Chatbots in AI Search: The Biggest AI Visibility Problems and How We'd Fix Them

Companies covered: Ada, Drift, Tidio, Landbot, Chatfuel, Intercom, ManyChat, Freshdesk, Zendesk Chat, LiveChat (Text S.A. (formerly LiveChat Software S.A.)

Answer Capsule

The November 2026 Chatbots benchmark compares 10 companies across 253 AI answers that met its inclusion rules. Intercom recorded 201 mentions and 92 recommendations (36.36% of the answers). Being mentioned means appearing in an answer; being recommended means receiving an affirmative endorsement.

We examine each company, compare the available monthly results, and set out a five-step plan for improving the information AI systems can draw on. The work starts with measuring visibility and checking conflicting claims, then moves to company pages, relevant publishers and buyer questions. The cited pages need review before any correction request, and the research does not establish what caused recommendation differences or promise future gains.

Executive industry diagnosis

Questions This Section Answers

• What does the Chatbots benchmark show about mentions, recommendations and placement?

The peer set splits into a clear leading tier and a long tail. Intercom, Tidio and Zendesk Chat are named in most answers and earn affirmative endorsements far more often than the rest, so they set the reference point buyers hear. Freshdesk and LiveChat are named frequently but receive endorsements at a much lower rate.

The gap between being named and being endorsed is the central business issue. Several companies are discussed in neutral terms without earning an endorsement, which suggests the answers treat them as options rather than preferred choices. That distinction matters more than raw visibility for anyone trying to win consideration. A practical first move is a cohort scorecard that records mentions, endorsements and placements separately for each company, so the team can see where reach exists without endorsement.

Fact checks and page reviews can then target the companies where the gap is widest, with current primary sources confirming any change before edits are proposed.

What changed since last month

Questions This Section Answers

• What changed across the usable monthly Chatbots snapshots?

Only months with verified results are shown. Monthly sample coverage is listed in Methodology.

The monthly figures come from separate answer samples, so the movement between them should be read as recorded change rather than like-for-like performance. Most companies show lower counts in the latest sample, and the recommendation rates move independently of mention totals. The comparability limit sits alongside the numbers, not inside them.

What the history does support is a usable snapshot series once the sample basis is documented. The agency would keep that record explicit, note which months are comparable, and avoid treating a single month's shift as a trend. Any interpretation of direction should wait for consistent sampling.

Company-by-company findings

Ada

Questions This Section Answers

• What measured visibility pattern and review priority does Ada have in Chatbots?

Ada recorded 48 mentions (18.97%) and 21 recommendations (8.30%) across 253 answers included in the benchmark. Top 3 placements numbered 12; first-place recommendations numbered 4. Framing among its mentions was 21 positive, 27 neutral and 0 negative.

From October 2026 to November 2026, Ada's mention count fell from 53 to 48, while its recommendation count fell from 40 to 21. The recommendation rate was 9.17% and 8.30% respectively.

AI answers contained 1 conflicting claim about Ada: 0 critical, 0 high, 1 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://ada.cx/.

Selected third-party review candidate: https://lindy.ai/blog/best-customer-service-chatbots.

Ada sits in the middle of the field: present in many answers, recommended in a smaller share, with few first-place positions. The recorded pricing conflict is the practical starting point, since two different pricing models appear in answers. A fact-check against current primary sources would establish which description is accurate before any correction is considered.

Proposed work is a conditional review of the owned page and the selected third-party article, focused on how pricing is described. The brief would ask for verified figures and clear plan structure, with no edit made until the source is confirmed.

Drift

Questions This Section Answers

• What measured visibility pattern and review priority does Drift have in Chatbots?

Drift recorded 33 mentions (13.04%) and 6 recommendations (2.37%) across 253 answers included in the benchmark. Top 3 placements numbered 4; first-place recommendations numbered 0. Framing among its mentions was 6 positive, 27 neutral and 0 negative.

From October 2026 to November 2026, Drift's mention count fell from 68 to 33, while its recommendation count fell from 44 to 6. The recommendation rate was 10.09% and 2.37% respectively.

