alternative means of deploying software through developers who actually exist and sit within their business functions.

Tim Butara: So why is it so important to do this citizen development part in house rather than, you know, in collaboration, in partnership, through outsourcing?

Anthony Maggio: Well, I think that what we see is that many of the people who contain the knowledge and the expertise to identify how to best leverage and gain benefits from AI, sit in house.

And so, you know, to speak to the added like flexibility that companies receive when AI is integrated through in house operations, you know, they really gain kind of benefits in terms of customization, in terms of understanding of their own business processes of understanding workflows and applying the technology to to those workflows.

You know, just as a, an example, like. I worked with recently with a large record label who was looking and interested in learning about how to implement AI to gain some efficiencies across the business. We looked at a variety of their processes, looked at opportunities, you know, in terms of how the work was being done today, you know, quickly identified that a lot of the.

The work that exists with within the record label, you know, goes into understanding data that that is currently sitting in documents and contracts. And today, it requires just a huge amount of manual effort for individuals to go read through those contracts, to extract specific terms out of the contracts, to enter them, you know, into spreadsheets.

Spreadsheets to take the data around things like royalty payments that are due to an artist or contractual terms around marketing and to integrate them and insert them into various, you know, business applications. Well, you know, these are workflows that the latest generation of LLMs is actually. Quite good at solving.

It's very easy now to, you know, to point an LLM at a document and to ask it to, to scan through the document, to extract the terms that are needed, to output them in a specific format, you know, to pull those results into, you know, into a structured form, you know, that's one example of where we are already seeing a lot of interest in these types of capabilities.

And of course, you know, when you build an application. In a no code tool like air table, you're not just running that type of action for one document. You can apply that, you know, that type of action to thousands of documents in a matter of seconds. So you're taking a workflow that was, you know, previously costing hours and hours of human labor, or was maybe outsourced to a third party vendor and now automating it, you know, within, within seconds, which can provide just a massive amount of business impact.

Tim Butara: And I mean, of course also kind of the education and reskilling, upskilling related to, you know, really empowering your own in house citizen developments that also involves some costs. But at least that's a long term investment into, you know, having more capable employees, having more capable teammates, and this is actually already a perfect transition into the next step.

part that I wanted to, to touch upon. And that's this culture and people aspect, you know, in organizations and companies and how like how that should change if it should change to, to allow these implementations to be as successful as possible and how can also leadership and management people, I guess, help to make this work successfully.

Anthony Maggio: Yeah. I mean, I think that. In many roles, comfort and familiarity with AI tools is going to be a critical skill for many to develop. And I think we're, we're still in, in early days there, but we're seeing, personally seeing, you know, countless examples of the type of people who, Have embraced this technology who have gone and learned about how to work with LLMs, about how to work with AI.

We're seeing that they are able to, you know, to really drive profound impact and to bring, you know, these type of unlocks and these efficiency gains and new insights into their, their organizations. And I think that that's something that, you know, You know, you really do need a kind of top down push from a leadership team, from management team to support, you know, there's one at the most basic level is, you know, just giving your team the air cover or the freedom to go experiment with these tools.

I think, you know, we still see many, many examples of, of companies who have shut down or kind of turned off all access to to AI products or to AI tools, and I think that, you know, that, that will largely kind of be. Looked back upon as a mistake because it is becoming so critical for, you know, really across every industry, across every category to be able to leverage this technology in a way that's going to unlock innovation and unlock efficiency and unlock, you know, really kind of workflow breakthroughs from, from their employees.

So that's number one is, you know, creating an atmosphere where this type of AI driven innovations, you know, not only allowed. But but encouraged, I think too, is that you, you actually need to go and, you know, give your teams and employees the right tools to be able to create and drive this transformation.

And, you know, I hear, you know, many examples of CEOs and executives who are saying they want to, they want their, their companies to be adopting AI in innovative ways. They want to drive. AI efficiency gains across the business. Then you actually look at how are you doing that? What are the technologies that you've enabled across your teams?

