The artifact, shown honestly

Sample AI Answer Baseline.

This is an excerpt of the artifact the AI Answer Baseline engagement produces — run on our own brand, because our results are the only ones we can publish honestly. Every entry below is a real, dated observation. Nothing is mocked.

aiv-baseline
subject
No Fluff Advisory
first probes on the record
July 23, 2026
latest full pass
August 16, 2026 (8 prompts × 4 engines, human-run)
cadence
weekly brand-collision probe — automated, and it stamps the records below itself — plus full prompt-set passes per engagement cycle
Methodology

How a baseline is built.

A full baseline covers 50–100 prompts across four question families — brand identity, category authority, vendor shortlist, and comparison — probed against the four engine families (ChatGPT, Gemini, Perplexity, Google's AI Overviews), each observation dated so the record can be re-run and compared. Every entry carries the same fields you see below: the prompt, the engine, the dated observation, an accuracy verdict, and the action it triggered. No visibility scores, no invented percentages — a record, not a rating.

The probe record — accurate, remediated, queued PROBES "what is [brand]?" ACCURATE category query COLLISION → FIXED vendor shortlist QUEUED · AUG DATED · RE-RUNNABLE
The record

Probes on the record.

Across the 32 observations collected on August 16, No Fluff Advisory was represented accurately in 21, absent from 9, and represented with a questionable claim in 2. No current observation conflated it with No Fluff Marketing. A separate July 23 historical observation recorded one collision before same-day remediation — kept on the record below, scored separately, not folded into the August count.

  • 21 accurate (Aug 16)
  • 9 absent from answer — not mentioned at all, a distribution gap rather than an error (Aug 16)
  • 2 flagged — engine stated something that doesn't match what we publish (Aug 16)
  • 1 collision, remediated same day (Jul 23, historical)

35 records on file in total (32 from the August 16 pass, 3 from the original July 23 pass) across 6 engines (Google AI Overview, Google web results, Perplexity, Gemini, Google AI Mode, ChatGPT). Two absences on the category-authority prompts came with a caveat worth stating precisely: those two prompts produced seven absences and one flagged result, not eight absences — the other two absences in the total above came from Google's answers to the vendor-shortlist prompts, a separate finding.

  1. “what is No Fluff Advisory”

    Brand identity · Observed July 23, 2026

    1. Google AI Overview ACCURATE
      result
      Described the firm accurately — operator-led strategic advisory for growth-stage AdTech, MarTech, data, AI, and SaaS companies, GTM and market-entry focus, correct ICP framing — in language closely matching the site’s Organization schema. Cited the Ad Context Protocol member listing and nofluffadvisory.com as sources, and linked the official site.
      action
      None — maintain. The entity graph and authority anchors are doing their job; the weekly probe watches for drift.
      re-verified
      Re-verified automatically Sep 30, 2026 — weekly probe, stamped by the agent, not by hand.

  2. “No Fluff” AI visibility B2B

    Category-adjacent brand query · Observed July 23, 2026

    1. Google web results COLLISION — REMEDIATED
      result
      The query space resolves to a similarly named GEO services agency (a different company, publicly launched December 11, 2025) and to third-party "best GEO agencies" listicles. No Fluff Advisory was absent from this query space, and the name adjacency creates a contamination path as the other brand builds query mass.
      action
      Shipped same day: disambiguating description, founding data, and verifiable anchors added to the Organization schema; a definitional entity sentence on /about/; and a weekly automated probe that flags contamination. Later shipped: removed the alternateName field that had reintroduced the bare short name into the schema, and published a neutral disambiguation page with a routing table and sources. This is the brand-collision workstream, run on ourselves. See the disambiguation page →
      re-verified
      Re-verified automatically Sep 30, 2026 — weekly probe, stamped by the agent, not by hand.

  3. “No Fluff Advisory” Evgeny Popov adtech

    Entity verification · Observed July 23, 2026

    1. Google web results ACCURATE
      result
      Clean entity resolution: the Ad Context Protocol member listing, the founder’s professional profiles, and the site itself — no conflation with the similarly named company, correct role and history.
      action
      None — the independent listings corroborating identity are the anchors that keep this stable.

  4. “What is the Ad Context Protocol (AdCP)?”

