Chatbots, or “conversational agents” are software applications that mimic written or spoken human speech for the purpose of simulating and interaction with a real person. Chatbots are most commonly used in the customer service space and are becoming more common partly due to the fact that they can provide round the clock accessibility and highly sophisticated human-like conversations.
Chatbots are becoming smarter and easier to install and use. We explored different kinds of chatbots to see what problems chatbots can solve and who should use them.
“Chatbots can reduce customer service costs by up to 30%.”
Types of chatbots
1. Problem-solving chatbots allow human chat agents to spend more time on difficult problems as opposed to handling inquiries that often can be answered by referencing guides on the website (FAQ). These chatbots provide:
automated answers
instant resolutions to clients
24/7 support to eliminate wait times and boost customer satisfaction
2. Chatbots that start a personalized conversation to understand the visitor’s intention and guide them to the best human resource for further assistance, like the Guide Bot. Chatbots can further:
provide clients with triaged options
connect visitors with the correct human resource
support specific job functions, like help book more meetings or boost lead generation
3. 100% customizable chatbots, like the Chatbot API, allows teams to scale chat conversations without inflating a human team, keeping conversations personalized and meaningful. This “bring your own bot” approach can:
automated answers
instant resolutions to clients
24/7 support to eliminate wait times and boost customer satisfaction
4. Chatbots to build trust with site visitors by asking for only the information needed to fill an inquiry. They work to:
While chatbots are all the rage, research needs to be done to accurately choose the best chatbot for your team. Critically assess what your team needs and what technological resources you have. Don’t be afraid to ask the hard questions!
Do your clients tend to ask easily answered questions or require more in-depth assistance?
How important is the voice and tone of the chatbot?
What resources are you prepared to put towards a chatbot solution?
Keep your site visitors and clients in mind throughout the vetting process so you can continue to deliver the best experience possible.
Growing with chatbots
Treat chatbots as a member of your team that needs continuous training to ensure continuous improvement. Research solutions fully and test when possible. Understand what your team needs from an internal standpoint so you can ask the right questions to chatbot providers.
We recommend keeping stats on both your chat agents and chatbots so you can compare customer satisfaction scores and see what clients are really asking your chatbot. Continue to refine your chatbot process. Chatbots are flexible and designed to fit different needs, use this to your advantage.
Chatbots + humans
Chatbots cannot be expected to perform the same duties as a real human. It is widely recommended that chatbots be used in conjecture with a human chat team. Doing this will allow your human team to focus on the important issues and increase your website conversion rate. Together, chatbots and humans can:
Serve more clients than ever before
Have more precise and meaningful conversations
Focus on the important issues
Drive conversion rates
Improve customer satisfaction
Keep learning
Ask questions and be curious. Companies using chatbots effectively enjoy elevated customer service and automated business processes. What is right for your team today may not suit your business needs in the future. Continue to refine and rethink chatbots in the marketplace to make informed decisions. Ready to begin?
In May 2018, the General Data Protection Regulation (GDPR) came into effect in Europe. Even firms that prepared early for the changes agreed that it’s a work in progress to design a data governance plan. Business experts shared some of what they learned post-GDPR at aSeptember 2018 conference in London:
Spotify realized that the best quality check was through looking at its data governance plan from a data subject’s trust perspective.
Adobe discovered that data subjects were more willing to provide data when the company was more respectful of how the data was used.
Omnicom Media found that to design a data governance plan is a continuous evolution rather than a milestone.
Dentsu Aegis Network determined that the data regulation landscape is ever-evolving as legislation and lawsuits set new legal precedents.
LiveRamp UK commented that a privacy-by-design approach is a better universal standard.
The overall insight is that, to design a data governance plan, the data subject must be approached as the central focus — and that the plan should be agile to respond to changes.
Ask important questions about your data
Businesses can take advantage of these insights to design a data governance plan that doesn’t just focus on data value, but also values its data sources. The plan requires solid preparation; lay the groundwork by asking essential questions about your data.
Whose data do you have?
You may have data of European and California citizens.GDPR acknowledges that it’s the strictest privacy policy in the world. TheCalifornia Consumer Privacy Act (CCPA) is right behind it. To make sure you’re in compliance, you have to sift through all of your data. You’ll want to know whose data you have and why you have it — which allows you to determine if you have quality data in the first place.
A study byMIT in 2017 estimated that companies lose 20% of their revenue because of data quality issues. Spring cleaning is mandatory to design a data governance plan that provides opportunity to prune away low-quality data.
Do you have a data subject-centric mindset?
The CCPA and GDPR are built to protect the data of the very stakeholders that drive business. Companies have ten days to respond to deletion and “right to know” requests by data-subjects under thedraft rules of the CCPA. You’ll need to be able to quickly access all of the correct applicable data and respond to consumer requests. When you design a data governance plan, think data subject-centric rather than merely data-centric.
