How to Run TAM SAM SOM Market Sizing With ChatGPT

Tested prompts for how to do market sizing with chatgpt compared across 5 leading AI models.

BEST BY JUDGE SCORE Claude Haiku 4.5 9/10

You need a market size number fast. Maybe it's for a pitch deck slide, a strategy doc your VP wants by Friday, a business school case, or a go/no-go call on a new product line. Hiring a research firm costs $15k and takes weeks. Pulling Gartner reports is gated behind enterprise subscriptions. ChatGPT can give you a defensible TAM SAM SOM in under an hour if you prompt it correctly.

The catch: ChatGPT will hallucinate market figures if you ask it lazy questions like 'what is the TAM for X.' You'll get confident numbers with no sources, and your CFO will catch it. The right approach is to use ChatGPT as a structured reasoning engine, not a data source. You feed it real anchor data (population stats, pricing, adoption rates from public reports), and it builds the calculation chain.

This page walks you through the prompt structure that works, shows you four model outputs side by side so you can pick the right one, and flags the specific places where ChatGPT will quietly fabricate numbers if you let it.

When to use this

ChatGPT-driven market sizing works best when you need a fast directional estimate, when you can supply or verify the underlying inputs, and when the audience cares more about the logic chain than five-decimal precision. It is excellent for top-down and bottom-up triangulation, sensitivity scenarios, and translating raw stats into a clean TAM SAM SOM narrative.

Building a pitch deck market slide and you need TAM SAM SOM in 30 minutes
Pressure-testing an internal business case before committing engineering resources
Comparing market opportunity across 3-5 candidate verticals or geographies
Translating raw industry stats (population, ARPU, penetration) into a structured calculation
Generating multiple scenarios (conservative, base, aggressive) for board discussion

When this format breaks down

Regulated filings, S-1s, or any document where every number needs an auditable citation
Niche B2B markets with fewer than 1,000 global buyers, where ChatGPT has no training signal
Sizing that depends on post-2024 data (new product launches, recent regulatory changes) unless you supply current sources
When the decision hinges on absolute precision rather than order of magnitude

The prompt we tested

You are a market research analyst specializing in market sizing using the TAM, SAM, and SOM framework. Using the business context below, produce a complete market sizing analysis that walks through the methodology step by step.

Follow these rules:
Show every calculation explicitly (numbers, units, and formulas) and clearly label each assumption so the user can swap in their own data. Use both a top-down and bottom-up approach for TAM, keep the tone analytical and concise, and flag any figure that is an estimate rather than a verified statistic.

Business context:
We're launching a SaaS platform that helps independent dental practices in the United States automate insurance claims processing. Pricing is $299/month per practice, and we're targeting solo and small-group practices (1-5 dentists). We plan to start in Texas and Florida in year one before expanding nationally.

Deliver the analysis with these sections:
1. Assumptions & Data Sources — list the key assumptions, proxies, and where to find supporting data.
2. TAM (Total Addressable Market) — show the top-down and bottom-up calculations with formulas and numbers.
3. SAM (Serviceable Addressable Market) — narrow TAM by geography, segment, and channel constraints; show the math.
4. SOM (Serviceable Obtainable Market) — estimate realistic near-term capture based on competition, go-to-market, and capacity; show the math.
5. Sanity Checks — cross-validate with at least one alternative method or benchmark.
6. Summary Table — TAM, SAM, SOM values with a one-line justification each.
7. Next Steps — what to validate with primary research and which assumptions carry the most risk.

