Four products on one platform
All four were designed and built by Doug Brown in 2026. Both main platforms have US provisional patent applications in his name.
Runs Google Local Services Ads accounts end to end: syncs every lead, transcribes and grades every recorded call on its own servers, watches verification and serving, re-analyses budget every two hours and finds each account's budget sweet spot.
HELM is its expert layer: a separate AI service bound to a 265-concept Local Services doctrine and a 36-rule list of advice it must never give. Its budget decisions are advisory today; they are not yet written to Google automatically.
Provisional patent 64/063,539, filed 12 May 2026.
ClickRadiusMakes a business readable and citable by AI search. It scores the site, generates the fixes and deploys them to Shopify, WordPress or any site by snippet, writes fact-checked content and measures citations across six AI engines.
Closed loop, white-label and priced for small businesses and agencies, where funded competitors mostly sell monitoring to enterprises.
Provisional patent 64/063,349, filed May 2026.
A $29-a-month (Pro $79) read-only view of a business's own Local Services account: leads, calls, rivals, geography and timing, with HELM answering questions. Built for every prospect who doesn't buy full management.
Its technology figure is small because it reuses CallRadius. Its value will come from subscribers.
ClickRadiusSIGNALThe self-serve companion to ClickRadius: tracks AI mentions, website citations, competitor share of voice, prompts and sources. Pro adds up to five prepared repair cards and one researched article each month.
Today it has its commercial-acceptance and business-profile foundations and an interactive sample workspace. It reuses the ClickRadius engine, so most of its value will come from subscribers.
Measured on the production systems
CallRadius, ClickRadius and CallRadius Signal were counted on the live server on 22 October 2026, and ClickRadius Signal on 26 October 2026. The August figures were re-measured the same way from the code as it stood on 9–11 August. "Logical lines" excludes blank lines and comments. Tests are counted separately and are not part of the value of the code.
| Measure | August 2026 | Today | Growth |
|---|---|---|---|
| CallRadius | |||
| Production code, same files as August (logical lines) | 25,441 | 185,064 | 7.3× |
| Full production code incl. sales flows and HELM (logical lines) | – | 212,873 | |
| API endpoints, counted the same way | 226 | 355 | 1.6× |
| Database tables | 36 | 166 | 4.6× |
| Automatic monitoring loops per account | 17 | 25 | 1.5× |
| Automated tests (lines) | – | 311,882 | |
| ClickRadius | |||
| Application code, all non-test JavaScript (logical lines) | ≈34,900 | 212,361 | 6.1× |
| API endpoints, counted the same way | 357 | 593 | 1.7× |
| Database tables | 76 | 232 | 3.1× |
| Application screens, same 28 pages (lines) | 31,017 | 42,450 | 1.4× |
| Automated tests (lines) · last recorded run | 5,481 | 268,734 · 7,392 of 7,392 passed | |
| CallRadius Signal (built after August) | |||
| Production code (logical lines) · endpoints · tables | – | 10,777 · 21 · 9 | |
| Automated tests (lines) | – | 6,134 | |
| ClickRadius Signal (started 24 October) | |||
| Production code incl. schema and workspace view (logical lines) · endpoints · tables | – | 2,452 · 11 · 8 | |
| Automated tests (lines) | – | 3,100 | |
Endpoint counts include each route's full path. Counted that way, August had 226 (CallRadius) and 357 (ClickRadius), more than the 218 and 279 the August report printed, which left out each route's prefix. CallRadius also has 74 more routes inside shared application pages, 429 in all. ClickRadius's last recorded full test run was on 17 October 2026. ClickRadius Signal's 2,452 lines are not inside ClickRadius's 212,361, which were counted before it was started.
Data the products have already gathered
A competitor can buy developers. It can't buy this operating history, which is why it is priced separately below.
How the figure is reached
Three standard approaches. The cost approach asks what it would cost to rebuild what exists today with 2026 methods, less anything not yet working. The income approach asks what the business could earn, weighted by how likely each outcome is. The market approach asks what comparable businesses sell for. Each adjustment is shown, including those that lower the figure.