And, you know, it turns out there's not a whole lot there to actually go and drive those gains. I think that's, you know, a big role that, that leaders will have to play as well. And then I think, you know, three is to really, you know, push on this sort of Mid level leadership teams to own the mandate of identifying opportunities to bring AI into their own workflows.

And, you know, that's where, you know, we believe that there's so much opportunity, this kind of long tail of AI business transformation where, you know, so much of the work that happens in most companies today does not exist in the. Three or four or five kind of core systems of record or core systems that have engagement that employees interact with today.

So, you know, every company is going to have like an HR system and a financial system, and maybe some, you know, engineering systems, but there's a huge body of work today That exists outside of those tools. And it's all of those, you know, those day to day workflows of managing the content, managing the planning, managing the strategy, you know, that exists outside of those tools that occupy a large percentage of knowledge workers time.

And, you know, those are the areas where. Where, you know, we, we refer to them internally as kind of the, the long tail of opportunity for AI. But I think that, you know, that's really where we're going to see the biggest benefits in terms of most companies who are able to, you know, really apply this technology in ways that it will accelerate those workflows, that it will, you know, unlock new insights out of those workflows.

Just as an example there, I mean, I think we work with a lot of teams in software development. And product management at Airtable, we've seen, you know, many examples of product teams applying AI across their user feedback workflows. So if you take, you know, all of the different sources for our product development team, where they're looking at user feedback that might be coming from community forums, might be coming from sales teams, from support tickets, from emails, you know, from lots of different calls and transcripts, you know, today, that is a.

Essentially, just a mountain of untapped data that provides a ton of insights right about what your customers and users are saying about your business. Well, that's a, you know, a perfect use case for AI to help a product team solve to be able to go, you know, take that unstructured data. Pull the relevant bits out of it that are informative for a product team to influence roadmap and priorities, and then to help structure them in such a form that you can actually understand the trends in that data, understand, you know, where there's opportunities to go drive breakthroughs and innovation.

And, you know, those are some of the areas where we're already seeing, you know, really massive impact from applying AI into these kind of, kind of common workflows within enterprises.

Tim Butara: Could you share any specific examples, like any specific success stories from air table? And also I'm interested which products and features are the ones that like empower citizen developers the most and give them like the most, the most options.

Anthony Maggio: Yeah. Why don't I start there actually? So I think, you know, the way that we have implemented AI within Airtable is somewhat unique. I think that, you know, a lot of what we see in other kind of point solutions and other tools are, you know, what I would call like relatively shallow. Implementations of AI.

It's okay. You know, I've got a I turned on in my chat tool and it can summarize a conversation or summarize a thread like that's great. You know, maybe saves you like a little bit of time here and there, but it's not really sort of leveraging your own company data or your own leveraging access to tools in unique ways.

What we've done is we're effectively now, you know, allowing Citizen developers to create custom AI features within the applications that they develop on Airtable. So, you know, as an example of this, and, you know, you asked for some examples of success stories. Like as an example of this, we work with the team at Amazon web services, the marketing team within AWS as an Airtable customer, they were one of our early customers in terms of adopting AI.

And they're using all of our, all of the anthropic models that are hosted on our AWS bedrock offering to further enhance their processes. And, you know, their major use case was marketing campaign planning. So they'd already built out an application in Airtable that manages all of their marketing campaigns.

It was integrated with Salesforce and Marketo brought together kind of all of their data to help marketers plan and execute B2B campaigns. And one of. The users who is responsible for maintaining this app thought to themselves, Oh, you know, wouldn't it be great if I could actually use AI to help marketers create better campaigns.

Right. And so they developed a custom AI function and air table that would look at all their past performing campaigns, identify where there are commonalities across the best performing campaigns. And then each time a marketer would create a new campaign in this application. This custom AI function will give them suggestions around how to improve the copy of their campaign, how to improve email subject lines, you know, how to change the, some of the targeting of the campaign to yield better results.