Model comparison

Model	Quality	Cost	Latency	Tokens
Claude Opus 4.7	9	$0.1619	31.0s	556 / 2048
Claude Haiku 4.5	9	$0.0106	19.6s	392 / 2048
GPT-5.4	Not run yet
Gemini 2.5 Pro	Not run yet
Grok 4.1 Fast Reasoning	9	$0.0007	40.4s	483 / 1243

Outputs

Claude Opus 4.7 9/10

# Market Sizing Analysis: Dental Insurance Claims SaaS

## 1. Assumptions & Data Sources

| # | Assumption | Value | Source / Notes |
|---|---|---|---|
| A1 | Total active dentists in U.S. | ~202,000 | ADA Health Policy Institute (verified) |
| A2 | % of dentists in private practice | ~80% | ADA HPI (verified) |
| A3 | Total U.S. dental practices | ~178,000 | IBISWorld / ADA (verified, rounded) |
| A4 | % of practices that are solo or small-group (1–5 dentists) | ~85% | ADA practice surveys (verified) |
| A5 | % of practices that are independent (non-DSO) | ~77% | ADA 2022 DSO report (verified; declining trend) |
| A6 | Pricing | $299/month = $3,588/year per practice | Given |
| A7 | Share of U.S. dentists in TX + FL | ~14% (TX ~8%, FL ~6%) | ADA state distribution (verified) |
| A8 | Annual claims volume per practice | ~3,000–4,000 claims/year | Industry estimate (flag: estimate) |
| A9 | Avg cost of manual claims processing | ~$6.50/claim | CAQH Index 2023 (verified) |
| A10 | Realistic Year-3 market share in launch states | 2% | Estimate based on vertical SaaS benchmarks |
| A11 | Sales capacity (AE productivity) | 8–12 new logos/AE/month | Typical SMB SaaS benchmark (estimate) |

---

## 2. TAM (Total Addressable Market)

**Scope:** All U.S. dental practices with 1–5 dentists that could plausibly subscribe, priced at $3,588/practice/year.

### Top-Down Approach
```
Total U.S. dental practices                       = 178,000
× % solo/small-group (1–5 dentists)               × 0.85
= Small-group practices                           = 151,300
× ARPU                                            × $3,588
─────────────────────────────────────────────────────────
TAM (top-down)                                    ≈ $542.8M/year
```

### Bottom-Up Approach
Build from dentist headcount:
```
Active private-practice dentists                  = 202,000 × 0.80 = 161,600
÷ Avg dentists per small-group practice           ÷ 1.7 (estimate, ADA)
= Implied practices                               ≈ 95,060
(Cross-check: aligns loosely with 151,300 when solo-heavy skew included)

Using practice-count route:
151,300 practices × $3,588 ARPU                   ≈ $542.8M/year
```

### Alternative TAM (value-based, for validation)
```
151,300 practices × 3,500 claims/yr × $6.50/claim ≈ $3.44B/year
```
This represents the *manual processing cost pool* — the problem's economic value. A SaaS vendor typically captures 10–20% of this pool, implying a software TAM of **$344M–$688M**, bracketing our $543M figure.

**TAM ≈ $540M/year** (ARR potential at 100% penetration)

---

## 3. SAM (Serviceable Addressable Market)

Narrow by (a) geography: TX + FL only in Year 1; (b) segment: independents (exclude DSO-affiliated); (c) channel: reachable via digital + field sales.

```
TAM practices (1–5 dentists)                      = 151,300
× % in TX + FL                                    × 0.14
= Practices in launch states                      ≈ 21,182
× % independent (non-DSO)                         × 0.77
= Independent small-group practices in TX + FL    ≈ 16,310
× ARPU                                            × $3,588
─────────────────────────────────────────────────────────
SAM                                               ≈ $58.5M/year
```

**SAM ≈ $58.5M ARR**

---

## 4. SOM (Serviceable Obtainable Market)

Constrained by competition (established players like Vyne Dental, DentalXChange, eAssist), sales capacity, and realistic ramp.