AI answers contained 4 conflicting claims about Drift: 0 critical, 4 high, 0 medium, 0 low. Of these, 4 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Another 1 possible conflicts were excluded because the company match was uncertain. They need an identity check before substantive review.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected third-party review candidate: https://intercom.com/learning-center/intercom-alternatives.

Drift's pattern is reach without endorsement. It is mentioned reasonably often but rarely recommended and never placed first in this sample, and most mentions are neutral. The high-priority conflict about whether the company is still active is the most consequential item, because that question shapes how buyers read every other answer.

Proposed work is a fact-check of the business-status claim against current primary sources, followed by a page brief for the selected third-party article if the evidence supports a change. The uncertain company match stays outside the correction queue until identity is confirmed.

Tidio

Questions This Section Answers

• What measured visibility pattern and review priority does Tidio have in Chatbots?

Tidio recorded 176 mentions (69.57%) and 91 recommendations (35.97%) across 253 answers included in the benchmark. Top 3 placements numbered 58; first-place recommendations numbered 21. Framing among its mentions was 95 positive, 81 neutral and 0 negative.

From October 2026 to November 2026, Tidio's mention count fell from 321 to 176, while its recommendation count fell from 267 to 91. The recommendation rate was 61.24% and 35.97% respectively.

AI answers contained 2 conflicting claims about Tidio: 0 critical, 0 high, 2 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://tidio.com/.

Selected third-party review candidate: https://kayako.com/blog/best-live-chat-software/.

Tidio is one of the strongest performers in the set, with high visibility and a large share of recommendations, though first-place positions are fewer. The recorded pricing conflicts are moderate in severity and concern plan figures, which are the kind of detail buyers check directly. Resolving them protects an otherwise strong position.

Proposed work is a pricing fact-check against current primary sources, then a page brief covering the owned site and the selected third-party comparison. Any edit would wait on confirmed figures and dates.

Landbot

Questions This Section Answers

• What measured visibility pattern and review priority does Landbot have in Chatbots?

Landbot recorded 16 mentions (6.32%) and 6 recommendations (2.37%) across 253 answers included in the benchmark. Top 3 placements numbered 3; first-place recommendations numbered 2. Framing among its mentions was 6 positive, 10 neutral and 0 negative.

From October 2026 to November 2026, Landbot's mention count fell from 25 to 16, while its recommendation count fell from 18 to 6. The recommendation rate was 4.13% and 2.37% respectively.

AI answers contained 0 conflicting claims about Landbot: 0 critical, 0 high, 0 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

No results are available for this part of the analysis.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://landbot.io/.

Selected third-party review candidate: https://kayako.com/tools/manychat-alternatives/.

Landbot has modest visibility and a small recommendation share, with few first-place positions. No conflicting claims were recorded, so the immediate opportunity is not correction but clarity: making sure the answers that do mention Landbot describe it accurately and in terms buyers use. The observed questions about free access and cheaper alternatives point to where explanation may help.

Proposed work is a page brief for the owned site and the selected third-party article, focused on plan and pricing clarity, with facts verified before any change is proposed.

Chatfuel

Questions This Section Answers

• What measured visibility pattern and review priority does Chatfuel have in Chatbots?

Chatfuel recorded 29 mentions (11.46%) and 10 recommendations (3.95%) across 253 answers included in the benchmark. Top 3 placements numbered 4; first-place recommendations numbered 3. Framing among its mentions was 12 positive, 17 neutral and 0 negative.

From October 2026 to November 2026, Chatfuel's mention count rose from 21 to 29, while its recommendation count fell from 17 to 10. The recommendation rate was 3.90% and 3.95% respectively.

AI answers contained 9 conflicting claims about Chatfuel: 0 critical, 0 high, 9 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://chatfuel.com/.

Selected third-party review candidate: https://setsmart.io/blog/manychat-alternative.

Chatfuel is mentioned more often than it is recommended, and the recorded conflicts cluster around starting prices. Those are moderate-severity items, but pricing is exactly what buyers compare, so inconsistent figures weaken an otherwise reasonable position. The observed questions about cheaper alternatives suggest buyers are weighing cost directly.

