📈 AI agents gut the outsourced customer-support industry
Klarna's press release — an OpenAI-built agent doing the work of 700 full-time reps, two-thirds of chats, $40M of profit — is the first production proof with measurable economics, and it reprices the entire outsourced-customer-support complex rather than just Klarna. Chamath's call: Teleperformance lost ~20% of its market cap ($1.7-1.8B) the day of the tweet and goes from ~$8B to $1B 'in short order' once two or three more clients publish results; Sacks says the next leg is voice/phone, which is the rest of the call-center industry.
0
CONVICTION
peaked 61.9 WATCH
EMBER
band
▼ SHORT CNXC
expression · bearish equities
HIT
outcome · R +37.0% · α +19.6%
2025-03-01
window closed
⚖ Why this verdict
fully deterministic — evaluate.py replays this from daily closes; nothing below is editable or hand-set
- Window: 2024-03-01 → 2025-03-01 — first mention + 12-month horizon, then the window locks.
- The call: ▼ SHORT CNXC (primary play). CNXC fell 37.0% over the window → direction-adjusted R = +37.0% (the call made money).
- Benchmark: holding SPY over the same window returned +17.3%
→ α = +37.0 − (+17.3) = +19.6%
— what this call made or lost against just owning the index. This is the number the verdict uses.
Stock-selection read: +54.3% — did CNXC move the predicted way relative to the market. The two only differ on a SHORT: this call was right about direction relative to the index but still cost money versus holding it.
- Rule fired:
▶ HIT — R ≥ +10% AND α ≥ +5 ✓ (R 37.0, α 19.6)· PARTIAL — R ≥ +5% OR α ≥ 0· MISS — everything else
- Credit: supporters of a HIT earn 1.0 each, opposers the inverse — this feeds the scoreboard weights. supported: Chamath, Jason, Sacks | proxy-sensitive: G→MISS (-59.4%)
Conviction timeline
bands: green ≥ 65 · watch ≥ 45 · ember ≥ 15
Plays vs SPY · % since first mention (2024-03-01)
Plays
| expression | symbol | kind | relevance | rationale |
|---|---|---|---|---|
| ▼ SHORT | CNXC | stock | PRIMARY | largest US-listed customer-experience BPO, direct comp to Teleperformance |
| ▼ SHORT | G | stock | adjacent | Genpact — labor-arbitrage business-process outsourcing exposed to the same substitution |
| ▼ SHORT | TTEC | stock | adjacent | pure-play outsourced customer-support seats, most levered to agent displacement |
Mention log
Chamath
equities w=0.84 · n=160
· E168 (2024-03-01) · explicit_prediction · strength 3
▶ 1:00:03
SUPPORT
“Yes, Klarna replaced 700 people and they saved $40 million of OPEX, but Teleperformance, while they were just doing their everyday work, lost $1.8 billion of their market cap at the exact same moment. ... But I just wanted to show you that the destruction was quite quick and it was pretty severe. And if two or three other big companies launch these kinds of tweets after real measurable results, Teleperformance will be a $1 billion company in short order.”
Sacks
equities w=0.89 · n=96
· E168 (2024-03-01) · explicit_prediction · strength 2
▶ 1:06:31
SUPPORT
“So, what Klarna seems to be talking about are email-based customer support cases. I think where this is going to go next is to phone. ... I think where it goes next is you'll call up the call center and you'll get a voice that sounds like a human, just talk to you. And you won't even necessarily realize that you're talking to an AI, because there are already these AI companies that can do generative voices, any language, any accent.”
Sacks
equities w=0.89 · n=96
· E179 (2024-05-17) · explicit_prediction · strength 2
▶ 26:18
SUPPORT
“I think their product roadmaps just became obsolete. Now that's not to say there isn't more work for them to do in workflow in terms of integrating the AI with customer support tools and doing that last mile of customizing the model for the vertical specific problems of customer support. But my guess is that hundreds of millions of dollars of R&D just went out the window.”
Sacks
equities w=0.89 · n=96
· E197 (2024-09-27) · explicit_prediction · strength 2
▶ 21:30
SUPPORT
“But you could imagine that once system integrators or enterprise SaaS apps go into these places, go into these companies, they integrate the data and then they map out the workflow, you could replace a lot of these steps in the workflow with agents.”