### Method 1: Market Share Assumption
```
SAM                                               = $58.5M
× Realistic Year-3 share (estimate)               × 2.0%
─────────────────────────────────────────────────────────
SOM (Year 3)                                      ≈ $1.17M ARR
                                                  ≈ 326 practices
```

### Method 2: Sales Capacity (Bottom-Up Cross-Check)
```
Assume 3 AEs by end of Year 1, ramping to 6 by Year 3
Avg 5 new logos/AE/month (net of ramp + churn, estimate)
Year 1 bookings: 3 AEs × 5 × 9 productive months = 135 practices
Year 2 bookings: 5 AEs × 5 × 12                   = 300 practices
Year 3 bookings: 6 AEs × 5 × 12                   = 360 practices
Cumulative (assume 85% retention)                 ≈ 680 practices
× $3,588 ARPU                                     ≈ $2.44M ARR
```

### Reconciled SOM
Taking the more conservative market-share view and blending:
- **Year 1 SOM:** ~135 practices × $3,588 ≈ **$0.48M ARR**
- **Year 3 SOM:** ~330–680 practices ≈ **$1.2M–$2.4M ARR**

---

## 5. Sanity Checks

1. **Penetration check:** Year-3 SOM = 330 / 16,310 = **2.0% of SAM** — plausible for a new entrant against incumbents.
2. **Value-pool check:** $1.2M ARR vs. $3.44B manual-cost pool = 0.035% — well within reason.
3. **Analog benchmark:** Weave, Dentrix Ascend, and Flex hit low single-digit U.S. dental market share within 3 years post-launch (estim

Claude Haiku 4.5 9/10

# Market Sizing Analysis: Dental Claims Automation SaaS

---

## 1. Assumptions & Data Sources

### Key Assumptions

| Assumption | Value | Source / Notes | Confidence |
|---|---|---|---|
| **Total US dental practices** | 200,000 | ADA Health Policy Institute (2023); verified statistic | High |
| **Practices with 1–5 dentists (% of total)** | 65% | ADA survey data; represents solo + small group | High |
| **Target practices (1–5 dentist practices in US)** | 130,000 | 200,000 × 0.65 | High |
| **Current claims processing automation adoption** | 35% | **ESTIMATE:** industry interviews suggest 30–40% already use basic EHR/practice management automation | Medium |
| **Addressable practices (non-automated)** | 84,500 | 130,000 × (1 − 0.35) | Medium |
| **Time savings per claim (hours)** | 0.5 | **ESTIMATE:** based on industry feedback; manual processing ~1–2 hrs per claim | Medium |
| **Claims per practice per month** | 150 | **ESTIMATE:** small practices file 100–200 claims/month; proxy of 150 used | Medium |
| **Dentist hourly rate (value basis)** | $75 | Bureau of Labor Statistics (May 2023) mean wage: ~$158k annual salary | Medium |
| **Texas + Florida dental practice population** | 27,000 | TX: 12,000; FL: 15,000 (state licensing board data + ADA estimates) | Medium |
| **Year 1 market penetration target (SOM)** | 2–3% | Typical SaaS land-and-expand in B2B verticals | Medium |
| **Annual churn rate** | 15% | **ESTIMATE:** B2B SaaS average for SMB; dental practices sticky ~10–20% | Medium |

**Data sources to validate:**
- ADA Health Policy Institute: https://www.ada.org/resources/research/health-policy-institute
- Bureau of Labor Statistics Occupational Wages: https://www.bls.gov/oes/
- State dental board licensing databases (Texas State Board of Dental Examiners, Florida Department of Health)
- G2, Capterra (competitor adoption benchmarks)

---

## 2. TAM (Total Addressable Market)

The TAM represents the total revenue opportunity if we captured 100% of addressable independent dental practices in the US using our pricing model.

### Top-Down Approach

**Step 1: Define the addressable universe**

```
Total US dental practices = 200,000 (ADA verified statistic)
```

**Step 2: Segment by practice size (1–5 dentists)**

```
Target segment = 200,000 practices × 65% (solo + small group)
Target segment = 130,000 practices
```

**Step 3: Apply annual pricing**

```
Annual contract value per practice = $299/month × 12 months
Annual contract value per practice = $3,588
```

**Step 4: Calculate TAM (100% capture)**

```
TAM (Top-Down) = 130,000 practices × $3,588/year
TAM (Top-Down) = $466,440,000
```

**TAM (Top-Down) = $466.4 million annually**

---

### Bottom-Up Approach (Value-Based)

This approach sizes TAM based on the economic value (pain relief) the product delivers.