Proposed work is a pricing fact-check against current primary sources, followed by a page brief for the owned site and the selected third-party article. Edits would be proposed only where verification supports them.

Intercom

Questions This Section Answers

• What measured visibility pattern and review priority does Intercom have in Chatbots?

Intercom recorded 201 mentions (79.45%) and 92 recommendations (36.36%) across 253 answers included in the benchmark. Top 3 placements numbered 67; first-place recommendations numbered 29. Framing among its mentions was 96 positive, 105 neutral and 0 negative.

From October 2026 to November 2026, Intercom's mention count fell from 374 to 201, while its recommendation count fell from 294 to 92. The recommendation rate was 67.43% and 36.36% respectively.

AI answers contained 9 conflicting claims about Intercom: 0 critical, 2 high, 7 medium, 0 low. Of these, 2 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://intercom.com/learning-center/intercom-alternatives.

Selected third-party review candidate: https://getmacha.com/blog/ai-chatbot-platforms.

Intercom leads the peer set on visibility, recommendations and first-place positions, so its main risk is accuracy rather than reach. The recorded conflicts include an ownership claim and a starting price, both of which could mislead buyers if left unresolved. The ownership item in particular deserves careful verification because it changes how the company is categorized.

Proposed work is a fact-check of the ownership and pricing claims against current primary sources, then a page brief for the selected owned and third-party pages. No correction would be proposed until the entity and date are confirmed.

ManyChat

Questions This Section Answers

• What measured visibility pattern and review priority does ManyChat have in Chatbots?

ManyChat recorded 73 mentions (28.85%) and 21 recommendations (8.30%) across 253 answers included in the benchmark. Top 3 placements numbered 15; first-place recommendations numbered 10. Framing among its mentions was 24 positive, 43 neutral and 6 negative.

From October 2026 to November 2026, ManyChat's mention count fell from 99 to 73, while its recommendation count fell from 80 to 21. The recommendation rate was 18.35% and 8.30% respectively.

AI answers contained 4 conflicting claims about ManyChat: 0 critical, 1 high, 3 medium, 0 low. Of these, 1 was rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://manychat.com/.

Selected third-party review candidate: https://chatfuel.com/features/pricing.

ManyChat has solid visibility but a much smaller recommendation share, and it is the only company here with recorded negative framing. The conflicts concern entry pricing and free-plan limits, which are practical details buyers rely on. Neutral and negative wording alongside pricing uncertainty suggests the answers are cautious rather than dismissive.

Proposed work is a fact-check of the pricing and plan-limit claims, then a page brief for the owned site and the selected third-party article. The negative framing would be reviewed for accuracy, not treated as a reputation problem to argue with.

Freshdesk

Questions This Section Answers

• What measured visibility pattern and review priority does Freshdesk have in Chatbots?

Freshdesk recorded 76 mentions (30.04%) and 27 recommendations (10.67%) across 253 answers included in the benchmark. Top 3 placements numbered 13; first-place recommendations numbered 1. Framing among its mentions was 28 positive, 48 neutral and 0 negative.

From October 2026 to November 2026, Freshdesk's mention count fell from 129 to 76, while its recommendation count fell from 87 to 27. The recommendation rate was 19.95% and 10.67% respectively.

AI answers contained 3 conflicting claims about Freshdesk: 0 critical, 0 high, 3 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://freshworks.com/customer-service/software/.

Selected third-party review candidate: https://zapier.com/blog/zendesk-alternatives/.

Freshdesk is mentioned often but recommended far less, and first-place positions are rare. Most mentions are neutral, which fits a pattern of being listed as an option without strong endorsement. The recorded pricing conflicts concern per-agent figures, a detail that matters in side-by-side comparisons.

Proposed work is a pricing fact-check against current primary sources, followed by a page brief for the selected owned and third-party pages. The aim is accurate, comparable plan information rather than a push for stronger wording.