And, you know, that's an area where they actually saw like pretty dramatic increases in open rates of emails, as well as in the performance of campaigns as a result of implementing this in a real kind of recurring workflow. So the examples are there, you know, that's one, but we've seen Many similar stories over the course of the last year or so, have we, as we've worked with our customers to implement AI into these bespoke processes, but I think what, you know, really kind of stands out about that use case is that it was extremely tailored to the organization.

It was leveraging a lot of their, you know, their past data. It was leveraging data, you know, that only they had access to performance data from their own systems. And that's what, you know, we look to as the model. For, you know, where you can really have much greater kind of business impact is by leveraging your own insights about a workflow, leveraging your own proprietary data, and then combining it with these custom AI functions through no code applications to, to really drive, you know, very sort of unique transformative and automated processes through through application development.

Tim Butara: And it sounds like this, this becomes kind of exponentially more important, the more complex and the larger the business in question is, or the team in question is, right? You know, for, for, I imagine that for, for somebody such as AWS, their, their marketing department, if they wanted to do this without the help of AI, it would take way too long time and way too much effort to actually be feasible.

So this is like a use case where AI, you know, it doesn't just ease a process. It actually introduces a totally new kind of benefit.

Anthony Maggio: That's exactly right. And I think, you know, in contrast to, we talked about like some of the other approaches that companies have taken, like implementing, you know, co pilots or chat bot tools, I think that, you know, the difference here is that you have one person who goes and builds this AI function, right.

With their knowledge of the workflow, with their knowledge of the way that a marketing team operates. And as soon as that AI function is created, And implemented into an application, it's immediately available to all of the marketers across the organization. Right? So you're not relying on every marketer to learn how to become an expert in prompt writing, right?

Or to remember what are the use cases where I should be using you know, a chat bot or a co pilot or where, you know, what are the use cases where maybe that's not going to be the best approach for you know, for this workflow. And so. That that's really, you know, another value and benefit that we see is that you can have, you know, relatively sort of small number of, of developers who become familiar with the capabilities of AI of these LLMs.

They can do a lot of the testing. They can understand, you know, how to get the best performance out of these models and how to incorporate them in their workflows in reliable ways that are consistently yielding High quality, high impact outputs. And then as soon as those functions are developed or those automations are developed, you can scale them to hundreds of individuals across a organization who are users of that application.

Tim Butara: Yeah. I mean, that does sound crucial, especially for the use cases that we highlighted during this conversation. And it was definitely a great one, Anthony. Thank you so much for joining us. It was a real pleasure. And I think that it will be very valuable. For anybody listening because these are such important topics and it doesn't need to be a large team such as the AWS marketing team for it to benefit from these insights.

But in case anybody would like to learn more about you, learn more about Airtables products or maybe connect with you, what, what are some of the best ways to do all that?

Anthony Maggio: Yeah, absolutely. You can learn more about Airtable's products at airtable.com. You'll see there all of our AI offerings. We also have, you know, recently rolled out some new offerings specifically for product development teams, our ProductCentral offering, which helps companies manage their end to end digital product development processes with AI baked in.

So those are the best resources to go and learn more about Airtable or to connect with me individually.

Tim Butara: Awesome. We'll make sure to include it in the show notes for easy access. And yeah, Anthony, as I said, the great conversation, great guests. Thank you for joining us today. It was great having you on.

Anthony Maggio: Thanks so much for having me.

Tim Butara: And well to our listeners, that's all for this episode. Have a great day, everyone. And stay safe.

Outro:
Thanks for tuning in. If you'd like to check out our other episodes, you can find all of them at agiledrop.com/podcast, as well as on all the most popular podcasting platforms. Make sure to subscribe so you don't miss any new episodes and don't forget to share the podcast with your friends and colleagues.