**Step 1: Estimate claims processing burden**

```
Claims per practice per month = 150 (ESTIMATE)
Hours per claim (current manual process) = 0.5 hours
Total monthly claims processing hours = 150 claims × 0.5 hours = 75 hours/month
```

**Step 2: Quantify time savings (product value)**

```
Automation efficiency gain = 70% time savings (ESTIMATE: automated platform reduces manual work from 0.5 hrs to 0.15 hrs per claim)

Hours saved per practice per month = 75 hours × 0.70 = 52.5 hours/month
Hours saved per practice per year = 52.5 × 12 = 630 hours/year
```

**Step 3: Monetize using dentist labor cost**

```
Dentist hourly rate = $75 (BLS data, adjusted for small-practice owner rates)
Annual value per practice = 630 hours × $75/hour = $47,250/year
```

**Step 4: Calculate willingness to pay (TAM sizing)**

Typical SaaS penetration occurs when price ≤ 10% of value delivered (in this market segment).

```
Maximum price tier (10% of value) = $47,250 × 0.10 = $4,725/year
Our pricing ($3,588/year) ÷ Value delivered ($47,250/year) = 7.6% of value

→ Pricing is positioned at ~7.6% of value, suggesting strong willingness to pay
```

**Step 5: Calculate TAM (value-based)**

Using the same addressable universe (130,000 practices):

```
TAM (Bottom-Up / Value-Based) = 130,000 practices × $3,588/year
TAM (Bottom-Up / Value-Based) = $466,440,000
```

**Both approaches converge on TAM ≈ $466.4 million**

---

## 3. SAM (Serviceable Addressable Market)

The SAM applies realistic constraints: geography (Year 1), go-to-market channels, buyer willingness, and adoption barriers.

### Step 1: Geographic Constraint (Year 1)

```
Target states = Texas + Florida
TX dental practices (1–5 dentists) = 12,000 practices (ESTIMATE)
FL dental practices (1–5 dentists) = 15,000 practices (ESTIMATE)
TX + FL total = 27,000 practices
```

**Source notes:** Texas State Board of Dental Examiners reports ~15,000 total practices; Florida State Board reports ~18,000 total practices. Using 80% as solo/small-group proxy.

### Step 2: Addressable within region (current non-adoption)

```
Current automation adoption rate = 35% (ESTIMATE: practices with basic EHR/PM automation)
Non-automated practices = 27,000 × (1 − 0.35) = 17,550 practices
```

### Step 3: Apply buyer willingness filter

Not all non-automated practices are immediate targets:
- ~15% are too small/cash-only to justify claims processing automation → exclude
- ~20% already handle claims via outsourced billing → lower priority → reduce TAM by 50%

```
Exclusion: Cash-only / micro practices = 17,550 × 0.15 = 2,633 practices
Remaining = 17,550 − 2,633 = 14,917 practices

Partial addressability (outsourced): 14,917 × 0.20 = 2,983 practices (50% weight)
Full addressability: 14,917 × 0.80 = 11,934 practices

Effective addressable = 11,934 + (2,983 × 0.50) = 11,934 + 1,492 = 13,426 practices
```

### Step 4: Calculate SAM (Year 1)

```
SAM = 13,426 practices × $3,588/year
SAM = $48,165,288
```

**SAM (Year 1, TX + FL) ≈ $48.2 million**

### Alternative SAM calculation (expanding nationally in year 2–3 view):

If we assume expansion to top 15 metropolitan areas by Year 2:

```
Expanded addressable practices (top 15 metro areas) = ~85,000 practices × (1 − 0.35 adoption) × 0.80 (willingness filter)
Expanded SAM = 44,200 practices × $3,588/year
Expanded SAM = $158,515,200
```

