Today, I want to concentrate on the absolute lowest hanging fruit for deploying generative AI: Ticket Deflection
But first, let’s get some background:
Currently, generative AI is all the rage, and every company is scrambling to test, prototype, and deploy generative AI projects.
The type of projects range from the very simple — like using ChatGPT — to the ultra complex — integrating ChatGPT-type AI with PII data and workflows.
Some use cases are simple like “Let’s use ChatGPT for X” and some are complex internal flowchart-type workflows like “Let’s try and get AI to do what Bob does”.
While these complex projects are admirable — and some of them are indeed pretty cool — it never hurts to show off to your boss a really cool demo.
However, in real life, most companies require immediate deployment and a clear return on investment for investing in generative AI.
And the return of investment is clearly there — IF — and that’s a big IF — a clear deployment blueprint is followed.
Why Should Ticket Deflection Be First?
Helpdesks are a pain for most companies because 90% of customer pain points can typically be answered by content that has already been generated and is available on the knowledge base, website, forums, or other knowledge sources (like Slack)
They are a pain for customers because customers don’t have the luxury of navigating your website and going through a needle in a haystack to find the answers they want.
No customer wants to navigate your website like a maze when they would rather prefer getting the answer in one shot (like their experience with ChatGPT)
What Is the Deployment Blueprint?
Let’s look at the three main considerations to get this deployed at your company.
Consideration 1: Return On Investment
When generative AI is deployed on help desks and company websites, the customer is clearly able to help himself rather than in creating senseless tickets.
This means that it greatly reduces the ticket volume and the time to resolution for customers.
In industry parlance, this is called “ticket deflection”.
And the clear ROI is: Time saved due to reduced ticket volume.
“Our ticket volume is down drastically since we implemented CustomGPT.ai on our website and helpdesk sites. It has made a material impact on our Support team. We’ve informed our private equity investors and they will be rolling this out across their portfolio”
Large PE-backed SaaS company.
Happy customers that are quickly able to find what they want, resulting in a higher NPS score.
As a case study, our company used to get 300 tickets a day to our helpdesk. We then deployed the chatbot on our website, which has indexed all of our public information like the website, helpdesk articles, guides and YouTube videos.
The volume of tickets has reduced by 90% — with a large fraction of the queries being answered by the bot.
This has freed up precious customer support resources to handle only the most extreme cases, which is good for the company.
Consideration 2: Employee Efficiency
When a customer support agent is handling a ticket, he himself has to browse through company databases to get to the resolution. He has to search through knowledge bases, dig for information in PDFs and guides to research and find the information he is looking for.
Debugging a customer support ticket takes time — trust me. I’ve suffered with this for months now. Anyone in customer support will tell you the pain endured to debug a ticket.
This results in a lot of time being spent to resolve a ticket. By making an efficient generative AI chatbot that knows ALL your company’s information AND making it accessible for internal use of employees, you get to immediately see the benefits — before putting it in front of customers.
And best of all, the bot is always up to date. It always knows the latest information — including those 3 blog posts the marketing department posted yesterday.
It can ingest and index every public piece of content (like that video of the CEO that was posted on Youtube yesterday).
And it is available 24/7.
The bot does not take a break, doesn’t go to sleep, take vacation days or stop on weekends.
And the bot speaks 92 languages.
24/7 availability around the globe in whatever language your customer speaks.
Consideration 3: Near Zero Risk.
Customer support is one of those rare things where the data and content is already public.
The knowledge base articles (like Zendesk) are already public.
The website is already public.
Every customer facing asset that you have, from PDFs to videos, to webinars, to audio to guides are all public information already.
Google has probably indexed all of this content anyway.
And so this makes it a near-zero data risk for your business.
Why is this important?
Because your security officer will have virtually no problem deploying a chatbot in this case, because the data is already public.
This makes this deployment blueprint one of the easiest, low-cost, low-risk, high-return deployments in generative AI.
Ninja tip: This near-zero risk profile when paired with CustomGPT.ai’s industry leading anti-hallucination solution is exactly the type of initial AI pilot that most companies are deploying.
Step-by-Step Deployment Guide
Well this is all good — so now how to go about deploying you ask? So here is a step-by-step guide to get you your 1st GenAI “win”.
Step 1: Build The Chatbot
You can build your chatbot — with no-code and zero IT resources — with your public information like website, public PDF documents, and help desks like Zendesk.
With an easy no-code tool like CustomGPT.ai, you can ingest content from your website or helpdesk and the content can be indexed and incorporated into the chatbot’s knowledge base.
