Avaya Connected Consumer Research | September 2026
Customer Experience on Their Terms
The new rules for AI, human expertise, and customer control
Consumers want less effort, connected conversations, and useful AI. They also want a say in how it all works. Explore the expectations shaping the next customer experience.
26 survey questionsU.S. consumers, age 18+September 2026
The story in five shifts
What customers want. What enterprises need to connect.
Choose a theme to explore the evidence and what it means for the service journey.
Shift 1 · Resolution
Solve the problem. Reduce the work.
Fewer than half of consumers say their latest issue was resolved during the first interaction. Waiting, transfers, and repetition make the journey harder.
What it means for enterprises
Design around the complete resolution journey. Connect the knowledge, people, and systems needed to finish the task.
Was your most recent customer-service issue resolved?
Yes, during the first interaction
47%
Yes, but it took more than one interaction
31%
Partially resolved
9%
No
10%
I do not remember
3%
Single response. Q1 · n=577 · p. 2
Shift 1 of 5
Put the research into practice
Same request. Different rules. Different experience.
Illustrative scenarios · not a product demo
Choose a situation and the authority you give AI. Then step through the experience—or bring in a person.
Go beyond the headline. Explore all 26 questions, compare response patterns, and jump to the source page.
QUESTION 01n=577Source: p. 2
The Resolution Gap
Think about the most recent time you contacted a company for customer service. Was your issue ultimately resolved?
Yes, during the first interaction
47%
Yes, but it took more than one interaction
31%
Partially resolved
9%
No
10%
I do not remember
3%
What the finding tells us
47% say their most recent issue was resolved in the first interaction; another 31% needed more than one interaction.
Percentages reproduce the report’s displayed chart values. Rounding can affect totals; questions allowing multiple selections can exceed 100%. Each finding identifies its respondent base and source page.
Enterprise implications
Build around the outcome. Keep the customer in control.
The report points to an operating model that connects AI, people, context, and workflows around each customer’s needs.
01 · Design for resolution
Connect the work behind the interaction.
Bring together knowledge, customer context, routing, and workflows so the customer can make progress without restarting the journey.
02 · Make authority explicit
Match AI’s permissions to the task.
Define which actions need approval, what can happen within pre-set limits, and how customers can review or recover from an outcome.
03 · Equip human expertise
Bring the person—and the context.
Make knowledgeable people easy to reach. Carry the customer’s intent and previous conversation forward so human help starts with understanding.
Avaya Infinity®
Orchestrate the enterprise behind the experience.
Avaya Infinity connects systems, data, AI, and workflows to help enterprises turn interactions into coordinated action.
About the research, methodology, and this experience
Study scope
Fielded in September 2026 using an online consumer research panel and a structured 26-question survey of U.S. adults aged 18 and older.
Questions 1–3: 577 respondents.
Questions 4–26: 558 respondents.
Sample balanced across age, gender, and region.
U.S. residents across the 50 states and the District of Columbia.
The findings provide directional insight into consumer expectations. They describe reported experiences and preferences; they do not measure the performance of a specific product.
Reading the findings
Question wording, response labels, and displayed percentages are drawn from Customer Experience on Their Terms, report MIS16061EN. Short interpretations are editorial summaries.
Rounding may prevent single-response totals from equaling 100%. Multiple-response questions can exceed 100%. “<1%” is retained as reported; its bar uses 1% as an upper-bound marker.
The scenarios are illustrations of the research principles. Example amounts, conversations, and workflow outcomes are fictional and are not survey results or a demonstration of configured product capabilities. Choices in this experience are processed locally and are not submitted.
Complete research index · all 26 questions and response tables
A text-first reference to all findings. These values match the interactive explorer.
Q1 · The Resolution Gap
Think about the most recent time you contacted a company for customer service. Was your issue ultimately resolved?
Single response · n=577 · source p. 2
Responses
Response
Percent
Yes, during the first interaction
47%
Yes, but it took more than one interaction
31%
Partially resolved
9%
No
10%
I do not remember
3%
47% say their most recent issue was resolved in the first interaction; another 31% needed more than one interaction.
Q2 · The Friction Stack
Which of the following, if any, made that experience more difficult? Select up to three.