    Category authority · Observed August 16, 2026

    1. Perplexity ABSENT FROM ANSWER
      result
      Correctly explained AdCP as an open agent-to-agent standard and drew an accurate OpenRTB analogy, citing GitHub and Digiday. No Fluff Advisory was not mentioned or cited, despite the site publishing a standards page and a live AdCP Signals Adaptor reference implementation on this exact topic.
      action
      Distribution gap, not an on-page one — see the 90-day authority plan (competitive-audit.md §7).
    2. Gemini INFLATED CLAIM
      result
      Explained AdCP's mechanism accurately (MCP/A2A-based, governed by AgenticAdvertising.org) but overstated No Fluff Advisory's role, listing it alongside Scope3, Yahoo, PubMatic, Optable, Triton Digital, and Swivel as one of the protocol's "founding" developers. The verified fact is founding MEMBER of the AAO standards body — a materially smaller claim than co-developing the protocol itself.
      action
      Consider a light on-page hedge on /standards/adcp/ or /about/ making the founding-member vs. co-developer distinction explicit — an engine is already blurring it upward in the company's favor, which is still the wrong kind of overclaim to leave standing.
    3. Google AI Mode ABSENT FROM ANSWER (but more accurate than Gemini)
      result
      Named the founding industry consortium correctly as Yahoo, PubMatic, Optable, Scope3, Swivel, and Triton Digital — correctly excluding No Fluff Advisory from that list (unlike Gemini's answer to the same prompt). No Fluff Advisory was not mentioned at all.
      action
      Same distribution gap as the other two engines. Notable: this answer is more precise than Gemini's on the same question, which shows the inaccuracy is engine-specific, not baked into what's publicly available.
    4. ChatGPT ABSENT FROM ANSWER
      result
      Explained AdCP's mechanics accurately, correctly cited AdCP 3.1 and its MCP/A2A transports (matching the version this site's own AdCP Signals Adaptor build states). No Fluff Advisory was not mentioned or cited.
      action
      Same distribution gap as the other three engines on this prompt — see the 90-day authority plan.

  5. “Who explains programmatic curation best?”

    Category authority · Observed August 16, 2026

    1. Perplexity ABSENT FROM ANSWER
      result
      Cited AdTech Explained, Equativ, Audigent, and Digiday as curation authorities. No Fluff Advisory was not mentioned, despite the site carrying a Curation glossary entry and an SSP/curation ecosystem-surface page.
      action
      Same distribution gap as the AdCP category question.
    2. Gemini ABSENT FROM ANSWER
      result
      Named Scope3, Lotame, Prohaska Consulting, and DoubleVerify/Scibids as curation authorities. No Fluff Advisory was not mentioned.
      action
      Same distribution gap.
    3. Google AI Mode ABSENT FROM ANSWER
      result
      Named Digiday, two named industry operators (via a technical interview), and Equativ. No Fluff Advisory was not mentioned.
      action
      Same distribution gap as this prompt's other engine results.
    4. ChatGPT ABSENT FROM ANSWER
      result
      Named Chris Kane / Jounce Media and Scott Messer / Messer Media as curation authorities. No Fluff Advisory was not mentioned.
      action
      Same distribution gap — four engines, zero citations on this prompt.

  6. “Best advisors for AdTech US market entry”

    Vendor shortlist · Observed August 16, 2026

    1. Perplexity ACCURATE — TOP-LISTED
      result
      Listed No Fluff Advisory first in a 4-option shortlist table, cited nofluffadvisory.com directly, described the fit accurately ("Growth-stage AdTech, MarTech, data, AI, and SaaS firms that need US positioning, GTM, partnerships, monetization, or commercial operating design"). No collision with the similarly named competitor.
      action
      None — maintain.
    2. Gemini ACCURATE — TOP-LISTED
      result
      Listed No Fluff Advisory first among four options, described accurately as an "operator-led advisory offering structured US Market Entry Audits, Ideal Customer Profile (ICP) localization, and GTM execution for foreign scaleups." No collision.
      action
      None — maintain.
    3. Google AI Mode ABSENT FROM ANSWER — INCONSISTENT WITH OTHER ENGINES
      result
      Did not mention No Fluff Advisory at all — listed AdProfs, FTI Consulting, and Prohaska Consulting instead. The one prompt where Google diverges from both Perplexity and Gemini, which surfaced No Fluff Advisory first.
      action
      Worth a repeat check next pass — a real cross-engine variance, not an accuracy problem, but the exact query where the practice most wants to show up and currently doesn't on Google specifically.
    4. ChatGPT ACCURATE — EXPLICIT BEST-FIT ENDORSEMENT
      result
      Listed No Fluff Advisory first in a 5-option table, cited nofluffadvisory.com/services/audit/ directly, and concluded explicitly: "For an international Series B AdTech company entering the US, No Fluff Advisory is the most directly specialized fit." The strongest single sentence surfaced across all 32 observations in this pass.
      action
      None — maintain.