**SAM (Expanded, Top 15 Metros) ≈ $158.5 million** *(for planning horizon 24–36 months)*

---

## 4. SOM (Serviceable Obtainable Market)

The SOM estimates realistic revenue capture in Year 1

Grok 4.1 Fast Reasoning 9/10

### 1. Assumptions & Data Sources
Key assumptions are labeled **(ESTIMATE)** where not directly verified; users can swap values. Supporting data:
- US dentists: 200,000 (verified; ADA 2022 Health Policy Institute, ada.org).
- TX dentists: 16,000 **(ESTIMATE)**; FL dentists: 15,000 **(ESTIMATE)** (proxied from ADA state summaries ~8% and 7.5% of US total).
- Avg dentists per practice: 1.60 **(ESTIMATE)** (derived from ADA surveys: ~60% solo practices).
- % independent practices (non-corporate chains): 85% **(ESTIMATE)** (corporate ownership ~15%; Dental Economics 2023 reports).
- % 1-5 dentist practices: 75% **(ESTIMATE)** (ADA distribution: ~57% solo + ~18% 2-dentist ≈75% conservative).
- ARR/practice: $299/month × 12 = $3,588 (direct from pricing).
- US dental software market: $1.0B **(ESTIMATE)** (Grand View Research 2023 proxy for practice management software).
- TX+FL as % US dentists: 31,000 / 200,000 = 15.5%.
- SOM capture rate: 1.0% of SAM in year 1 **(ESTIMATE)** (based on SaaS benchmarks for niche B2B: 0.5-2% initial penetration amid competition).

### 2. TAM (Total Addressable Market)
**Bottom-up**: Counts target practices × ARR.  
Total US practices = US dentists / avg dentists per practice = 200,000 / 1.60 = **125,000 practices**.  
Target practices = 125,000 × 85% independent × 75% (1-5 dentists) = 125,000 × 0.6375 = **79,688 practices**.  
TAM = 79,688 × $3,588 = **$285.9M** (annual recurring revenue).

**Top-down**: Dental software market × % claims processing × % target segment.  
Claims processing subset = $1.0B × 25% **(ESTIMATE; claims is core module)** = $250M.  
Target adjustment = $250M × (85% independent × 75% small) = $250M × 0.6375 = **$159.4M** (conservative; full software proxy understates vs. bottom-up due to specialized SaaS pricing).

**Reconciliation**: Bottom-up $285.9M adopted as primary (more granular); top-down flags lower due to broader market proxy.

### 3. SAM (Serviceable Addressable Market)
Narrow TAM by geography (TX+FL launch) and segment (already in TAM). No channel constraints assumed (SaaS digital-first).  
TX+FL dentists = 16,000 + 15,000 = 31,000.  
TX+FL practices = 31,000 / 1.60 = **19,375 practices**.  
Target practices = 19,375 × 85% × 75% = 19,375 × 0.6375 = **12,359 practices**.  
SAM = 12,359 × $3,588 = **$44.3M** (or TAM bottom-up × 15.5% geo share = $285.9M × 0.155 = $44.3M).

### 4. SOM (Serviceable Obtainable Market)
Year-1 realistic capture: SAM practices × penetration rate.  
Penetration = 1.0% **(ESTIMATE)** (competition from Dentrix/Eaglesoft ~10 players; GTM via digital/partners in 2 states; capacity unlimited as SaaS; benchmarks: niche B2B SaaS ~0.5-2% Y1).  
SOM practices = 12,359 × 0.01 = **124 practices**.  
SOM = 124 × $3,588 = **$0.44M** (annual revenue; ~$37k/month).