| Step | Basis | Effect | Running total |
|---|---|---|---|
| 1 · Cost approach: rebuild cost, less obsolescence | |||
| CallRadius core | August's measured rebuild cost (centre $1.30M) × 1.52 growth in what the system does | $1.98M | $1.98M |
| ClickRadius | August's measured rebuild cost (centre $1.50M) × 1.51 growth in what the system does | $2.26M | $4.24M |
| HELM expert service | Separate 12,900-line AI service plus the curated 265-concept doctrine; not in August | $0.30M | $4.54M |
| CallRadius Signal | 21 endpoints, subscriptions with trial, Google connect, 10-view portal; reuses CallRadius | $0.15M | $4.69M |
| ClickRadius Signal | 11 endpoints and 8 tables so far, sample workspace; scaled against CallRadius Signal by endpoints and code | $0.07M | $4.76M |
| Rebuilding with 2026 methods | A rebuild today would use AI coding agents. Studies of enterprise teams find 0–26% gains; this build itself (one founder directing AI agents over about six months) points to far more for new code. 35% sits between. | −35% | $3.09M |
| Built but not yet live | ClickRadius card payments, both Signal enrollments and HELM's automatic budget writes are not yet on for customers | −10% | $2.78M |
| Google's Local Services change | Google is moving LSA into Performance Max; some CallRadius controls lose their surface | −10% on CallRadius, HELM, CallRadius Signal | $2.64M |
| Product direction | The Local Services and AI-search knowledge that decided what to build (August: $200K–350K, midpoint) | +$0.28M | $2.92M |
| Assets money can't buy | Operating data above, two May 2026 patent priority dates, months of live calibration, time to market | +15% | $3.36M |
| Cost approach | $3.36M | ||
| 2 · Income approach: outcomes weighted by likelihood | |||
| Plan | The partnership briefing's planning case: $3.75M monthly revenue by month 60. Valued today at $46.9M; likelihood 4% | $1.88M | $1.88M |
| Half plan | Half the planned subscribers: $1.88M monthly revenue by month 60. Valued today at $22.8M; likelihood 8% | $1.82M | $3.70M |
| Slow growth | Grows steadily from today's early base to $150K monthly revenue over five years. Valued today at $1.3M; likelihood 33% | $0.42M | $4.11M |
| Stall | Doesn't scale; the technology is sold for about $0.75M in year 2. Valued today at $0.44M; likelihood 55% | $0.24M | $4.36M |
| Income approach | $4.36M | ||
| 3 · Market approach | |||
| Revenue multiple | Small software businesses under $1M of annual revenue sell for about 2.5–4× revenue. Applied to today's early CallRadius revenue it gives a small floor that credits nothing for ClickRadius, the patents or the platform. | floor only | 0% weight |
| 4 · Conclusion | |||
| Appraised value | 50% cost approach + 50% income approach | $3.86M | |
Sensitivity of the cost approach to the AI-era rebuild discount: −15% gives $4.29M, −25% $3.82M, −35% $3.36M (used), −50% $2.65M, −60% $2.19M. The reasonable range of $3.2M–$4.6M combines the −50% and −25% cases with the cautious and generous likelihoods above.
Why the figure doesn't follow the line count
Measured the August way, both codebases are 6–7 times larger than in August. Scaling the August value up by that would give a much larger number. It would not survive a buyer's review, for three reasons:
- Much of the new code is protection, not features. Safety checks, verified writes to Google, audit records and release gates make the system trustworthy with live budgets. They are worth having, but they do not multiply what a customer can do.
- Code written with AI help runs longer than the hand-written code the cost benchmarks were based on.
- A buyer's expert would argue it, so this appraisal applies the reduction itself, at a rate the evidence supports.
For the same reason, the nearly 590,000 lines of automated tests are not counted as value, though they make the products far cheaper to keep improving.
What buyers and investors are paying in 2026
These are not like-for-like comparisons, which is why they carry no weight in the figure. Each company had a team, paying customers or venture funding. They show that both categories are attracting serious money, and what proven traction is worth.
| Company | Category | Event | Value |
|---|---|---|---|
| Profound | AI-search visibility | Series D, 15 Sep 2026, Sequoia and Kleiner Perkins | $1.8B |
| Profound | AI-search visibility | Series C, 24 Feb 2026, Lightspeed | $1.0B |
| Scrunch | AI-search visibility | Acquired by Sitecore, Jun 2026, about $19M raised (price reported by Bloomberg) | ≈$225M |
| Peec AI | AI-search visibility | In talks to raise, Jun 2026; reported ARR above $10M | ≈$200M pre-money |
| XFunnel | AI-search visibility | Acquired by HubSpot, Oct 2025 (reported price) | ≈$30M |
| Avoca | AI for home-service contractors | $125M+ raised, Apr 2026; calls, booking, marketing for HVAC and plumbing | $1.0B |
| ServiceTitan | Contractor software (public) | Market value, 22 Oct 2026 | ≈7.4× sales |
| Semrush | Search marketing (public) | Adobe acquisition, announced Nov 2025 | $1.9B, ≈4× ARR |
Version 1 used the 2026 median AI seed pre-money valuation ($21.0M, PitchBook-NVCA) as a market approach. It is not used here: seed valuations are prices for preferred stock in companies that raised money with teams and usually traction, and there is no round here to calibrate against.