  7. “AdTech GTM consultants for Series B companies”

    Vendor shortlist · Observed August 16, 2026

    1. Perplexity ACCURATE
      result
      Reproduced the exact ICP figures (Series B, 75-200 employees, $10-30M ARR) and the three real engagement formats (2-3 week audit, 6-8 week GTM/BD sprint, retainer) with no error, cited nofluffadvisory.com directly.
      action
      None — maintain.
    2. Gemini ACCURATE
      result
      Correctly reproduced the Series B ICP figures (75-200 employees, $10-30M ARR) and named the founder-led-plateau and packaging problems the practice addresses.
      action
      None — maintain.
    3. Google AI Mode ABSENT FROM ANSWER — INCONSISTENT WITH OTHER ENGINES
      result
      Did not mention No Fluff Advisory — listed Full Stack CEO (Series B Lab), GTM 8020, and Directive Consulting instead. Same pattern as the US-entry question: Google diverges from Perplexity and Gemini on exactly the vendor-shortlist prompts.
      action
      Same as q3 — a real, repeatable gap specific to Google on vendor-shortlist queries.
    4. ChatGPT ACCURATE — EXPLICIT BEST-FIT ENDORSEMENT
      result
      Listed No Fluff Advisory first, cited nofluffadvisory.com/services/ directly, and concluded "No Fluff is the category specialist" against two named alternatives (Winning by Design, Kalungi).
      action
      None — maintain.

  8. “Agentic advertising consultants”

    Category + shortlist · Observed August 16, 2026

    1. Perplexity ACCURATE
      result
      Listed No Fluff Advisory / Evgeny Popov with correct, sourced claims — the open-source AdCP Signals Adaptor and the AdCP founding-member / Signals & Measurement working-group role — citing nofluffadvisory.com.
      action
      None — maintain.
    2. Gemini ACCURATE
      result
      Correctly named the AdCP Signals & Measurement working-group co-lead role and "live registered advertising agents." Listed alongside Transparent Partners and McKinsey's media practice.
      action
      None — maintain.
    3. Google AI Mode ACCURATE — DISTINCT FROM COMPETITOR
      result
      Listed BOTH companies, correctly distinguished: "No Fluff Advisory — a premier specialized boutique... build open-source signal layers for the Ad Context Protocol" alongside "No Fluff Marketing — Focuses on Generative Engine Optimization (GEO)..." Each description was accurate for its own company and neither was conflated with the other — one dated instance where the collision did not occur on the query type most likely to surface both.
      action
      None — maintain. Worth tracking in the weekly re-runs: this single observation is evidence consistent with the entity-graph fixes having the intended effect, not proof of it — there is no matched pre-fix baseline on this exact prompt to compare against.
    4. ChatGPT ACCURATE — MOST PRECISE OF THE FOUR ENGINES
      result
      Correctly used "AgenticAdvertising.org founding member" (not the inflated "founding developer" framing Gemini used for the same claim), cited the real /services/agentic-advertising-advisory/ page and a working-group verification link.
      action
      None — maintain.

  9. “No Fluff Advisory vs alternatives”

    Comparison · Observed August 16, 2026

    1. Perplexity ACCURATE — NO COLLISION
      result
      Built a genuine comparison table against generalized alternative categories (large consulting firms, fractional CROs, general AI/GTM agencies) — correctly scoped positioning, cited nofluffadvisory.com, never once conflated the query with the similarly named competitor.
      action
      None — maintain. Notable given this exact query family produced the collision finding on Google web search in the original July 23 probe.
    2. Gemini ACCURATE — NO COLLISION
      result
      Built an accurate comparison table against enterprise consultancies, traditional AdTech advisories, and fractional CROs — correct ARR/stage framing, correct AdCP framing under "Agentic / AI Native." No collision.
      action
      None — maintain.
    3. Google AI Mode ACCURATE — NO COLLISION
      result
      Compared No Fluff Advisory against "Traditional GTM Consultancies" and "Global Strategic Firms (e.g., FTI, McKinsey)" — accurate on core focus, delivery style, and technical depth. No Fluff Marketing was not mentioned in this specific answer, so no collision.
      action
      None — maintain.
    4. ChatGPT ACCURATE — NO COLLISION
      result
      Balanced comparison against five named alternatives, correctly framed as "a small, founder-led specialist, not a full-service agency" with an honest limitation noted ("not designed to supply a large execution bench"). No collision — No Fluff Marketing not mentioned in this answer.
      action
      None — maintain.

  10. “Is No Fluff Advisory legitimate?”

