How To Leverage GPT-4 In Your Internal Business Processes: An AI Guide
December 6, 2023
April 18, 2023
Recent advancements in generative AI—like GPT-4—are changing what’s possible for companies and people to achieve with technology.
Company leaders and internal teams are, in turn, working overtime to find a way to harness the power of these tools in a way that maximizes their potential, which is not only to change what’s possible for people to achieve with technology, but to transform how we work in general.
As with other waves of technological progress, it seems likely that those who succeed—who adopt and adapt—will thrive, while those who don’t will fade away.
This post will detail how you can be sure your organization and team does the former, specifically by building tools like GPT-4 into the infrastructure of your day-to-day business processes.
A paradigm shift in process improvement—and in how people interact with technology
First, a word on why exactly generative AI like ChatGPT and other large language models (LLMs) are so powerful. It has to do with their unprecedented capacity for resolving employees’ problems.
This capacity represents a paradigm shift in how people interact with and utilize technology. Take search engines, for example. For a long time, people have turned to search engines when they have questions or problems. What do search engines return? Research. They organize information. This is fine, but it then puts the onus of figuring out what to do with that information onto us.
The power of AI is it allows end users to leap frog from research to something far more useful: resolution.
And not just any kind of resolution, but resolution that’s differentiated, rapid, personalized, and context-aware.
The paradigm shift in the broader technology landscape hinges on this huge experiential difference between research and resolution. At the end of the day, that’s what people truly want and need out of technology—for their tools to understand and promptly resolve their requests, questions, and problems. Now that people are beginning to feel that difference, there’s likely no going back.
People are going to demand this functionality moving forward, which is why AI is going to upend enterprise software. Companies that don't adapt will die out. In order to improve process experiences—and in turn increase process adoption, maximize compliance, and maximize efficiency—this is the next logical step.
The question then, is how?
How to leverage GPT-4 in your operations? Build it into your core internal processes—all your processes.
The truth is, to get the most out of generative AI, you need to leverage it operationally. That is, you should make it a core part of your internal processes.
Why? The answer’s simple. To get the absolute most out of AI, you need to leverage it in a way that’s structured, accessible, and repeatable at scale. That comes down to building it into your everyday essential processes. In many ways, core internal processes make up what companies actually are. Certainly, as an organization, you can only be as innovative, resilient, and adaptive as your processes are strategic and effective.
So how can you begin to incorporate something like GPT-4 into your day-to-day operations? That’s something we’ve been helping organizations with at Tonkean for some time now. Earlier this year we released a suite of capabilities aimed at helping internal service teams—like legal and procurement—build the power of GPT-4 into their key internal processes: LegalGPT, ProcurementGPT, and InvoicesGPT.
These tools equipped procurement and legal teams with the ability to create workflow automations that, in the case of procurement, are capable of doing things like fully automating all tail-spend processing, strategic sourcing, and “3 Bids and a Buy.”
ProcurementGPT does this by automatically identifying appropriate purchase categories, analyzing vendors’ websites, extracting their benefits and capabilities, suggesting alternatives, and generating RFI/RFPs—ultimately reducing unnecessary spend. (LegalGPT, for its part, fully automates the triage and intake of all incoming legal requests, including routing them to a human when special circumstances arise. InvoicesGPT connects to any source you need, such as your email inbox, Drive folder, and so on, to automatically handle all incoming invoices.)
Also of note: what these tools all have in common is they not only expedite task completion—rather, they do it automatically, and according to your own predefined processes and rules. Tonkean wraps around whatever systems and policies your organization already relies upon, which eliminates the need for change management. And by design, Tonkean provides tracking for all processes with dashboards, which ensures accuracy, completion, and compliance.
All of which is an example, we think, of building the power of GPT-4 effectively into the backend of your internal processes—equipping internal service teams themselves with the superpower of far more rapid, intelligent, and capable process resolution.
But what about the experience of employees and requesters on the front end? To truly harness the power of AI in your everyday operations, don’t you want to leverage its potential for process resolution into your process experiences end-to-end?
What do you do when you want to learn something, dig up the answer to a question, or find something on the internet? You Google it.
But why is that? It’s because that’s the simplest, fastest, most frictionless way to get what you need. A search engine is like a “front door” to the internet. It’s a portal to all the world’s collected digital information, and to access all of it, you go to that door and waltz right on through.
Inside most organizations, employees and requesters don’t enjoy anything similar. To get what they need from Legal when they need a new NDA, for example, they have to jump through hoops, fill out forms, learn how to navigate new interfaces, or otherwise suffer tedious process experiences that pull them away from the work they really want to and should be focusing on.
The Tonkean AI Front Door changes all that.
The Tonkean AI Front Door is an AI-powered intake experience for submitting—and proactively resolving—internal requests. It’s a tool with a perfectly simple interface that belies its incredible power. It looks like a search bar or a chat window nestled on a homepage. But it’s so much more.
It’s capable, among other things, of the following:
Meeting employees where they are. The Tonkean AI Front Door supports frictionless service interactions with a plain-language user experience—and is accessible to requesters via email, Slack/Teams, or a customized web portal. (Meaning they can access it from wherever they like to work.)
Providing intelligent assistance that’s personalized and context-aware. The Tonkean AI Front Door delivers tailored support that understands employees’ intentions moment-to-moment. Customize response patterns for your unique requirements and policies. Automate actions on existing platforms and extract both structured and unstructured data. Apply comprehensive role-based access control to response output.