You can even upload PDFs like product manuals and other content assets you have into the chatbot’s data index.
Ninja tip: This can be done with no-code – which means you dont need expensive developers or technical people. Any employee can get this done, without requiring technical expertise.
Step 2: Internal Use
Once the chatbot is built, run it internally with your employees — specially the customer support staff — and get them to love it.
No form of deployment is possible until you make champions out of internal users.
Every deployment we have seen starts with employees using the chatbot, falling in love with it, seeing the benefits, and then expanding usage of generative AI around the company.
Step 3: Customer Facing
Once the employees and internal evangelists share their feedback and ratify the chatbot’s capabilities, then — and only then — put it in front of customers.
Remember: You need some internal champions who can see the value of “ticket deflection” and customers helping themselves.
At that point, your ROI will really start flowing in terms of:
- Time saved by customers
- Reduced ticket volume
- Quicker time-to-resolution.
And then, if you are really smart about what you want your chatbot to do, you can use it to drive leads, sales, and other revenue drivers of your business.
Step 4: Customer Analytics.
The beauty of putting a chatbot in front of customers is that their conversations give you a goldmine of information that can then be analyzed to derive key insights relevant to your business.
These insights can be utilized by every aspect of your company from:
- Customer Support: See what content is missing — resulting in the bot responding with “I don’t know”
- Product Marketing: See how customers are reacting to product features — specially how they are phrasing the pain points in their own words.
- Product Development: Customers are clearly telling the bot what features they want — this is especially effective because the bot is “judgment free”. Customers dont mind telling a bot things that they would never tell a human.
- Inbound Marketing : The queries in the bot serve as a goldmine of information for blog posts, SEO and other inbound marketing efforts.
- Executive Decision Making: Insights from the chat logs can be used to decide roadmaps, overall customer sentiment, NPS and other metrics.
- Legal and Compliance: The chat logs can be analyzed from a “compliance” viewpoint to see if the chatbot is causing any AI risks.
This simple deployment blueprint, which can be executed in hours, is the absolute lowest hanging fruit in generative AI right now.
And this is being ratified — as we speak — by hundreds of live deployments that we are seeing in the field.
High ROI, low cost, near zero risk. This is the magic combination.
That makes this the absolute lowest hanging fruit in generative AI.
While large companies struggle to deploy the simplest of generative AI use cases (aka: the boil-the-ocean approach), smart companies that are taking this crawl-walk-run-fly approach are reaping the benefits.
These early “wins” and learnings from deployments will then allow “early adopters” to create future advanced uses for Generative AI — giving you an edge against competitors.
Case Studies
To see real-life case studies with results, see these examples:
Dlubal Software
Dlubal Software is a global leader in advanced structural analysis and design software used by structural and civil engineers.
At Dlubal, we are constantly looking for ways to optimize our processes and improve the customer experience. The development of an assistant using Customgpt’s facilities was a crucial step forward. The assistant has enabled us to offer 24/7 support while improving accuracy and speed of response. This has led to a noticeable increase in customer satisfaction and even faster support. At the same time, our support team has seen a significant increase in the efficiency of our customer service. The assistant provides precise answers in real time and significantly reduces the workload of our support team.
George Dlubal, CEO at Dlubal Software’
Read the full case study here >>>
Divorce Online
Online Legal Services Limited, founded by Mark Keenan, runs Divorce-Online. This UK-based service offers a quick, affordable, and stress-free online divorce process.
“We have been able to reach a lot more customers out of hours and sales have doubled due to the excellent responses given.”
Mark Keenan, CEO & Founder, Online Legal Services Limited
Read the full case study here >>>
White Papers
If you are the technical type, here are some white papers we have written that should help you deploy AI at your company:
- The Absolute Lowest Hanging Fruit In Generative AI Is… — Published in Artificial Intelligence in Plain English
- During Off-Hours, The AI Takes Over Customer Service — Published in Towards AI
- 18 Roadblocks To AI Adoption — Exclusive Surveys & Exec Interviews — Published in Towards AI
- AI Chat Logs — The Hidden Goldmine Your Company Hasn’t Discovered Yet. — Published in Operations Research Bit
- Delete your legacy chatbot — Improving engagement and customer experience with GenAI. — Published in Bootcamp
What’s Next?
To get started with deploying this blueprint, here are two options:
- Self-service: You can start a 7-day free trial on one of our subscription plans. This option is more suited for do-it-yourself small businesses.
- Talk to Sales: For larger companies that prefer a sales-assisted AI pilot (involving solution design, procurement, legal, dedicated support, etc), our Enterprise plan is recommended.