Select up to three · n=577 · source p. 4
Responses
Response
Percent
None of these
8%
Long wait time
51%
Having to repeat information
47%
Being transferred between people or departments
49%
Switching between channels or devices
11%
The representative did not have the information needed to help me
14%
An automated system or AI could not resolve my issue
25%
I could not easily reach a person
22%
Something the company said it would do did not happen afterward
8%
The website or app made the task difficult
7%
Long waits (51%), transfers (49%), and repeated information (47%) are the most selected sources of friction.
Q3 · The Resolution Priority
What would have made the biggest improvement to that experience? Select one.
Single response · n=577 · source p. 6
Responses
Response
Percent
Getting help more quickly or waiting less time
37%
Getting my issue fully resolved without needing another interaction
23%
Reaching the right person when I needed help
17%
Not having to repeat information I had already provided
7%
Being able to complete the task myself more easily
5%
Clear follow-through until my issue was resolved
4%
None of these
4%
Other
4%
Faster help leads at 37%, followed by complete resolution without another interaction at 23% and reaching the right person at 17%.
Q4 · The Knowledgeable Human Advantage
Thinking about customer service more generally, which of the following would most improve your experiences with companies today? Select up to two.
Select up to two · n=558 · source p. 8
Responses
Response
Percent
None of these
1%
Easy access to a knowledgeable person when needed
71%
Automated systems that can fully resolve simple requests
32%
Carrying my information across channels so I do not have to start over
29%
Companies that identify and address service problems before I contact them
13%
Clear follow-through until my issue is resolved
26%
Easier websites and apps for completing tasks myself
13%
71% select easy access to a knowledgeable person as an improvement they want; 32% also select automation that fully resolves simple requests.
Q5 · The Continuity Expectation
If you moved a customer-service conversation from one channel or device to another, for example from chat on a laptop to a phone call, which would you prefer?
Single response · n=558 · source p. 10
Responses
Response
Percent
Automatically continue with the previous conversation and context once the company verifies it is me
67%
Ask me each time before carrying the previous conversation and context into the new interaction
13%
Carry over the previous conversation and context only when I specifically request it
11%
Start a new interaction without carrying over the previous conversation or context
3%
No preference
6%
67% prefer automatic continuation after identity verification, while 24% prefer continuity with a prompt or an explicit request.
Q6 · The Anywhere Journey
If a company could continue a customer-service conversation on any of the following, which would you be willing to use? Select all that apply.
Select all that apply · n=558 · source p. 12
Responses
Response
Percent
None of these
1%
Computer or laptop web browser
59%
Company Smartphone app
42%
Mobile web browser
40%
Text message (SMS)
44%
Phone call
48%
Messaging app I already use, such as WhatsApp Messenger, or iMessage
18%
Smart speaker or voice assistant
6%
Smart TV or streaming device
7%
Vehicle infotainment system
5%
Smartwatch or other wearable device
4%
Consumers are willing to continue service across familiar channels, led by a computer browser (59%), phone calls (48%), and SMS (44%).
Q7 · The Proactive Service Mandate
In which of the following situations would you most want a company to proactively contact you before you had to contact them? Select up to three.
Select up to three · n=558 · source p. 14
Responses
Response
Percent
None of these; I generally prefer companies to wait until I contact them
4%
A possible fraudulent or unauthorized transaction
78%
A service outage affecting me
42%
A flight, delivery, or appointment disruption
41%
A billing error or unexpected charge
39%
A known problem or recall affecting a product or service I use
24%
A subscription or contract is about to renew
16%
A lower-cost plan or option is available to me
20%
The company believes I may be interested in purchasing something
3%
78% want proactive contact about possible fraud, compared with 3% who select outreach about something the company thinks they may want to buy.
Q8 · The Permission Boundary
Which of the following would make proactive customer service from a company feel less acceptable to you? Select all that apply.
Select all that apply · n=558 · source p. 16
Responses
Response
Percent
None of these would make proactive service less acceptable to me
3%
Contacting me too often
64%
Using information about me that I did not expect the company to use
58%
Using my precise location
41%
Using information from another company or service
42%
Analyzing my interactions when I did not expect it
36%
Taking action for me without my permission
52%
Contacting me through a device or channel I did not expect
29%
Trying to sell me something when I am seeking service or support
48%
Other / not sure
3%
Frequent contact (64%), unexpected use of information (58%), and action without permission (52%) can make proactive service less acceptable.