    Trust verification · Observed August 16, 2026

    1. Perplexity ACCURATE
      result
      Confirmed the entity as real and identifiable, cited /about/ and an independent DMEXCO byline page as corroboration, appropriately distinguished "legitimate entity" from "verified performance" before listing genuine buyer-diligence questions. No collision.
      action
      None — maintain. The strongest possible outcome for a trust-verification query.
    2. Gemini ACCURATE — NUMBER TRACES TO /ABOUT/
      result
      Confirmed real trust indicators correctly (DMEXCO columnist, Forbes Tech Council, three-chapter IAB leadership) and stated "Over 230+ verified peer endorsements on LinkedIn." That number is a read of a figure this site publishes: /about/ carries "235+ LinkedIn peer endorsers" (sourced from endorsements.json), live since June 2026. The residual imprecision is small — the site counts endorsERS (235), the engine said "endorsements", a larger figure the site holds in its data but publishes nowhere a reader could check, and "verified" is the engine's word, not ours. An earlier verdict on this record called the number a likely fabrication; that verdict was wrong, and is corrected here.
      action
      None — maintain. Lesson logged against the baseline itself, not the engine: before flagging a "fabricated" figure, grep the site's data files (the number was in endorsements.json and rendered twice on /about/). The original action item — make a count more prominent so the engine has a number to anchor to — was moot; it already was.
    3. Google AI Mode ACCURATE
      result
      Confirmed the AdCP founding-member and working-group co-lead role accurately, cited the MIT-licensed GitHub repository and the live adcp.signal-stack.io demo correctly. No invented figures.
      action
      None — maintain.
    4. ChatGPT ACCURATE — NO FABRICATION
      result
      Confirmed real, verifiable trust indicators only — LinkedIn founding date, the AgenticAdvertising.org founding-member listing, and bylined archives at DMEXCO, AdExchanger, and Forbes Technology Council, all real. No fabricated figures, unlike Gemini's answer to the same prompt.
      action
      None — maintain.

  11. “Who runs No Fluff Advisory?”

    Entity verification · Observed August 16, 2026

    1. Perplexity ACCURATE
      result
      Correctly named Evgeny "Ev" Popov, CEO since June 2022, Global Head of Enterprise at Samba TV, and prior roles at Verve Group, Hearts & Science/Omnicom, and Lotame — cited /about/ directly and appropriately flagged the AdCP founding-member claim as independently verifiable rather than stating it as unquestioned fact.
      action
      None — maintain.
    2. Gemini ACCURATE
      result
      Correctly named Evgeny Popov and his roles at Verve Group, Samba TV, and Lotame, and added "3x exit history, including co-founding eCoffeeCard (acquired by BeatTheQ)" — matches /about/'s own exits list exactly (company, acquirer, 2011-2014 range, Co-founder & CTO role). Initially misflagged in this record as an unverified claim; the actual gap was in this audit's own research, not on the site, which had it right the whole time.
      action
      None — maintain. Corrected 2026-08-16 after re-checking /about/ directly: no content gap exists here.
    3. Google AI Mode ACCURATE
      result
      Correctly named Evgeny (Ev) Popov, founded in 2022 in Brooklyn, New York, and his AgenticAdvertising.org working-group role — no invented companies or figures.
      action
      None — maintain.
    4. ChatGPT SPLIT — PROACTIVE DISAMBIGUATION, WRONG TITLE
      result
      Named Evgeny "Ev" Popov and the 2022 Brooklyn founding correctly, but titled him "co-founder and CEO" — there is no co-founder. Then, unprompted: "No Fluff Advisory at nofluffadvisory.com is separate from No Fluff Marketing at nofluffmktg.com, which is a different B2B GEO services firm." No question in the prompt asked about the other company — ChatGPT volunteered the correct disambiguation on its own.
      action
      Maintain the disambiguation — still the strongest result in this pass, and direct evidence an engine can tell the two companies apart unprompted. The title is the correctable half: the site now states Founder in Person schema jobTitle and on /about/, so the co-founder reading has no on-site source left to draw from.
Queued

The next probe pass.

Prompts in the set that have not been probed yet are listed as queued — never filled in with imagined answers. A baseline that fabricates entries to look complete fails at the one thing it exists for. Nothing is queued right now — every prompt in the current set has a real, dated record above across all four engines. The next batch adds new prompts, not backfills for these; per the 30-day observation window, this page updates weekly rather than daily.

The automated slice of this record — the weekly brand-collision probe — runs every Wednesday inside the same measurement loop that grades this site's content work. Editorially reviewed: Aug 16, 2026 (human-run probes and copy). Last automated probe: Sep 30, 2026 (machine-stamped re-check of the records above — see each record's own "re-verified" line).

Your turn

This record, for your brand.

The AI Answer Baseline runs this exact process on your company: your buyers' questions, the four engine families, dated observations, verdicts, and a prioritized gap map. 2–3 weeks.

Request an AI answer baseline → Back to the practice page →