Where is your data?
It’s necessary to identify data that exists in multiple systems, platforms, or storage locations. This requires assessing current data flows as well as quantifying and mapping its management. Assess current data flows along your data supply chain through the collection, processing, distribution, integration, storage, and deletion of data.
Quantify and map present data management and governance — from customer-facing processes to back end office functions. The areas that extract data value and those that are liable for data risk, in particular, are guidelines for razor focus.
Why do you store data?
Ask the important questions about your data sources, data quality, data regulatory risk, and data tech and tools. Also, revisit permissions of all data citizens (any employee who handles data). What service providers or vendors do you share data with — and why? While taking stock of the reasons for data, keep in mind the five “why’s” of data governance.
The 5 “Why’s” of data governance
1. Data regulation
There is no way you can ensure GDPR and CCPA compliance unless you maintain efficient data governance and data management systems. Whether or not you collect data from citizens of the EU or California, expect this to be the name of the game to stay competitive.
Design a data governance plan persistent in the face of any evolving privacy laws. Think long-term to create a data governance plan poised to scale while staying flexible in anticipation.
Takeaway:Be certain you understand the scope of these privacy laws.
2. Data integration
There are few lines of business that aren’t data-driven in some respect. All lines of business must now take ownership of data, this kicks off with education. If data owners, data stewards, and data users don’t understand the importance, they likely won’t be too invested. All internal data stakeholders must be equally invested. Only then can fluid communication thrive in order to meet the response deadlines required by data privacy laws.
Takeaway:Educating internal stakeholders is a fundamental aspect of a robust data governance plan.
3. Data volume
Massive data volume increases risk — both through data breaches and inability to comply with privacy regulations. As the amount of data increases, scalable master data management (MDM) systems become essential. Creative data management, and lean systems built to trim, are critical for unhindered growth. These systems cut unnecessary data and retain only quality data with inherent business value. Before you can structure a plan, you need to know where all your data is, what it is, and why it’s necessary.
Takeaway: Data inventory is a prerequisite to design a data governance plan.
4. Customer focus
From now on, customer data must be treated as though it’s borrowed – you no longer own it. All data now expressly belongs to the customer, and it’s the customer’s right to instruct on what they want you to do with it. All organizations must respond to a customer’s directive regarding their data, and in real-time.
You’ll also need to vet any vendors that process data for you. Under the GDPR, if they aren’t compliant, then you aren’t either. Likewise, CCPA necessitates revisiting your vendor contracts. Map your vendors and define your role with them ahead of time.
Takeaway:Cultivate customer empathy into your design.
5. Tech
A good data governance plan not only takes advantage of available tech, but also searches for areas where new tech can automate scalable data processes. As technologies evolve, the role of AI, machine learning, and other emergent technologies in data management and governance, will increase. Assess your technical assets and decide where adopting additional tools can improve data flows and processes.
Takeaway:Research new tech and be prepared to budget for future tech and tools.
Which data governance plan is best?
There is no one-size-fits-all template for designing a data governance plan and team. The data governance model you design must be flexible enough to mold into your business framework as it exists today. It could be disruptive to impose a model out of convention, especially if it’s not complementary to your business structure.
However, this is not to say that you don’t want to emulate structures that might work well with segments of your business. You are free to customize. Take advantage of the experiences of companies that have walked this path before. Explore data governance plan theory and research data governance plan frameworks that have already been deployed.
What are data governance models?
While the strictness of the CCPA and GDPR are relatively new, data governance is not. Data governance, like many business models, has gone through a variety of iterations. Most are variations of the three most common frameworks:
Top-down — centralized
Bottom-up — distributed or decentralized
Hybrid — centralized control with distributed management/decentralized ownership
The shape of the plan that works best for your company depends directly on your industry, size, business model, and business culture.
For example, a smaller firm with a traditional business model and hierarchical business culture may fit well with a top-down, centralized approach. But for larger firms that practice agile methodology, decentralized ownership may be much more effective and scalable.
The current state of any master data management (MDM) model and governance initiatives already in place will also be instrumental when designing a data governance plan.
Who is your data governance team?
Examine the architecture of your firm in preparation for designing a data governance plan. You’ll want to visualize the integration of adata governance team. The design of your governance plan and governance team are not mutually exclusive.
Highlight existing stakeholders that could be essential candidates for roles in your data governance team. How you assign those roles and titles should be in line with your company’s architecture, rather than by convention.
Set data governance plan priorities
After you’ve assessed your current set up, reviewed key objectives, and explored data governance plans and team models, it’s time to set data governance priorities and goals. It’s critical to take specific steps before you can move on to others.