Q9 · The Permissioned Automation Model
Suppose a company's AI assistant detects that you were charged twice for the same purchase and could correct the duplicate charge. Which approach would you prefer?
Single response · n=558 · source p. 18
Responses
Response
Percent
Tell me what it found and let me correct the issue myself
25%
Prepare the correction and ask me to approve it
37%
Correct it automatically if this falls within rules I previously set
16%
Correct it automatically and notify me afterward
13%
I would not want AI to take this action for me
9%
The largest group, 37%, wants AI to prepare the correction and ask for approval; 16% prefers action within previously set rules.
Q10 · The Memory Contract
If a company's AI could retain information between customer-service interactions, which approach would you prefer?
Single response · n=558 · source p. 20
Responses
Response
Percent
Remember relevant information between interactions automatically
40%
Ask my permission before remembering information
22%
Let me choose which types of information it may remember
17%
Remember information only while my issue is still unresolved
11%
Do not retain information between interactions
6%
Not sure
4%
40% prefer relevant information to be remembered automatically, while others favor permission, information-level choices, or memory limited to an unresolved issue.
Q11 · The Sensitivity Gradient
For each type of information below, how comfortable would you be allowing a company's AI to retain it for future customer-service interactions?
Single response per information type · n=558 · source p. 22
Products or services I use
Response
Percent
Comfortable with it being retained automatically
43%
Comfortable only with explicit permission
34%
Not comfortable with it being retained
18%
Not sure
4%
Purchase or order history
Response
Percent
Comfortable with it being retained automatically
49%
Comfortable only with explicit permission
34%
Not comfortable with it being retained
15%
Not sure
3%
Previous customer-service problems
Response
Percent
Comfortable with it being retained automatically
49%
Comfortable only with explicit permission
32%
Not comfortable with it being retained
15%
Not sure
4%
Communication preferences
Response
Percent
Comfortable with it being retained automatically
56%
Comfortable only with explicit permission
29%
Not comfortable with it being retained
13%
Not sure
3%
Preferred store, branch, or service locations
Response
Percent
Comfortable with it being retained automatically
55%
Comfortable only with explicit permission
27%
Not comfortable with it being retained
14%
Not sure
3%
Personal schedule or appointment information
Response
Percent
Comfortable with it being retained automatically
34%
Comfortable only with explicit permission
34%
Not comfortable with it being retained
28%
Not sure
3%
Payment or financial information
Response
Percent
Comfortable with it being retained automatically
28%
Comfortable only with explicit permission
34%
Not comfortable with it being retained
34%
Not sure
4%
Health-related information
Response
Percent
Comfortable with it being retained automatically
28%
Comfortable only with explicit permission
30%
Not comfortable with it being retained
38%
Not sure
3%
Current or recent location information
Response
Percent
Comfortable with it being retained automatically
32%
Comfortable only with explicit permission
33%
Not comfortable with it being retained
32%
Not sure
3%
Family or household information
Response
Percent
Comfortable with it being retained automatically
27%
Comfortable only with explicit permission
29%
Not comfortable with it being retained
40%
Not sure
3%
Information from my conversations with the company
Response
Percent
Comfortable with it being retained automatically
40%
Comfortable only with explicit permission
40%
Not comfortable with it being retained
17%
Not sure
3%
56% are comfortable with communication preferences being retained automatically, compared with 28% for payment or financial information and 28% for health-related information.
Q12 · The Delegated Action Threshold
Which of the following actions, if any, would you personally be comfortable authorizing an AI assistant to complete without asking you for another approval at the time of the action? Select all that apply.
Select all that apply · n=558 · source p. 24
Responses
Response
Percent
None of these
18%
Update my profile or contact information
40%
Reschedule an appointment
33%
Make changes to an existing order
29%
Process a return or replacement
35%
Initiate a refund when I qualify
41%
Cancel a subscription or service
24%
Correct a billing error
34%
Change or upgrade a service plan
12%
Make a purchase under a spending limit I set
8%
Choose among options based on preferences I set
14%
Refunds (41%) and profile updates (40%) lead the actions consumers would be comfortable pre-authorizing; purchases under a set spending limit receive 8%.