Zendesk Chat

Questions This Section Answers

• What measured visibility pattern and review priority does Zendesk Chat have in Chatbots?

Zendesk Chat recorded 152 mentions (60.08%) and 70 recommendations (27.67%) across 253 answers included in the benchmark. Top 3 placements numbered 46; first-place recommendations numbered 19. Framing among its mentions was 74 positive, 78 neutral and 0 negative.

From October 2026 to November 2026, Zendesk Chat's mention count fell from 249 to 152, while its recommendation count fell from 192 to 70. The recommendation rate was 44.04% and 27.67% respectively.

AI answers contained 1 conflicting claim about Zendesk Chat: 0 critical, 0 high, 1 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://zendesk.com/service/messaging/live-chat-software/.

Selected third-party review candidate: https://zapier.com/blog/best-customer-support-chat-apps/.

Zendesk Chat is a strong performer, with high visibility and a substantial recommendation share, though first-place positions are fewer than its overall reach would suggest. The single recorded conflict concerns starting price, a moderate item that is easy for buyers to check. Resolving it protects an otherwise consistent position.

Proposed work is a pricing fact-check against current primary sources, then a page brief for the selected owned and third-party pages, with any edit contingent on confirmed figures.

LiveChat (Text S.A. (formerly LiveChat Software S.A.)

Questions This Section Answers

• What measured visibility pattern and review priority does LiveChat (Text S.A. (formerly LiveChat Software S.A.) have in Chatbots?

LiveChat (Text S.A. (formerly LiveChat Software S.A.) recorded 105 mentions (41.50%) and 26 recommendations (10.28%) across 253 answers included in the benchmark. Top 3 placements numbered 18; first-place recommendations numbered 11. Framing among its mentions was 30 positive, 75 neutral and 0 negative.

From October 2026 to November 2026, LiveChat (Text S.A. (formerly LiveChat Software S.A.)'s mention count fell from 154 to 105, while its recommendation count fell from 119 to 26. The recommendation rate was 27.29% and 10.28% respectively.

AI answers contained 0 conflicting claims about LiveChat (Text S.A. (formerly LiveChat Software S.A.): 0 critical, 0 high, 0 medium, 0 low. Of these, 0 were rated critical or high priority; the critical claims are included in that total. These ratings help prioritize fact-checking. They do not establish that a company or publisher made an error.

Another 8 possible conflicts were excluded because the company match was uncertain. They need an identity check before substantive review.

No results are available for this part of the analysis.

Both columns quote AI answers, not verified company facts. Resolve the entity, date and source before deciding whether a correction is warranted; different dates can explain an apparent conflict.

These are recorded answer observations, not search volume, verified demand, or proof of missing website content.

Selected owned-page review candidate: https://livechat.com/.

Selected third-party review candidate: https://zendesk.com/service/messaging/live-chat-software/.

LiveChat has good visibility but a much smaller recommendation share, and most mentions are neutral. No confirmed conflicts were recorded, though several possible ones were excluded because the company match was uncertain. Those need an identity check before any substantive review, since the parent-company naming adds ambiguity.

Proposed work is an identity check on the excluded items, then a page brief for the owned site and the selected third-party article. Only confirmed matches would move into the review queue.

The CiteWorks five-step improvement plan

Step 1. Establish the baseline

Questions This Section Answers

• What should the agency deliver when it works to establish the baseline for Chatbots?

Build a cohort scorecard that records mentions, recommendations, placements and framing separately for each company, with the sample basis noted for every month. Each record gets an owner, an evidence requirement and a review status, so future comparisons rest on documented ground rather than memory. This is proposed work, not a measured result.

Step 2. Verify disputed facts

Questions This Section Answers

• What should the agency deliver when it works to verify disputed facts for Chatbots?

AI answers contained 33 conflicting claims across 8 companies. Another 9 possible conflicts were excluded because their company matches were uncertain. The conflicting claims need fact-checking before any corrections are proposed.

Work through the recorded conflicts company by company, starting with the high-priority items and the pricing claims. Each check resolves the entity, date and source before deciding whether a correction is warranted. The excluded uncertain matches stay outside the queue until identity is confirmed. Confirmed edits, no-change decisions and open questions are logged separately.