Who built this conviction
each voice's total force on the score — supports and opposes from every mention, weighted exactly as the replay applied them · share = % of all mention-driven movement
⏳ decay drained -70.4 over the idea's life — that's time passing, attributed to no one
Score events
| episode | kind | Δ | after | note |
|---|---|---|---|---|
| E168 2024-03-01 | init | +35.2 | 35.2 | E168 born by Chamath (explicit_prediction x3) [w=0.84] |
| E168 2024-03-01 | reinforce | +12.9 | 48.1 | E168 Sacks support x2 (new voice) [w=0.89] |
| E169 2024-03-08 | decay | -1.9 | 46.2 | E169 silent |
| E170 2024-03-15 | decay | -1.8 | 44.3 | E170 silent |
| E171 2024-03-22 | decay | -1.8 | 42.5 | E171 silent |
| E172 2024-03-29 | decay | -1.7 | 40.8 | E172 silent |
| E173 2024-04-05 | decay | -1.6 | 39.2 | E173 silent |
| E174 2024-04-12 | decay | -1.6 | 37.6 | E174 silent |
| E175 2024-04-19 | decay | -1.5 | 36.1 | E175 silent |
| E176 2024-04-26 | decay | -1.4 | 34.7 | E176 silent |
| E177 2024-05-03 | decay | -1.4 | 33.3 | E177 silent |
| E178 2024-05-10 | decay | -1.3 | 32.0 | E178 silent |
| E179 2024-05-17 | reinforce | +9.0 | 41.0 | E179 Sacks support x2 [w=0.89] |
| E180 2024-05-24 | decay | -1.6 | 39.4 | E180 silent |
| E181 2024-05-31 | decay | -1.6 | 37.8 | E181 silent |
| E182 2024-06-07 | decay | -1.5 | 36.3 | E182 silent |
| E183 2024-06-14 | decay | -1.5 | 34.8 | E183 silent |
| E184 2024-06-20 | decay | -1.4 | 33.4 | E184 silent |
| E185 2024-06-29 | decay | -1.3 | 32.1 | E185 silent |
| E186 2024-07-04 | decay | -1.3 | 30.8 | E186 silent |
| E187 2024-07-12 | decay | -1.2 | 29.6 | E187 silent |
| E188 2024-07-19 | decay | -1.2 | 28.4 | E188 silent |
| E189 2024-07-26 | decay | -1.1 | 27.3 | E189 silent |
| E190 2024-08-02 | decay | -1.1 | 26.2 | E190 silent |
| E191 2024-08-09 | decay | -1.0 | 25.1 | E191 silent |
| E192 2024-08-16 | decay | -1.0 | 24.1 | E192 silent |
| E193 2024-08-23 | decay | -1.0 | 23.2 | E193 silent |
| E194 2024-08-30 | decay | -0.9 | 22.2 | E194 silent |
| E195 2024-09-06 | decay | -0.9 | 21.3 | E195 silent |
| E196 2024-09-20 | reinforce | +12.5 | 33.9 | E196 Sacks support x3 [w=0.89] |
| E196 2024-09-20 | reinforce | +14.5 | 48.3 | E196 Jason support x2 (new voice) [w=0.97] |
| E196 2024-09-20 | reinforce | +7.8 | 56.1 | E196 Chamath support x3 [w=0.84] |
| E197 2024-09-27 | reinforce | +5.8 | 61.9 | E197 Sacks support x2 [w=0.89] |
| E198 2024-10-03 | decay | -2.5 | 59.5 | E198 silent |
| E199 2024-10-11 | decay | -2.4 | 57.1 | E199 silent |
| E200 2024-10-18 | decay | -2.3 | 54.8 | E200 silent |
| E201 2024-10-25 | decay | -2.2 | 52.6 | E201 silent |
| E202 2024-11-01 | decay | -2.1 | 50.5 | E202 silent |
| E203 2024-11-08 | decay | -2.0 | 48.5 | E203 silent |
| E204 2024-11-16 | decay | -1.9 | 46.5 | E204 silent |
| E205 2024-11-23 | decay | -1.9 | 44.7 | E205 silent |
| E206 2024-12-07 | decay | -1.8 | 42.9 | E206 silent |
| E207 2024-12-13 | decay | -1.7 | 41.2 | E207 silent |
| E208 2024-12-20 | decay | -1.6 | 39.5 | E208 silent |
| E209 2025-01-04 | decay | -1.6 | 38.0 | E209 silent |
| E210 2025-01-11 | decay | -1.5 | 36.4 | E210 silent |
| E211 2025-01-18 | decay | -1.5 | 35.0 | E211 silent |
| E212 2025-01-25 | decay | -1.4 | 33.6 | E212 silent |
| E213 2025-01-31 | decay | -1.3 | 32.2 | E213 silent |
| E214 2025-02-07 | decay | -1.3 | 30.9 | E214 silent |
| E215 2025-02-15 | decay | -1.2 | 29.7 | E215 silent |
| E216 2025-02-21 | decay | -1.2 | 28.5 | E216 silent |
| E217 2025-03-01 | decay | -1.1 | 27.4 | E217 silent |