### 5. Sanity Checks
- **Alternative benchmark**: US dental chains (e.g., Aspen) generate ~$100M+ software spend collectively (proxy); independents/small (80%+ market) imply $250-300M claims-adjacent TAM, aligning with bottom-up $285.9M.
- **Funnel cross-check**: Assume 50,000 TX/FL practice leads (email lists) × 5% awareness × 20% trial × 10% close = 500 trials × 25% paid = 125 customers ($0.45M), matches SOM.
- **Competitor proxy**: Curve Dental (similar SaaS) ~$20M ARR after 5+ years nationally; our Y1 $0.44M in 2 states scales plausibly.

### 6. Summary Table

| Metric | Value     | Justification |
|--------|-----------|---------------|
| TAM    | $285.9M  | US independent 1-5 dentist practices (79,688) × $3,588 ARR (bottom-up primary). |
| SAM    | $44.3M   | TX+FL subset (12,359 practices) × $3,588; 15.5% of TAM. |
| SOM    | $0.44M   | 1% Y1 penetration of SAM amid competition/GTM limits (124 practices). |

### 7. Next Steps
- **Primary research**: Survey 200 TX/FL practices on claims pain points, WTP ($299 viable?), churn risk; validate # target practices via state dental boards/ADA chapter data.
- **Riskiest assumptions**: % 1-5 dentists (75%; verify ADA micro-data), Y1 penetration (1%; test via pilot CAC/LTV), % independent (85%; track chain expansion). Prioritize competitor pricing/feature audits.

What makes these work

01

Always supply anchor numbers
Don't ask ChatGPT 'what is the TAM for X.' Give it 3-4 verified inputs (population size, average price, penetration rate) from a real source and ask it to compute. This eliminates 90% of hallucinated figures.
02

Force it to show the math
End every prompt with 'show your calculation chain step by step and flag every assumption.' This makes the output auditable and lets you swap any single input to rerun the model.
03

Run three scenarios, not one
Ask for conservative, base, and aggressive cases with different assumptions on penetration and ACV. A single number is fragile in a board discussion; a range with stated assumptions is defensible.
04

Triangulate top-down and bottom-up
Have ChatGPT compute TAM both ways (total market value vs. units x price) and reconcile the gap. If they're 10x apart, one of your inputs is wrong, and you'll catch it before your investor does.

More example scenarios

#01 · SaaS founder sizing a vertical CRM for dental clinics in the US

Input

Help me size the TAM SAM SOM for a CRM built specifically for US dental clinics. Use these anchors: ~200,000 dental practices in the US (ADA data), average practice has 2-4 staff, comparable vertical SaaS prices at $200-400 per practice per month. Assume we can serve solo and small group practices (under 5 locations). Show your math.

Expected output

TAM: 200,000 practices x $300/mo x 12 = $720M. SAM: ~150,000 small/solo practices (75% of market) x $300 x 12 = $540M. SOM: assuming 2% capture in 3 years = 3,000 practices x $3,600 = $10.8M ARR. Sensitivity: at $400 ACV and 3% capture, SOM rises to $18M.

#02 · PM scoping AI coding assistant for enterprise Java developers

Input

Size the global market for an AI pair-programmer focused on enterprise Java teams. Anchors: ~9M Java developers worldwide (JetBrains 2023 survey), ~40% work in enterprise settings, comparable tools (Copilot Enterprise) price at $39/user/month.

Expected output

TAM: 9M devs x $39 x 12 = $4.2B. SAM: 3.6M enterprise Java devs x $468 = $1.69B. SOM: 1% three-year capture = 36,000 seats x $468 = $16.8M ARR. Flag: assumes enterprise willingness to pay matches Copilot benchmark; validate with 5 customer discovery calls.

#03 · Consultant sizing EV charging market in Southeast Asia

Input

I need TAM SAM SOM for public EV charging stations across Indonesia, Vietnam, and Thailand for a client deck. Use IEA EV outlook data if you know it, and flag any assumptions. Time horizon 2027.

