<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=275133924453494&amp;ev=PageView&amp;noscript=1">

Try Solvemate without risks! We are willing to guarantee your satisfaction, or your money back

Learn more

Go beyond the chatbot hype and boost your customer service with automation to improve your CX.

Discover our features to create meaningful conversations in an instant, scalable and cost effective way.

We integrate with the best! Solvemate is built to fit into your existing tools and workflows. Find out more.

Jul 16, 2020 8:12:00 AM

Not All Chatbots Are the Same: How to Narrow Down Your Search


Chatbots have been all the hype for a while now. They can be used for a variety of situations in different industries, so there are a lot of options out there. It can be challenging for companies to pick the most appropriate chatbot for their business needs.

Many companies struggle with the fact that a lot of chatbot vendors seem to be selling virtually the same product and they all promise great results and benefits. In reality however, there are wildly different technologies and use cases behind the lofty marketing and sales pitches. It can be a daunting task for business leaders to effectively distinguish between the different products.


To get the desired results from a chatbot, we recommend taking a two pronged approach: nailing down the business problem that needs solving and understanding the technological solutions that exist to solve it. After all, chatbots are just like any other business tool: they need to deliver concrete results to be justified. Implementing a chatbot just because it is in vogue won’t get your business very far. So, if you’re thinking of buying a bot, it’s a great idea to first define the use cases. Then you can start exploring the market for fitting (technical) solutions.


We talk to a lot of business people about chatbots every day and have found that many executives feel overwhelmed by the variety of choice. Obviously, we’ll recommend our products if we think they’re suitable, but occasionally we end up recommending other providers such as Drift for lead qualification for SMEs, or CallDesk for companies that exclusively use voice-based channels, if their products are a much better fit. Most chatbots work best in specialised areas, so it’s smart to first define the areas of your company that would benefit most from automation.


At Solvemate, we’ve launched chatbots for dozens of customer service organisations in different industries around the world which all have slightly different use cases. Our experiences have taught us a number of things that can help business leaders get started. We’ve prepared a handy infographic that will enable you to make informed decisions. Our step-by-step guide explains the seven chatbot characteristics to consider before starting a chatbot project


What Kind Of Chatbot@2x (1)

1. What medium will the chatbot use?


You probably have an idea about this already, but it is an important aspect to clarify so that a number of options can be ruled out.


What channel are you trying to cover with your new chatbot? Voice-based or text-based channels? Or both? If your whole service department is based on phone service and you aren’t planning on changing that, then investigate voice-based bots rather than chatbots.


2. What is the general purpose of the chatbot?


Which department should the bot support and why, what will be its purpose? Some chatbot vendors specialise in certain verticals or departments. 


If you are looking to automate or enhance retail sales, e-commerce bots could be the way to go. Some bots, for example, specialise in guided selling. They ask questions and present the customer with suitable options from the company’s catalogues. Other bots also provide product recommendations based on the customer’s browsing history in the online shop. 


Customer support bots can take part of the workload off your service department and reduce the number of repetitive requests that have to be handled by a human. Lead qualification bots are popular amongst B2B companies. Basically, there is probably a chatbot available for every use case!



3. How much of the chat flow is rule-based vs. dynamically generated?


Depending on your circumstances, you might need a rigid system of “if-then” rules or trainable, changing responses.


Is the chat flow rule-based or dynamic? In other words, is it a fixed flow or does it change organically, based on training and mathematical calculations? 


“Rule-based” means there are hard-coded rules built into the bot that result in a chat interface with a fixed flow. This approach offers little flexibility but can be quickly and easily set up, whereas the dynamic model is more adjustable but requires assistance and a proper onboarding from Customer Success Managers to get started.



4. Does the flow change automatically based on usage?


Following on from point 3: Is the chat flow changed automatically or manually? If you choose a flow that doesn’t change automatically, you need to make sure that a member of your team is made responsible for adjusting the chat flow when needed. 


Many bot providers claim to use artificial intelligence (AI) or machine learning, but not every use case needs these technologies. Don’t get swept away by the idea of having the most cutting-edge technology: it’s more important to figure out what will work best for the use case you have in mind. If a simple, rule-based chatbot is all you really need, then a machine learning application could blow an unnecessary hole in your budget. 