Step 3. Review owned pages

Questions This Section Answers

• What should the agency deliver when it works to review owned pages for Chatbots?

Assign page briefs to the supplied owned URLs, asking for verified facts, clear plan and pricing descriptions, and language that matches how buyers frame their questions. No edit is proposed until current primary sources confirm the detail. Each brief records what was checked, what changed and what was deliberately left alone.

Step 4. Review third-party sources

Questions This Section Answers

• What should the agency deliver when it works to review third-party sources for Chatbots?

Prepare source-review packages for the selected third-party pages, verifying current contents before any request is made. A citation does not establish that a page influenced an answer, so requests should rest on accuracy rather than assumed leverage. Confirmed errors, suggested clarifications and publisher responses are tracked separately.

Step 5. Address buyer questions and measure again

Questions This Section Answers

• What should the agency deliver when it works to address buyer questions and measure again for Chatbots?

Turn the observed buyer questions into content briefs that answer them directly, using verified facts and plain comparisons. Delivered work is tracked separately from answer outcomes, so completed briefs and pages are not confused with measured visibility gains. A fresh scorecard at the end records what changed and where comparability limits still apply.

Third-party authority and citation opportunities

Questions This Section Answers

• Which cited publisher pages support targeted source review in Chatbots?

These pages are candidates for review, not confirmed errors. Citation appearances are counted across the captured answers and are not added again for each company.

The publisher list should be prioritized by relevance and by how many of the selected companies each page covers, not by citation counts alone. Pages spanning several companies offer broader review value, while single-company pages allow deeper accuracy checks. Citation appearances show where answers point, not that a page shaped the wording.

Before any request, the team would verify each page's current contents and confirm which claims are actually inaccurate. Requests would be framed around specific, evidenced corrections rather than general prominence. Responses and no-change decisions are logged so the record stays honest about what was asked and what was granted.

What we'd prioritize first

Start with the high-priority conflicts, since the business-status and ownership questions affect how buyers categorize companies at all. Pricing conflicts come next because they are common, checkable and directly comparable. The uncertain company matches wait for identity confirmation rather than entering the queue early.

The first month produces the baseline scorecard, a scoped fact-review register and a page and source shortlist, each with an owner and evidence requirement. The second month turns approved briefs into justified requests, separating confirmed edits from no-change decisions. The third month delivers buyer-question content and a fresh scorecard, with completed work recorded apart from answer outcomes.

Methodology and conclusion

Questions This Section Answers

• What do the Chatbots sample and source-review limits allow us to conclude?

This report covers 2026-11. The exact dates when the original AI answers were collected have not been verified. The visibility rates use 253 answers that met the benchmark's inclusion rules. A company is recommended only when the answer both endorses it and uses positive wording. Top 3 and first-place counts are part of those recommendations, not extra recommendations.

An answer can mention or recommend several companies, so company counts cannot be added to find the number of distinct answers. The net sentiment score is positive minus negative mentions, divided by all company mentions; it describes answer wording, not customer satisfaction.

The wider citation analysis covers 800 captured answers and 4946 appearances of cited URLs. Citation appearances across all captured answers differ from the number of benchmark answers citing a page and the number that also recommend a company. The current contents of cited pages and disputed company claims still need independent verification.

Monthly sample coverage:

July 2026 measured a different set of companies, so it is excluded from this comparison. August 2026 measured a different set of companies, so it is excluded from this comparison.

The answer samples differ between months. These comparisons show how the recorded results changed; they do not establish why they changed.

The results describe how often companies appeared and were endorsed across the included answers, not why. Because the monthly samples differ, the comparisons show recorded change and do not establish cause. Recommendation requires both endorsement and positive wording, and placements sit inside recommendations rather than adding to them.

The sensible next step is disciplined verification: confirm disputed facts against current primary sources, brief the pages that matter, and re-measure on a documented basis. No outcome is promised, and completed work is reported separately from any movement in the answers.