Until you’ve mapped your data, you can’t identify areas of liability in data storage, management, or governance, a prerequisite to moving forward with your design. Estimate time and human resource costs for each goal so you can create a roadmap with achievable milestones. A timeline with reachable objectives will serve you well if you make a proposal to the executive level to gain sponsorship.
Use the GDPR and CCPA as templates for future data privacy law
Spotify and Adobe discovered that approaching data governance is better done with empathy for the consumer. Omnicom and Dentsu realized that data privacy regulations will continuously evolve with legislation revisions and lawsuits setting precedent. The GDPR and CCPA set the bar fairly high on data privacy laws, and we can expect ensuing regulation to follow their lead.
Consumers are aware of these laws. Transparency about your data governance approach is paramount. An honest appraisal of your current data management and governance systems, while keeping an eye on the rights of your customers, and the other on regulatory policy, is an essential first step to design a robust data governance plan.
Here is the update with what the team has been working on in the last weeks:
New Answer Bot Features:
• We have added an integration with Salesforce Community Knowledge with our Knowledge Base and Answer Bot options. This can now be enabled under the Integrations -> Knowledge Base tab additionally to Salesforce Solutions and the classic Knowledge options.
• We have added a preliminary Analytics tab for the Answer Bot:
We are currently working on enhancing this report with additional options to view and drill down into a list of search terms that were entered by your visitors. Stay tuned for these updates.
• The multi-step Guide Bot and Answer Bot options are now officially released and no longer considered ‘Beta’. Details on the new and improved functionalities can be found in this FAQ article.
Updates:
• We added a new option to update the agent links in Hub when a visitor entered their email address during the chat with a {hubVar:email} variable. You can read more on how to use this feature here.
• The Info Capture Bot now has an Alias option to display to the visitors, similar to the other bot options.
• Improved the UI of the multi-step Guide Bot to not show the input area to the visitor in between steps
• Improved the UI of the visitor chat box to always show a semi-transparent scroll bar.
Resolved Issues:
• Fixed an issue where some messages were not logged consistently in the chat transcript
• Fixed an issue where agents responses were not visible in the visitor’s side after they attempted to upload a file of an unsupported type.
• Fixed an issue where agent links in Hub were not updating when toggling between chats.
• Fixed an issue where a chat coming from the Guide Bot was not set to offline when no agents were available
• Fixed an issue where an agent who was also the account owner could not transfer chats
• Fixed an issue where SnapEngage cookies showed a Chrome warning about a cross-site resource with SameSite attribute
• Fixed an issue with minified SnapEngage code where the AMD resolver failed in certain configurations.
• Fixed an issue where a close button appeared in the chat after interacting with a Bot.
• Fixed an issue where a SMS case sent even though “do not send chat transcript to destination” option was Selected.
our team has been working hard on updating our bot offering with some new and improved functionality. Please see this help article for more details, such as:
the new multi-step Guide Bot:
Guide Bot users can chain together up to 100 logical steps to generate more visitor engagement with contextually appropriate messaging and guide customers to the right solution instantly.
Create complex dialog trees and deliver highly-personalized capabilities (i.e. dynamically inject content into bot messages with JavaScript variables).
You can also use Guide Bot’s multi-step dialog to label incoming conversations. This reduces manual agent work and allows users to trigger automated workflows in integrated CRM or Help Desk systems.
Enterprise users can even use Guide Bot to dynamically route chats to the most appropriate agents who specialize in specific skills using Agent Tags.
It’s also possible to use Guide Bot to route incoming chats to additional bots for advanced customization and flexibility.
Other Bot Updates
The new Guide Bot –
now features a ‘Live Preview’ option to test the flow of the bot steps,
includes a configurable offline message option similar to the Answer Bot,
will not take up a chat agent seat anymore but a separate bot seat which can be added in your subscription page.
The pre-chat prompt which can be configured in the Design Studio tab also includes the Guide Bot avatar and alias similar to the Proactive Chat functionality.
The Chat Agents tab will show the enabled/disabled state of the configured bots in the overview.
Adding bots now does not automatically add a Priority Tier anymore.
The “Approval Bot” functionality was renamed to Approval Checker and removed from the Chat Agents tab bot overview. The functionality remains the same.
Other Updates
Chat Routing: Transferred chats which are broadcasted to another widget, broadcasted from the bot or an agent to the same widget will now also trigger the auto-responder functionality if no agent responded in the predefined time frame.
We have updated the automated ‘thank you for your message’ email to be sent from [email protected] to comply with DMARC regulations.
Fixed an issue with the visitor JavaScript code snippet not being properly minimized to improve loading speeds.
Fixed an issue where the post-chat survey was not showing for API based bots.
Fixed an issue where not-responded chats were not showing in the Analytics Wait Time report.
Fixed an issue where CSV exports for sub-admins were not taking the correct comma or semicolon setting into account.