Q13 · The Control Spectrum
If an AI assistant acts on your behalf, which control model would you prefer?
Single response · n=558 · source p. 26
Responses
Response
Percent
Approve every individual action
33%
Pre-approve certain types of actions, regardless of the specific situation
15%
Set rules or limits and allow the AI to act automatically within them
20%
Allow the AI to handle most actions independently, but ask me when an action falls outside normal circumstances
9%
Allow the AI to decide independently
<1%
Do not allow AI to act for me
22%
33% prefer approving each action, 20% prefer rules or limits, and fewer than 1% prefer letting AI decide independently.
Q14 · The Approval Threshold
Which of the following actions would you most want to approve each time, immediately before an AI assistant completes them?
Select up to three · n=558 · source p. 28
Responses
Response
Percent
None of these would require immediate approval if the AI stayed within rules I had previously approved
13%
Make a purchase
59%
Agree to a recurring charge or price increase
49%
Change payment or financial information
44%
Open or close an account or service
19%
Cancel a subscription or service
15%
Accept a contract, agreement, or new terms
20%
Share sensitive personal information with another company
17%
Make or change an important healthcare-related decision
15%
Financial actions lead the approval priorities: purchases at 59%, recurring charges or price increases at 49%, and payment changes at 44%.
Q15 · The Reversibility Principle
How would the ability to reverse or undo an AI action affect what you would allow it to do?
Select the answer that best describes your approach · n=558 · source p. 30
Responses
Response
Percent
I would allow AI to act on its own only when the action could easily be reversed
37%
I would allow some actions that cannot easily be reversed, but only after I approve them
18%
Reversibility would matter for important or high-impact actions, but not for routine actions
12%
Whether an action could be reversed would not significantly affect what I allow
5%
I would not allow AI to take actions on my behalf
28%
The largest group, 37%, would allow independent AI action only when it could easily be reversed.
Q16 · The AI Role Ladder
What is the most advanced role you would currently be comfortable allowing AI to play when you deal with companies?
Single response · n=558 · source p. 32
Responses
Response
Percent
Answer: Provide information when I ask
39%
Assist: Help me complete tasks
18%
Anticipate: Identify what I may need and recommend next steps
13%
Act: Take actions I have approved in advance
6%
Represent: Interact with companies on my behalf within limits I set
2%
None: I would prefer not to use AI when dealing with companies
22%
Answers and assistance are the most common comfort limits; only 2% select representation as the most advanced AI role they would currently allow.
Q17 · The Transparency Dial
After an AI takes an action on your behalf, how much information would you generally want the company to provide about what happened?
Single response · n=558 · source p. 34
Responses
Response
Percent
Only notify me if something went wrong
23%
Confirmation that the action was completed and the final result
28%
A brief summary of what the AI did
11%
A summary of what the AI did and why
10%
A detailed record of the actions and decisions
20%
Not sure
8%
The leading preference is completion confirmation at 28%, but customers vary widely in the amount of detail they want after AI acts.
Q18 · The Disclosure Expectation
At a minimum, when should a company tell you that AI is being used in customer service?
Single response · n=558 · source p. 36
Responses
Response
Percent
Whenever AI is involved in serving me
64%
When I am directly interacting with AI
16%
Only when AI makes a decision or takes an action that affects me
12%
I do not need to be told as long as the service works
3%
Not sure
4%
64% want disclosure whenever AI is involved in serving them, extending the expectation beyond direct chatbot interactions.
Q19 · The Safeguard Stack
If an AI can take actions on your behalf, which safeguards would most increase your confidence?
Select up to three · n=558 · source p. 38
Responses
Response
Percent
None of these
9%
I control which actions it can take
59%
I can set financial or other limits
38%
Important actions require my approval
52%
I can easily undo an action
30%
I can see what it did and why
16%
A person can step in when needed
16%
The company verifies my identity before allowing the AI to act
12%
I can turn autonomous actions off
10%
The company takes responsibility and fixes the problem if something goes wrong
14%
Control over permitted actions leads at 59%, followed by approval of important actions at 52% and financial or other limits at 38%.
