Early revenue traps and how Accel picks lasting AI companies

This article summarizes an episode of 20VC’s video series featuring Miles Clements, a partner at Accel.

Miles Clements, a partner at Accel/ Photo credit: Accel

Clinging to old financial rules in a fast moving technology market is a dangerous game. Miles Clements, a partner at the investment firm Accel, warns that rigid pricing formulas often prevent investors from backing the most successful new founders. This problem is even more acute as AI changes how companies scale.

Rather than obsessing over early revenue headlines, Accel prioritizes real, repeat engagement: how quickly users see value and how deeply a product becomes part of daily work.

The problem with strict plans

Finding lasting companies means investors sometimes must abandon their safety nets. When firms refuse to adjust pricing models they can reject talented founders just to feel financially secure.

Taking a calculated risk often yields better results than strictly following a textbook. Clements recalls losing out on early investments in hugely successful companies like ServiceTitan and Rippling because the initial asking prices seemed too high at the time.

“We lost it because we sort of got cute on price,” he notes regarding ServiceTitan, “and then that went on to be a US$9 billion company.”

Dealing with market changes
Missing out on these massive successes can easily cause investors to panic and abandon their entire strategy. Following a big missed opportunity, many firms either overpay for hyped up companies or completely stop investing while they wait for the market to calm down.

To avoid this emotional rollercoaster, a solid investment plan must focus on the basics of a good business instead of constantly chasing immediate wins.

Using a baseball metaphor, Clements shares his firm’s guiding philosophy: “Focus on hitting singles and doubles, and let the home runs take care of themselves.”

Rethinking a product’s value

This focus on the basics is especially critical today because traditional sales figures are no longer a reliable measure of success in the AI market. When early growth numbers are artificially inflated by excitement, judging a new business with old rules becomes impossible.

Instead, investors need a new way to measure how deeply a product integrates into its users’ daily lives. Clements suggests looking at two specific areas to separate temporary hype from genuine value:

• Time to value: measures how quickly a new customer can start using a tool and immediately see its benefits.
• Durability of value: measures how much the tool becomes a required part of daily work, which keeps customers from switching to a competitor.

Many early AI tools got a lot of attention but quickly lost users because people simply did not stick with them. However, software designed specifically for computer programmers proved to be different.

Finding the real starting point