5. How does the input work?


This is perhaps the most complex question of all. The input method will have the biggest impact on the usability and the success rate of your chatbot. What kind(s) of input would you like to have? How should the user interact with the chat? Should they be able to type freely or simply be offered multiple choice options? Or a mix of both? As always, this really depends on your use case. Some contexts favour free text, while others benefit from a multiple-choice environment.


Let’s say for example that a customer wants to troubleshoot a device they bought. They don’t know why it’s not working. The method of elimination—based on questions and multiple choice answers—has a much higher chance of solving the issue than if the customer were typing freely and the chatbot was searching for keywords the customer may or may not use.




Natural language processing (NLP): 

Bots that “understand” free text are also called NLP bots. They identify patterns in human language to determine an intent and reply accordingly. Users can type requests in their own words and the bot will try to find the right answers. It’s an impressive technology, but it’s not yet fully mature so the error margin is usually quite high.





Multiple-Choice Questions:

Another option is to present the user with a series of multiple choice questions that they can click their way through. This approach is usually used for rule-based chatbots with a static decision tree. In many cases this results in tedious interfaces with lengthy chat flows where users have to go through every option, even those that don’t apply.





Natural Language Processing and Multiple-Choice Questions: 

A third input method is a mix of the previous two approaches. Ideally, it combines the best of both worlds. Users can type their messages or click through questions. The bots respond with answers or follow-up questions based on probabilities and they improve with every conversation through machine learning. The result is a dynamic chat flow where users can find the answers they need faster.

6. What is the business model?


Is the chatbot provider a digital agency or a SaaS company? Digital agencies sell projects, whereas SaaS companies sell software licences. Generally speaking, these are the two types of business model on the market.


Custom developed by an agency:

These companies sell projects, not software. This business model functions like a “digital agency” for chatbots. The companies make money based on the amount of work (or hours) billed, not the software they develop. While a custom-built chatbot presents many advantages, there are numerous pitfalls and risks, so it’s often not worth the effort and the investment. Ultimately, the company buying the custom solution carries all the risks and overheads—the company selling the solution is a mere “contractor”. Maintaining custom built software typically involves continuous spending and resource allocation and a much longer timeline for achieving a positive ROI. However, specially developed chatbots might be a relevant option for companies or industries with high security concerns.



Companies selling chatbots as SaaS products are selling licences for their software. In practical terms, it means they have already developed a standardised piece of software that just needs to be configured for their customers’ use cases. These configurations can be carried out by either the SaaS company or the customer themselves. The pricing model is based on how much the software is used—not the number of hours spent on, project management tasks, meetings or development work. While out-of-the-box solutions have their limitations, they tend to be more profitable in the long run, as they can yield a positive ROI sooner. Unlike custom solutions, SaaS solutions can save costs from day one, as the customer does not have to bear the overheads of the actual development work. 


Many SaaS providers also offer shorter contracts or opt-out clauses, as well as tools for solid ROI calculations. So the commitment is manageable if the chatbot doesn’t end up working for your business.





7. Is the solution applicable to all industries or only specific verticals?


As different industries and departments work with different tools and integrations, your chatbot must be able to be integrated with the relevant technologies. So, is the chatbot you are looking at specialised for one industry (your industry?) or is it applicable for all industries?


You know the service stack your teams work with, so it’s important to check whether your future bot can be aligned with the software programs they are using. Again—make sure the bot in question is actually compatible to your use case.



The best chatbot is the one that best suits your use case


When looking at different chatbot vendors, it is important to check whether they are actually applicable for your industry, your customer service stack, your customers, the way your company communicates and the contexts in which you are communicating.


Many chatbot providers claim to make bots based on AI or machine learning. Not all of them actually do. Plus you may not actually even need these technologies, depending on what you’re trying to achieve with your chatbot.




Chatbots can deliver real benefits that will improve your customer service and save you time and money. But if your chosen bot is not cut out for what you want it to do, it could be an expensive mistake.


So before spending money on an agency or a software, we recommend using our infographic to define your chatbot’s key characteristics and KPIs, and making sure you understand the scope of the maintenance work that your choice will entail. Good luck!


Anna is a former Solvemate and workout enthusiast in her free time. When she isn’t busy attending fitness bootcamps, she's probably cooking Indian food or discovering new restaurants in her hometown Berlin. Being a true geek, she favours reading over dancing at Berghain. A Londoner at heart, she still embodies British chic and enjoys a good cuppa (caffeine-free) tea.