Q20 · The Recovery Imperative
Imagine a company's AI made an incorrect change to your account, but the mistake could be fully corrected and caused no financial harm. Which response from the company would do the most to restore your confidence?
Single response · n=558 · source p. 40
Responses
Response
Percent
Correct the mistake quickly
37%
Clearly explain what went wrong
13%
Connect me with a knowledgeable person
13%
Let me review and approve the correction
6%
Apologize and explain what the company is doing to prevent it from happening again
12%
A mistake like this would permanently reduce my trust
14%
Not sure
5%
Quick correction is the leading confidence-restoring response at 37%, while 14% say even this fully correctable mistake would permanently reduce trust.
Q21 · The Escalation Imperative
If you had tried several times to resolve an issue with an AI assistant and the issue was still not resolved, what would you most want to happen next?
Select one · n=558 · source p. 42
Responses
Response
Percent
Not Sure
5%
The AI should automatically connect me with a person and carry over the conversation and information I already provided
64%
The AI should ask whether I want to be connected with a person
18%
The AI should try a different approach before offering a person
3%
The AI should offer other self-service options
2%
I should be shown the available options and choose what happens next
8%
After repeated AI failure, 64% want automatic connection to a person with context carried over, and another 18% want to be offered a human connection.
Q22 · The Context Handoff
If an AI assistant connects you to a customer-service representative, which information would be most important for the representative to know already?
Select up to two · n=558 · source p. 44
Responses
Response
Percent
Not Sure
5%
The reason I contacted the company
51%
Information I already provided
60%
What the AI already tried
20%
The outcome or resolution I am trying to achieve
22%
Relevant account or service information
10%
I would rather explain the situation myself
8%
The handoff priorities are information already provided at 60% and the reason for contacting the company at 51%.
Q23 · The Agent-to-Agent Frontier
Imagine your personal AI assistant could contact a company's AI on your behalf, using only account information you had authorized and taking only actions you had approved in advance. How comfortable or uncomfortable would you be with that?
Single response · n=558 · source p. 46
Responses
Response
Percent
Not Sure
6%
Very uncomfortable
25%
Somewhat uncomfortable
17%
Neither comfortable nor uncomfortable
12%
Somewhat comfortable
19%
Very comfortable
21%
Under the question's authorized, pre-approved conditions, 40% are comfortable with personal AI representation and 42% are uncomfortable.
Q24 · The Delegation Sweet Spot
Which of the following would you be most likely to ask a personal AI assistant to handle with a company for you?
Select up to two · n=558 · source p. 48
Responses
Response
Percent
I would not use a personal AI assistant for company interactions
18%
Wait on hold or navigate customer-service systems
39%
Check the status of an order or request
48%
Schedule or reschedule an appointment
18%
Cancel or change a service
9%
Resolve a billing or account issue
16%
Compare prices, plans, or offers
20%
Status checks lead potential delegation at 48%, followed by waiting on hold or navigating service systems at 39%.
Q25 · The AI Preference Divide
Imagine two companies offer similar products, prices, and overall customer-service quality. One uses AI extensively to automate customer service, while the other relies more heavily on traditional self-service tools and human representatives. Which would you prefer?
Single response · n=558 · source p. 50
Responses
Response
Percent
Not Sure
5%
Strongly prefer the company that uses AI extensively
26%
Somewhat prefer the company that uses AI extensively
15%
No preference
10%
Somewhat prefer the company that uses less AI
15%
Strongly prefer the company that uses less AI
29%
Preferences are divided: 26% strongly prefer extensive AI use and 29% strongly prefer less AI when products, prices, and overall service quality are similar.
Q26 · The Customer Service Blueprint
Thinking about everything covered in this survey, which of the following would most improve customer service for you?
Select up to two · n=558 · source p. 52
Responses
Response
Percent
None of these
5%
Easy access to a knowledgeable person when needed
72%
AI that resolves routine requests quickly
28%
Carrying my information across channels so I do not have to start over
33%
Companies that identify and address service problems before I contact them
15%
AI that can take approved actions for me
7%
A personal AI assistant that deals with companies for me
5%
Easier websites and apps for completing tasks myself
15%
72% prioritize knowledgeable human help, 33% prioritize information continuity, and 28% prioritize AI that resolves routine requests quickly.