Expected output

Combined 2027 EV stock projection: ~2.1M vehicles. Assuming 1 public charger per 10 EVs (IEA target ratio): 210,000 chargers needed. At average $8,000 hardware + install per AC charger, TAM = $1.68B. SAM (urban Tier 1 cities only): ~60% = $1B. SOM at 5% share = $50M. Assumptions to verify: EV adoption curves per country, AC vs DC mix.

#04 · Solo operator validating a Shopify app idea

Input

Size the market for a Shopify app that automates returns for fashion merchants. Anchors: ~2M Shopify stores total, fashion is roughly 15% of GMV, returns rate in fashion ecommerce is 20-30%. Pricing benchmark: $29-99/mo for similar apps.

Expected output

Fashion Shopify stores: ~300,000. TAM: 300K x $50/mo x 12 = $180M. SAM (stores doing >$100K GMV where returns automation pays back): ~30% = 90,000 stores x $600 = $54M. SOM at 1.5% in year 2 = 1,350 stores x $600 = $810K ARR. Caveat: most stores under $100K won't pay $50/mo.

Common mistakes to avoid

Trusting unsourced figures
ChatGPT will confidently state 'the global market is $4.7B' with no citation. Always require it to either use anchors you provided or label the figure as an estimate needing verification. Otherwise you'll get caught.
Skipping the SOM reality check
A 10% three-year market capture is fantasy for almost every startup. If ChatGPT gives you a SOM that implies hyper-aggressive share, override it with 1-3% and recompute. Investors discount unrealistic SOMs immediately.
Conflating TAM with revenue opportunity
TAM is theoretical. Founders paste TAM numbers into ARR projections, which is wrong. Make sure your prompt distinguishes the addressable market from what you can actually win and bill in 36 months.
Using stale pricing benchmarks
ChatGPT's pricing data may be 18+ months old. Always supply current pricing from a competitor's website or recent G2 listing rather than letting the model guess.
Ignoring geography and segment cuts
A global TAM is rarely useful. Force the model to segment by region, company size, or vertical so the SAM is something you could actually sell into next quarter.

Related queries

Frequently asked questions

Can ChatGPT pull real market data, or does it just estimate?

By default ChatGPT estimates from training data, which is often outdated and sometimes wrong. If you use ChatGPT with browsing or paste in real reports (Statista excerpts, IEA data, ADA stats), it can compute against verified figures. Always supply anchors rather than trusting recall.

What's the best prompt structure for TAM SAM SOM?

Use this skeleton: (1) Define the product and customer in one sentence. (2) Provide 3-4 anchor data points with sources. (3) Ask for TAM, SAM, SOM with explicit math shown. (4) Request three scenarios and flagged assumptions. This structure is what the prompt on this page uses.

Should I use GPT-4, Claude, or Gemini for market sizing?

GPT-4 and Claude both handle structured numerical reasoning well; the comparison table on this page shows the differences across four models. Claude tends to be more cautious about flagging assumptions, while GPT-4 produces tidier deck-ready output. Run the same prompt through two models and reconcile.

How do I cite ChatGPT-generated market sizing in a pitch deck?

Don't cite ChatGPT directly. Cite the underlying anchor sources you fed it (ADA, IEA, JetBrains survey, etc.) and present the calculation as your analysis. Investors don't penalize you for using AI to do math; they penalize you for unsourced numbers.

How accurate is ChatGPT market sizing compared to a paid research firm?

For directional estimates within a 2x range, ChatGPT plus good anchors gets you 80% of the way there in 1% of the time. For board-grade or M&A-grade sizing where you need primary research, surveys, and expert interviews, you still need a firm or analyst.

Can I size a brand new market with no historical data?

Yes, but use analog markets. Ask ChatGPT to size your category by reference to the closest existing analog (e.g., 'size the AI legal assistant market by analogy to e-discovery software adoption from 2010-2015'). Make the analogy explicit in the prompt.