Delivering a flexible experience without compromising on policies. The AI Front Door provides AI-guided workflows that can open tickets, launch processes, and bring the right people into the loop. Enforce policies with a no-code, fully auditable toolset. Save time and effort with intelligent issue triage. Streamline process by orchestrating across existing systems
Anticipating and resolving requests before they come up. Finally, going beyond intake, Tonkean monitors applications and data sources to predict requests based on historical patterns. Preemptively reach out to users with suggested next actions. Provide instant resolution for simple requests. Launch cross-platform orchestration sequences for complex requests.
How you actually use the AI Front Door will differ depending on your existing policies and systems, but here are a few common ways it works in practice:
Document / Content creation. Example: Type "I need an NDA for ACME" into the Front Door and Tonkean springs into action, checking for existing documents, generating a new one if needed, and even initiating e-signature workflows with a single click. (See above.)
Purchase request. Example: Request a new MacBook and Tonkean instantly identifies your purchasing policies, suggests options from your catalog—Ariba, let’s say—and auto-fills necessary information to kick off the process seamlessly.
Data lookup and drill down. Example: Ask about the status of a deal and Tonkean rapidly searches connected data sources, presents the most relevant information, and suggests follow-up actions, all within a single conversation. (Tonkean remembers what you’ve already told it, and so if you want Tonkean to create an NDA for the deal, it will trigger that workflow using the data you’ve already provided it in the thread.)
Deflecting requests intelligently. Example: Request legal to review a contract via email, and Tonkean checks if a signed copy already exists, providing it as an alternative before diving into a new review process.
Proactive resolution. Doors open more than one way. In the event that you’ve changed the status of a new deal inside Salesforce—but did not take the sort of additional action you’ve taken in similar instances in the past—Tonkean will reach out to you inside whatever system you spend most of your time in. Example: Tonkean notices you've changed a deal's stage and, based on past behavior, proactively creates an MSA, asking if you need any revisions or changes, saving you time and effort.
A new standard for AI-powered business process automation and a new future for business operations
Compare the above experience with that of searching for something on Google. A Google search will offer up links and leads that you can follow.
But Tonkean’s AI Front Door understands what you need, gives you the pertinent forms and documents, automatically kicks off related processes, and holds your hand until it’s all resolved—and in a manner that’s specific and customizable to the particulars of your organization.
Another useful comparison: How do you request a new laptop from your company currently? What form do you need for a deal NDA, and where is it located? How do you check the status of a deal your team is working on?
You probably don’t know. And most of the people in your organization probably don’t, either. Most likely, the answers to those questions reside inside someone’s head, or maybe in a document or spreadsheet somewhere. Immediately, you have to start asking around or dig through your email or Drive folder or Salesforce entries. But all of that requires time, brainpower, and friction.
And now it’s all completely unnecessary because of the AI Front Door, which reduces the time it takes for employees to go from request to resolution from hours or sometimes days to seconds.
The answers exist, in the context in which you need them. You just have to knock on the front door to get to them.
A boon to process experience.
Among the many benefits of leveraging AI in this way inside your organization is its potential for improving the felt experience of interacting with your core internal processes.
At a high level, building AI into your everyday essential processes is important because it’s the only true way to get the most out of this powerful new technology—which power we’re really only just now waking to the true potential of.
But it’s also important to tap into for its potential to transform process experiences—and, in turn, categorically increase process adoption.
Inside organizations, internal process adoption is generally very low. In legal and procurement, for example, employees often ignore internal processes altogether. The reason for this is very simple: Most internal processes are actively challenging to follow. They’re complex, fragmented, and unappreciative of the needs or preferences of the end user.
This is not a small problem. Compliance, transparency, and efficiency depend on high process adoption. Low process adoption, in turn, hinders efficiency, reduces visibility, lowers productivity, slows time-to-value, and increases risk.
How, then, can internal teams begin to increase internal process adoption? The only way is to improve the experience of engaging with those processes—specifically by creating process experiences that are both easy to follow and that provide real value to requesters.
There’s no better way to do that than with AI, and more specifically with tools that allow you to put structure around AI’s power.
There’s no magic box—to get the most out of GPT-4, you need structure and strategy.
It’s true that the AI Front Door is so effective in part because of GPT-4’s incredible abilities, but like all AI, GPT-4 requires data to function and structure around it to deliver consistently cogent results.
That’s what incorporating GPT-4 into your operations effectively and at scale demands. And that’s precisely what a tool like Tonkean now provides.
Too often, people think of AI as a magic box. Generative AI like ChatGPT certainly sometimes seem like magic. But the reality is that AI works best when it has narrowly defined tasks and pristine data fueling its algorithms.
That’s Tonkean’s approach to this AI Front Door. It’s fed by all of your company’s data, contacts, documents, defined processes (which you can make in Tonkean), and connected applications (like email, chat tools, CRM, and so on).
In a way, you can think of Tonkean as the essential structure you need to convert that data and the raw potential of GPT-4 into something that serves an operational purpose.
The importance of such a structure cannot be understated. The chatbots that represent the primary way people have experienced the power of GPT-4 so far cannot drive process automation or digital transformation on their own. As cool as it is to have a chatbot pass the bar, it remains basically a toy.
But instrumented strategically into your operations, it goes from a toy to something like a steam engine—something that makes possible a whole new way of doing business.
Whether it’s Tonkean or some other platform like it, this is the kind of structure you should be supporting your AI adoption efforts with. You can be sure that the organizations succeed in digitally transforming their operations with AI—and who end up designing the future of work and of our economy in the process—will have done precisely that.
Legal intake is inevitable, but managing it doesn’t have to be a slow, painstaking, manual process. With AI, automation, and intake orchestration, you can streamline intake processes and reclaim hours of time and effort.