What moves the figure up
The biggest increases come from proof that customers pay, because that raises the likelihoods in the income approach and brings revenue multiples into play.
Paying customers
Turn on card payments for ClickRadius and open enrollment for both Signal products. Removing the "not yet live" deduction alone adds about $0.3M; paying customers move the income approach far more.
Full patent applications
Both provisionals must be converted by May 2027. A filed full application, and later a grant, turns 67 unexamined claims each into examinable IP.
Agency channel at work
Signed agencies reselling CallRadius and ClickRadius prove distribution, the single biggest lever on the income approach.
Earlier valuations
| Date | Covered | Result | Status |
|---|---|---|---|
| 1 May 2026 | ClickRadius only | $3.2M fair value (range $1.8M–5.5M) | Superseded |
| 10 Aug 2026 | CallRadius and ClickRadius | $4.5M–9M fair value; $3.0M–3.5M rebuild basis | Superseded |
| 22 Oct 2026 (v1) | CallRadius with HELM, ClickRadius, CallRadius Signal | $6.4M | Superseded: too high |
| 26 Oct 2026 (v2) | CallRadius with HELM, ClickRadius, CallRadius Signal, ClickRadius Signal | $3.86M (range $3.2M–4.6M) | Current |
Version 1 was re-reviewed to come as close as possible to what an independent appraiser would conclude. Its market leg ($6.83M, 40% weight) rested on venture seed valuations; its 15% AI-era rebuild discount was too small for a build of this kind; it counted database tables as growth in function; and it made no deduction for features not yet live. Version 2 corrects each of these, adds an income approach, and adds ClickRadius Signal. The August "$4.5M–9M fair value" relied on similar market and scenario evidence and is superseded for the same reasons.
Limitations
- Internal appraisal, not certified. It is built to approximate what an independent credentialed appraiser (ABV, ASA or CVA) would conclude. A certified (USPAP) valuation for a transaction, tax or litigation needs that appraiser, and theirs governs.
- No vendor bids. The rebuild cost is modelled from the August bottom-up pricing and measured growth, not from quotes against a written specification.
- Judgment calls are visible. The 35% AI-era discount, the 10% not-yet-live deduction, the 10% Google adjustment, the 15% premium, the income likelihoods, the 30% rate, the 3× exit multiple and the 50/50 weighting are reasoned choices, each shown above with sensitivities.
- Platform dependency. CallRadius and CallRadius Signal depend on Google's Ads API and policies; ClickRadius and ClickRadius Signal depend on how AI engines choose sources.
- Key person. One founder designed and built all four products.
- Ownership of the IP must be confirmed and assigned to the company before any certified appraisal or transaction.
Sources
- Code, endpoint, table and data counts: measured on the production servers, 22 Oct 2026 (ClickRadius Signal 26 Oct 2026); August figures re-measured from the 9–11 Aug 2026 code.
- Income scenarios: the partnership briefing's monthly model (planning and half cases), 24 Oct 2026.
- Earlier valuations: Technology Asset Appraisal, Rev. R3, 10 Aug 2026; ClickRadius Valuation Analysis, 1 May 2026; Technology Valuation v1, 22 Oct 2026 (all internal).
- Small-software multiples: FE International; summary of Acquire.com 2026 listings (indicative).
- Venture-capital rates of return by stage: AICPA Accounting and Valuation Guide, Appendix B (subscription).
- Profound Series D: Profound newsroom, 15 Sep 2026 · Series C: Fortune, 24 Feb 2026
- Scrunch: Bloomberg, 3 Jun 2026; Adweek (Sitecore did not confirm the price)
- Peec AI: The Next Web, 29 Jun 2026 · XFunnel: MarketScreener
- Avoca: Pulse 2.0, 27 Apr 2026 · ServiceTitan: StockAnalysis, 22 Oct 2026
- Semrush: Adobe, 19 Nov 2025; Semrush FY2025 results
- Seed valuations (v1 only, not used here): PitchBook-NVCA Venture Monitor, Q3 2026
- AI and developer productivity: MIT / Microsoft / Accenture field experiments; METR, Feb 2026; Computer Weekly on Gartner, AI coding costs
- Developer rates: Arc, May 2026
- Google LSA move to Performance Max: Search Engine Journal, 20 Jul 2026; Google Ads API docs
- Consumers and AI: BrightLocal, 10 Mar 2026; Alphabet Q2 2026